WPS4376

P olicy R esearch W orking P aper 4376

Absenteeism and Beyond:

Instructional Time Loss and Consequences

Helen Abadzi

The World Bank

Independent Evaluation Group
Sector, Thematic, and Global Evaluation Division
October 2007
Policy Research Working Paper 4376

Abstract
Studies have shown that learning outcomes are related Pernambuco, 71 percent in Morocco, and 78 percent in
to the amount of time students engage in learning tasks. Tunisia.
However, visits to schools have revealed that students Instructional time use is a mediator variable that is
are often taught for only a fraction of the intended time, challenging to measure, so it often escapes scrutiny.
particularly in lower-income countries. Losses are due Research suggests that merely financing the ingredients of
to informal school closures, teacher absenteeism, delays, instruction is not enough to produce learning outcomes;
early departures, and sub-optimal use of time in the students must also get sufficient time to process the
classroom. A study was undertaken to develop an efficient information. The quantity-quality tradeoff that often
methodology for measuring instructional time loss. Thus, accompanies large-scale enrollments may be partly due to
instructional time use was measured in sampled schools instructional time restrictions. Time wastage also distorts
in Tunisia, Morocco, Ghana, and the Brazilian state of budgetary outlays and teacher salary rates. To achieve the
Pernambuco. The percentage of time that students were Millennium Development Goals students must get more
engaged in learning vis-à-vis government expectations of the time that governments, donors, and parents pay
was approximately 39 percent in Ghana, 63 percent in for.

This paper—a product of the Sector, Thematic, and Global Evaluation Division (IEGSG), Independent Evaluation
Group (IEG)—is part of a larger effort in the department to develop innovative indicators and improve the assessment of
learning outcomes in the Bank's education lending portfolio. Policy Research Working Papers are also posted on the Web
at http://econ.worldbank.org. The author may be contacted at [email protected].

The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development
issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the
names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those
of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and
its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent.

Produced by the Research Support Team

Acronyms
BNPP Bank-Netherlands Partnership Program - Global and Regional Initiatives
CNIPRE Centre National de l’Innovation Pédagogique et des Recherches Educatives
(National Center of Innovation and Educational Research)
EARC Educational Assessment and Research Centre (Ghana)
EFA Education for All
FTI Education for All Fast-Track Initiative
HDNED World Bank Human Development Network
IBE International Bureau of Education (UNESCO)
IEG Independent Evaluation Group
OECD Organization for Economic Cooperation and Development
PASEC Programme d’analyse des systèmes éducatifs
PPAR Project Performance Assessment Review
PIRLS Progress in International Reading Literacy Study
SAEPE Education Evaluation System of the State of Pemambuco
SACMEQ Consortium for Monitoring Educational Quality
SPSS Statistical Package for the Social Sciences
TIMSS Trends in International Mathematics and Science Study
WBI World Bank Institute
USAID United States Agency for International Development
UNESCO United Nations Educational, Scientific, and Cultural Organization
 i

Contents
ACKNOWLEDGEMENTS......................................................................................................III

EXECUTIVE SUMMARY ....................................................................................................... V

1. THE PEDAGOGY OF POVERTY ...............................................................................1

2. THE EFFECTS OF INSTRUCTIONAL TIME USE: A LITERATURE REVIEW..........4

Time Loss at the School Level ............................................................................................... 8

Teacher Absenteeism ............................................................................................................. 9
Use of Classroom Time (Time on Task) .............................................................................. 13
Time Spent Learning Prescribed Curricula (Academic Learning Time) .......................... 16
Relationships between Instructional Time, Poverty, and Achievement .......................... 16

3. INSTRUCTIONAL TIME DATA COLLECTION AND FINDINGS .............................21

How Much of the Intended Time Was Spent on Learning Activities?.............................. 22

How Often Were Teachers Absent or Late? ....................................................................... 23
Was There Recuperation of Canceled Classes? ................................................................ 25
What Policies Would Increase Teacher Attendance and Punctuality?............................ 26
How Often Were Students Absent?..................................................................................... 26
How Much of the Classroom Time Was Used for Learning Activities? ........................... 27
How Much Homework Did Students Complete? ................................................................ 30
Trends in Linkages between Learning Outcomes and Poverty Status............................ 31
Methodological Lessons and Data Limitations .................................................................. 31

4. IMPROVING INSTRUCTIONAL TIME USE: POLICY IMPLICATIONS ...................33

Economic Implications of Instructional Time Use ............................................................. 34

Teacher Attendance and Accountability............................................................................. 36
Effective Supervision - Strengthening the Supervisory Chain......................................... 37
Improving Local Governance and Stakeholder Involvement............................................ 39
Training and Consciousness Raising on the Consequences of Time Loss ................... 39
Options for Increasing Student Achievement Given Limited Time Use .......................... 41
Instructional Time Measurement for Evaluation and Research ....................................... 42

REFERENCES......................................................................................................................73
 ii

Tables
Table 2-1. Instructional Hours in African Countries Using “Split-Shift” Schooling .................................................. 9
Table 2-2. Provider Absenteeism Rates by Country and Sector...........................................................................12
Table 2-3. Classroom Activities during an IEG Mission Visit in Brazil (2002).......................................................14
Table 3-1. Instructional Time Use in the Four Countries of the Study...................................................................23
Table 3-2. Solutions adopted by Tunisian and Moroccan principals during teacher absence...............................26
Table 3-3. Percentage of Time Spent in Instructional Tasks................................................................................27
Table 3-4. Average percentages of Time Use in 4th Grade (common for all countries) .......................................30
Table 3-5. Homework Efforts Reported by Students and Parents ........................................................................30

Figures
Figure 1. Expected and Actual Number of Days Students were Engaged in Learning ......................................... vii
Figure 1-1. Instructional Time Loss Model ............................................................................................................. 2
Figure 2-1. Instructional Time Indicators in Basic Education of Some Middle Eastern Countries......................... 6
Figure 2-2. Argentine students carrying textbooks from the library during class time ............................................ 8
Figure 2-3. Not all absenteeism is illegal; teachers may show up but not teach when expected.........................11
Figure 2-4. Call for a strike of indefinite duration in Argentina ..............................................................................12
Figure 3-1. Percentage of time lost at the country level ........................................................................................22
Figure 3-2. Expected and registered number of hours taught in Pernambuco......................................................25
Figure 3-3. Interactive teaching: brief lecture, practice, discussion......................................................................28
Figure 3-4. Whom were teachers addressing in class? ........................................................................................29
Figure 3-5, Use of Instructional Materials by Country ...........................................................................................29
Figure 4-1. Classroom instruction as a black box mediating input and outcome variables ...................................33
Figure 4-2 Student groups engaged and supervised (Indonesia) .........................................................................38
Figure 4-3. Groups unsupervised, engagement uncertain (Brazil)........................................................................38
Figure 4-4: Indonesia: “Benefit from work time as much as possible”..................................................................41
Figure 4-5. How many students seem off task in this caricature?.........................................................................42

ANNEXES

ANNEX A. STUDY ACTIVITIES, METHODOLOGY, AND DETAILED FINDINGS ............ 45

ANNEX B. SURVEY FORMS AND INSTRUCTIONS .......................................................... 64

ANNEX C. POLICY DISCUSSION EXERCISE .................................................................... 72

 iii

ACKNOWLEDGEMENTS
The author would like to acknowledge the contribution and kind collaboration of colleagues
in agencies and countries where the research took place: Fatima Merrichko, Souad
Abdelwahed, and the late Moncef Mahersi (Centre National d’Innovation Pédagogique et de
Recherches en Education-CNIPRE, Tunisia); Mohammad Sassi, Said Boudega, and El-
Mokhtar El-Hadri (Ministry of Education, Morocco); James O. Afrani, Edward Dogbey,
Dorothy Konadu (Ministry of Education, Ghana); Secretary Mozart Neves Ramos, Teresa
Notario, Maria Maciel, Arnildo Ananias de Oliveira (Coordenação da Secretaria Executiva
de Desenvolvimento da Educação and the Gerência de Avaliação, State of Pernambuco,
Brazil).

Thanks are also due to the staff of UNESCO-International Bureau of Education: Massimo
Amadio, Aaron Benavot, and the late director Cecilia Braslavsky; to World Bank staff Paul
Gertler, Robin Horn, Halsey Rogers, Michel Welmond, Linda English, Andrea Guedes, Deon
Filmer, Yuri Obara, Benoit Millot, and Dandan Chen; to academics Jane Stallings, David
Markham, Stephanie Knight, and Eugene Schaffer, as well as research assistants Kenny
Khoo and Aashti Zaidi.

Finally the author thanks the Government of the Netherlands for the grant that made this
study possible.
 v

EXECUTIVE SUMMARY
To fulfill the millennium development goals and attain universal primary school completion by
2015, the donor community has made intense efforts to increase the education budgets of low-
income countries so that they can pay for buildings, salaries and materials to provide quality
education. The attainment of the Education for All goal depends on the ability of countries to
translate financing into time and opportunities for students to learn.

However, the delivery of basic educational services to the poor often falls far short of
expectations. Curricular objectives must be covered within specific numbers of hours or days.
But in low-income areas worldwide, from Brazil to Niger, only a fraction of the intended
instructional time is used for learning tasks. Schools may close informally before or after
holidays, start late in the day or end early. When teachers are present, they may be engaged in
activities other than teaching. One outcome of limited instruction is an inability to read fluently
(if at all) until the later primary grades. Lacking parental support, poorer students tend to fall
behind early and repeat grades or drop out illiterate.

To assess the magnitude of the problem and find methods to measure it, a team of World Bank
staff obtained grant funds from the government of the Netherlands. First the existing research
was reviewed, that highlighted the following issues:

y Worldwide, governments define the number of days or hours that schools should teach
specific material, usually 850-1,000 instructional hours or 180-220 days per school year,
aside from breaks and extracurricular activities. (In poor countries, however,
overcrowded classes may be split into two, with resulting 40% loss in instructional time
and reduced learning outcomes.) Instructional time surveys are few in lower-income
countries and may have measured just one aspect of time loss. For example, schools were
found to be open 70% of the official time in Mali and 114 days of the official 200 days in
Honduras, while a Dominican study reported an overall time use of 65%;

y Instructional time wastage is often linked to impoverished environments. Studies in the

US found that schools serving the poor often have lower time-on-task and spend this time
less effectively. For example, socioeconomic differences were found in the amount of
time students spent learning to read, with classroom interruptions and disruptions more
frequent in low-income areas. Poorer students were found to spend 5% less time per day
engaged in academic tasks than schools in better-off areas and need to attend school 1.5
months during the summer break in order to attain an equivalent amount of engaged
learning time; time wastage accumulated over the years places poorer students at risk;

y Teacher absenteeism has figured prominently as the culprit in the wastage of education
investments. Surveys have shown figures ranging from 11% in Peru, 21% in Indonesia,
27% in Uganda, 30% in Kenya. Overall, 20-25% of school staff may be absent on any
given school day, with higher rates in rural schools.
 vi

The absenteeism findings have received much attention and have led to calls for increased
teacher accountability and incentives to improve attendance. But absenteeism is only one of
the reasons for limited instruction. School closings and poor instructional time use inside the
class are important ones as well. Wastage in small amounts constitutes significant loss of
time in the long run. For this reason, an instructional time loss model was developed from
research evidence, and some parameters were estimated.

Intended class time as allotted by a government (e.g., 200 days, about 1000 teaching hours)
Remaining after school closures (strikes, weather, inservice training, extra
holidays, weather)
Remaining after teacher absenteeism and tardiness

Remaining after student absenteeism

Class time devoted to any learning
task
Learning time relevant
to curriculum

How much of the theoretically available school time is in fact used to engage students in learning
activities?

The study studied instructional time loss in four countries: Ghana, Morocco, Tunisia, and the
Brazilian state of Pernambuco. A methodology was developed in 2003, and data were collected
in 2004-2005 through surprise school visits, classroom observations, and surveys. Instructional
time losses varied by country. In Tunisia, the country with the most efficient use of time, students
were engaged in learning 79% of the available time. In Ghana, however, students were engaged
in learning only 39% of the time, and in Pernambuco 63% of the time.1 Morocco had indicators
somewhat lower than those of Tunisia (using 71% of time for learning), but school closures may
have been underreported. Translated into number of days effectively available for learning, losses
are palpable. For example, only 76.3 days were devoted to learning tasks of the 197 officially
available to Ghanaian students, while for Tunisian students 148.1 of the 190 days officially
available were devoted to learning tasks. In effect, Tunisian students get twice as much of the
intended classroom time as Ghanaian students.

Within each country, time use among schools showed substantial variability. Most of them used
the time rather well, but a minority had substantial losses and skewed the averages. Across
countries, teacher delays and absences (legitimate or otherwise) resulted in losses of 11.5 to 43
days during the school year. Classes lost were rarely recuperated. The number of days schools
were closed for various reasons during the school year was reported as low, from 1.4 day in
Morocco to 4.8 days in Pernambuco. However, no details were asked about the timing and
reasons for closures or about partial cancellations of classes, so these figures may be
underestimates. For example, in Morocco schools may cancel classes for several days so that
students can study for final exams.

1. In the summary and abstract percentage of time loss figures have been rounded for brevity.
 vii

The most significant block of time lost tended to be inside the classroom. Ghanaian teachers
spent 70% of the time engaging students in learning, while Tunisian teachers spent 87% of the
time. The non-instructional time (some of which is necessary) was taken up with organization or
management activities that included giving instructions, textbook distributions, socialization, or
the teacher being out of the room. It ranged from 13% in Tunisia to 28% in Ghana.

Figure 1. Expected and Actual Number of Days Students were Engaged in Learning

250

200 195 204 203

197 194 190 185
200
175 176 171
145 148
150
126
Days

109
100
76

50

0
Pernam buco (Brazil) Ghana Morocco Tunis ia

Days in school year Days af ter Closures No. days schools operated Days students engaged in learning

Engagement rates of 70-80% in class may seem substantial, but the activities observed might not
result in optimal learning. Lecturing made up much of the interactive learning time, which if
prolonged can result in limited student attention and subsequent recall. Instructional materials
could focus attention and help students retain specific items, but they were infrequently used.
Students were taught mainly through blackboard and notebooks, with manipulables and
audiovisual equipment used only about 3% of the time. Textbooks were used from 23% of the
time in Pernambuco to 7% of the time in Ghana, where few exist. The average class size was
modest (from 25 students in Pernambuco to 31 in Morocco) and could allow for interactions with
individual students. However, teachers in all countries addressed the entire classroom 70-90% of
the time, so some students had limited opportunities to get feedback and reinforcement for
learning. Furthermore, less than 5% of the time was spent in group activities in any country.
Overall, students mainly listened to subject matter through “chalk and talk” presentations; they
had few opportunities to contemplate it through activities that would result in better recall.

Possibly due to poorly organized learning activities, 19% students in Ghana and 21% in
Pernambuco were found to be “off task”, that is uninvolved with class activities. By comparison,
only about 10% of Tunisian students and 9% of Moroccan students were uninvolved during
class. Disengagement from learning activities (student off-task rate) has been linked to
achievement outcomes in US research. Therefore, Tunisia and Morocco offered not only more
class time in learning tasks, they also engaged a higher percentage of students in learning than
Pernambuco and Ghana. Arguably the former countries teach students more efficiently than the
latter.
 viii

Test scores available for schools in Pernambuco showed significant relationships between
achievement and various time indicators. Socioeconomic data, available in Pernambuco and
Morocco, also suggested relationships with instructional time measures. Particularly in Morocco,
poverty rate and rural school locations were positively correlated with school closures, interactive
and passive instruction, time spent in management, and off task behaviors. In general, interactive
learning time was negatively correlated with urban poverty. Though these findings are
preliminary and do not establish causality, they are in accordance with earlier research suggesting
that the schools of the poor offer less instructional time to their students.

This study was implemented in only three countries and one state. It did not have a large sample
of very low-income countries where textbook scarcity forces children to spend most of their time
copying from the blackboard. Nevertheless, it has shown that under some circumstances the time
and opportunity to learn are very limited. This phenomenon may account to some extent for the
disappointing learning outcomes that the poor children show worldwide. Making better use of
instructional time is key to achieving Education for All.

Thus, time is a mediator variable that has escaped scrutiny and measurement thus far. It is not
enough to provide the ingredients of instruction and assume that they will be used in class.
Students must get sufficient time to master the instructional objectives intended in specific
subjects. Furthermore, inputs like teaching aids must be employed within the timeframe
available to students, or they may not promote student learning. Other sources of systemic
inefficiency exist, such as dropout, repetition, and high staffing costs. But these may be causes or
effects of time wastage. If so, the time devoted to learning the material prescribed by the
curriculum may be at the crux of educational “quality”. The quantity-quality tradeoff that is
often mentioned in relationship to rapid expansion of education in low-income countries may be
mitigated if measures are taken to give students the instructional time they need, even as class
sizes increase.

Equally important, instructional time loss has budgetary implications. A class hour in a
particular school corresponds to a slice of a country’s recurrent expenditure budget. In some
countries only a small fraction of government and donor resources may be converted into
learning activities. To ensure that government and donor investments have the expected impact,
policy dialogue must focus on the processes needed to supervise and improve time use at all
levels. Research on instructional time in low-income countries is very limited, but the scope is
large. Issues with policy implications include:

y Monetarizing the losses and consequences of time loss, and adjusting economic
indicators on this basis;
y Collecting data on the policies and practices that result in school closures, such as late
teacher assignments, payment policies, class cancellations for impromptu training,
cultural events, and lengthy exam preparation; developing effective action plans to
minimize them in each country.
y Making principals explicitly accountable for keeping schools open and monitoring
teachers’ presence and classroom instructional time; establishing accountability with
district staff who interface with them to monitor time use at the local and central levels;
y Ensuring that all students in all grades have textbooks that they can take home and use for
homework and that are available during class to facilitate time use; informing
 ix

governments about the time use problems that arise when just a few textbooks are
available for classes.
y Negotiating with and informing the teacher unions about the consequences of time loss
on the poor, publicizing local-level linkages between time use and academic
performance, exerting social pressure against political influence that enables non-
performing teachers to continue drawing salaries.
y Specifically training teachers to use time well, particularly through dissemination of
effective classroom management practices that can keep students occupied all the time;
training principals and supervisors to recognize the major components of instructional
time use with brief observations and to use this system in their work.
y Publicizing the cost of time loss at the local level and making communities aware of how
well children’s time is used. In particular sensitizing lower-income parents to the signs
that show whether their school teaches efficiently and to the simple means they can use to
ascertain whether their children learn basic skills (e.g., oral reading fluency).
y If feasible, structuring development policy lending disbursements against significant
improvements in instructional time use.

Instructional time parameters as monitoring indicators. The instructional time parameters are
useful for monitoring institutional performance and accountability. Well- managed schools
should stay open the number of days that governments specify, have few teachers absent, and
ensure that students of absent teachers are engaged in learning. Visitors should find school
registries with numbers of days and hours per subject matter approximating those specified by
the government. Reduced teacher absenteeism should constitute evidence of effective school-
based management. Finally, time on task in class should constitute evidence of effective
supervision by principals, inspectors, or supervisors. Additional work is needed to develop
modular surveys that accurately and inexpensively estimate instructional time loss. The benefit
may be worth the cost and effort. Measurements obtained in this study suggest that an overall
80% use of time could be used as a standard to attain for lower-income countries.

The effects of time use must be tested more extensively vis-à-vis learning outcomes. Much
needs to be learned about the effects of time use on the delivery of curricula, on test scores and
on reading fluency acquisition in the all-important early grades. Memory issues, such the rate
and efficiency of consolidating new information need to be studied in relationship to the amount
and distribution of learning time, in order to identify the activities that maximize long-term recall.

The future of the Education for All initiative may depend on how seriously governments and
donors take instructional time wastage. Hopefully this research is only the beginning.
 1

1. THE PEDAGOGY OF POVERTY

Slowly and laboriously the car traverses the sandy road to a small town two hours away
from the capital of Niger, in search of schools built by a World Bank project. A newly
constructed building appears ahead, and the Bank’s evaluation mission eagerly approaches,
but in mid-morning it is closed. Half an hour later another school appears, with several
children playing in the yard. A teacher sees the car approaching and calls them inside. The
mission enters a room with bare walls and a single textbook that is on the teacher’s desk.
Words in French are written on the blackboard, and visitors ask students to read them.
Two students seem to decipher them, but when the same words are written elsewhere on
the blackboard, they cannot. The teacher says that he was assigned to the school only a
month before, and that the class had not functioned for several months.

A World Bank evaluation mission visits a school in rural Bangladesh. The school is closed
when the mission arrives and opens an hour behind schedule. The principal explains that
because of the rain, students could not walk out of their houses earlier. But only about 25%
of the registered students finally arrive, and the classrooms clearly have no seating space for
all those who are registered. It is because of the rain that students are absent, teachers say.
Just yesterday, most of them were there. Probably tomorrow they will all be there again.

A researcher with a group of university students learning to conduct classroom observations

requests permission to visit classes in a primary school of Pernambuco in northeast Brazil.
Permission is granted, but when the team arrives in the school, they find no students or
teachers. The principal explains that she misunderstood the request and thought that the
building was needed for a function, so she canceled all classes for two days. The group then
proceeds to a large secondary school in downtown Recife. They arrive in the first hour of the
afternoon shift. The students noisily loiter the hallways because many teachers are late.
Inside other classrooms students are talking to each other in groups, but it is unclear whether
they are doing group work or just chatting. In two classrooms, teachers sit on the desk
correcting papers, while secondary-level students (who receive no textbooks from the
government) are copying math problems and language texts from the blackboard. In a school
of about 12 classrooms, only one had a teacher clearly interacting with students on an
instructional topic. After spending an afternoon in search of instruction to observe, trainees
could hardly get any practice.

The Independent Evaluation Group (IEG), a semi-autonomous department of the World

Bank in charge of evaluating completed projects, visits many schools and observes classes.
Time use in many countries has been a cause for concern. Schools may close informally,
sometimes for local functions. The school day may start late or end early for various reasons,
and schools may be effectively operating for only 2-3 hours a day per shift. Teacher
absenteeism is rampant, due to strikes, legal leave, departure to home towns before holidays,
late arrivals and early dismissal of classes. And when teachers are in school, they may be
chatting among themselves and abandon classes or allow inordinately long recesses. During
class, much of the time may be spent taking attendance, handing out textbooks, doing small
chores, and copying to and from the blackboard. As a result, public schools of low-income
 2

areas may offer students just a fraction of the time governments and donors put at their
disposal. The time wastage constitutes failure to deliver a basic service to the poor. 2

A striking symptom of this failure in some countries is students’ inability to read fluently (if
at all) until the later primary grades. 3 Lacking parental support, poorer students tend to fall
behind early and repeat grades or drop out illiterate. In middle-income countries like Brazil,
poor time use means that some students are surrounded by piles of books that they cannot
read. Performance is often worse in multilingual countries that must teach in official
languages, particularly languages with complex spelling systems like English, French, and
Portuguese; there is simply not enough time to teach the basics.

When and how is time lost? International studies have mainly focused on teacher
absenteeism (Chapter 2). But absenteeism constitutes only one source of loss in the delivery
of basic educational services. Students’ opportunity to receive information and to process it
is reduced in multiple stages. Sometimes the loss happens in large blocks, due to strikes or
late teacher assignments. But time is frequently lost in small increments; a day here, an hour
there. It is lost minute by minute when teachers fail to start work promptly in class, distribute
textbooks, or leave the room to chat with others (Figure 1.1). Kids are often happy to have
extra vacations or to get off school early, so no one may complain. However, they are the
ones who ultimately lose the most; they may sacrifice current earnings to spend years
nominally attending school and lose future earnings as well as social status if they lack the
skills needed for the labor market.

Figure 1-1. Instructional Time Loss Model

Intended class time as allotted by a government (e.g., 200 days, 1000 teaching hours)
Remaining after school closures (strikes, weather, inservice training, extra
holidays, weather)
Remaining after teacher absenteeism and tardiness

Remaining after student absenteeism

Class time devoted to any learning task

Learning time relevant to
curriculum

The wastage of instructional time should be an important concern for those who finance
education. Governments, donors, and parents pay for educational inputs, buildings, salaries,
teacher training, textbooks, and materials in order to provide quality education. The
financing is expected to translate into time and opportunities for students to learn, and it is
commonly expected that 100% of the investments will be used for student learning. Thus, a
class hour in a particular school corresponds to a budgetary fraction given the amount of time
schools officially operate (for about 4-5 hours in about 200 days – see Chapter 2). The
degree of budgetary wastage resulting from time loss has probably not yet been calculated.

2.World Development Report 2004.

3. IEG 2006.
 3

Time wastage is rarely measured through all the stages in which it is lost, yet such measures
are feasible. Unlike many educational measures, time is a well-understood concept that also
has equal units and an absolute zero. It also offers the possibility of international
comparisons. A variety of methods have been used to assess instructional time, and many
studies have focused on specific stages of time loss. Measuring the multiple steps of time
loss can be a complex undertaking. Documenting large blocks of time loss (such as strikes)
may be easy, but measuring smaller amounts of time requires more intricate methodology.

The opportunity to assess an integrated model of time loss arose with a grant from the Bank-
Netherlands Partnership Program for Global and Regional Initiatives (BNPP). 4 Research
undertaken during this grant attempted to answer the following questions:

a) How much of the classroom time do students spend engaged in learning compared
to the theoretically available time?
b) What is the relationship between the various measures of time use and academic
achievement?
c) What policy options will improve use of time at central government and at local
levels? And what costs and savings might be associated with better time use?
d) How can the amount of time spent on learning be measured in ways that are
simple and easy to administer, reliable, comparable across countries and
monitorable across time?

Rationale for IEG involvement. IEG had raised the issue of instructional time use
repeatedly during Project Performance Assessment Reviews (PPARs) in many countries,
but there was no easily usable methodology to permit a holistic measurement of time
loss. The BNPP grant made it possible to develop it and study its feasibility. So the study
took place in countries where PPARs had been conducted (Tunisia in 1998 and Morocco
in 2001), as well as Ghana, which had been the focus of a 2004 impact evaluation. To
pilot the use of instructional time data as a baseline for later project evaluation, the
Brazilian state of Pernambuco was included in the study. An education project had been
prepared and was about to become effective as the study started. In Pernambuco,
supervisors were trained after data collection on techniques that would help increase
instructional time in class. The effects on learning outcomes would be studied at project
completion, around 2010.

Subsequent chapters provide the research background, results, and methodological challenges
involved in the measurement of instructional time loss. A follow-on, more detailed study,
also financed by the Government of the Netherlands, was under implementation as this
document was being prepared.

4. A team of staff members (Helen Abadzi-IEG, Robert Prouty-HDNED, and Benoit Millot-SASHD)
received a grant of US$227,500, that became effective on July 23 2003 and was completed on June 30
2005, after an extension of one year. The staff volunteered their time for this project and could only travel
in connection to grant activities during trips undertaken for other purposes.
 4

2. THE EFFECTS OF INSTRUCTIONAL TIME USE: A

LITERATURE REVIEW
In the 1960s, some economically oriented studies examining the inputs and outputs of
schooling in the US concluded that classroom events made little difference in the lives of
children; the school was essentially a “black box” between the inputs and the outputs. 5
However, this view was challenged soon. An early model of school-based learning 6
conceptualized achievement as a combined outcome of two time variables: (a) the
amount of time a learner is engaged in learning (i.e. the time a pupil is involved in
learning tasks); and (b) an individual’s learning rate (i.e. the amount of time needed for
learning). Mathematical learning models also tried to show that optimal learning
performance could be achieved by giving students sufficient time to learn. 7 It was
understood that with sufficient time, students should retain what they practice and think
about. 8 In the late 20th century, neuroscientific research led to a better understanding of
learners’ need for prolonged and repeated exposure to stimuli. The cellular process
called long-term potentiation, for example, helps increase the strength of synaptic
connections that result in learning. 9 Thus, the rationale for prolonged and distributed
learning sessions with feedback and contemplation of the material rather than one-time
exposure or mere “quality time” has been firmly established.

The instructional time concept (sometimes called “opportunity to learn”) spurred a wave of
research in the 1970s and 1980s in the U.S. Researchers 10 attempted to quantify and
compare learning time in various instructional settings, and to measure loss due to
disrupting factors (e.g., inadequate educational policies, poorly organized or delivered
class lessons). Time loss was documented through classroom observation instruments (e.g.
those of Stallings, Virgilio, and others). Concerns about using time well entered pre-service
and in-service training during the 1980s, under mottos such as “Keep students occupied all of
the time!” Classroom management techniques have been developed to help teachers control
the class and use time well.

How is time spent in school? Primary school curricula can be classified into six subject
areas: language, mathematics, natural science, social sciences, aesthetic education and
physical education. These subject areas receive between 80% and 90% of overall
instructional time during the first six years of schooling. On average, one-third of all
instructional time in primary schools is devoted to language instruction; of this, about 25% of
time is for national/official languages and 8% to foreign languages (local languages are
infrequently taught). About 20% of instructional time is devoted to math. The mean
instructional time devoted to the arts, sciences, physical education and the ‘social
sciences’ (history and geography) is usually 10% for each subject area. Some systems
may also include religious/moral education, hygiene/health education, vocational education/

5. Coleman 1966, Jencks 1972

6. Carroll 1963
7 Atkinson 1972
8. Doyle 1983, Ben-Peretz and Bromme 1990
9. Rabenstein et al. 2005. See Abadzi 2006 for a review of neuropsychological issues, including feedback
and memory consolidation.
10. For example, Smith 2000 and Wang 1998
 5

practical skills.11 In secondary education, the trends are similar. At the global level,
instructional time is devoted to languages (30%), math, science, and computer technology
(about 30%), and social sciences (about 13%). The remaining time is devoted to arts,
physical education, vocational skills, and religious/moral education. In addition to academic
subjects, middle-income countries and private schools may occupy students during lunch and
offer extracurricular activities. For example, US elementary schools may only spend 65-70%
of the day in teaching academic subjects.12 (These extracurricular hours were taken not taken
into account in this study.)

How much instructional time are schools expected to offer students? Students ought to go to
school for the amount of time necessary to master instructional objectives in each grade.
Education ministries define the number of days or hours that schools should function and
teach specific material (Figure 2-1). Guidelines are typically found in the weekly timetables
of various countries and in policy documents regarding the length of the school year, the
school day and the class period. The 2005 UNESCO EFA Global Monitoring Report
recommends 850 to 1,000 hours annually (aside from breaks and lunch periods), and the
Education for All Indicative Framework expects at least 850 (or about 200 days at 5 days per
week). 13 Several East Asian countries provide more than 1,000 instructional hours of
teaching (documentation provided by the UNESCO International Bureau of Education).
In terms of days, the length of the school year varies. In the early grades of primary
education (grades 1-4), median instructional hours tend to be higher in the education systems
of Sub-Saharan Africa, Latin America and the Caribbean, and Western Europe and North
America. They tend to be lower in Central Europe and the former Soviet Union and, to a less
extent, in East Asia and the Pacific and the Arab States. 14 In Pakistan and Nepal, the
primary school year lasts for 180 days, rising to 190 days in Zambia, 200 days in
Bangladesh, and 220 days in India. In terms of actual rather than instructional hours, the
global mean may range from 705 hours in grade 1 to about 830 hours in grade 6. 15

11. These curricular structures have remained remarkably stable between 1920 and 1985, but the proportion
of instructional time devoted to ‘modern’ subjects such as mathematics, natural sciences and foreign
languages has increased and the teaching of history, geography and civics as separate subjects has been
reduced in favor of the more interdisciplinary ‘social studies’ (Benavot and Amadio 2004, Benavot 2006)
12. Perie et al. 1997; Roth et al 2003
13. UNESCO 2005; an instructional hour lasts 40-50 minutes or sometimes less (as in Ghana, where it lasts
30 minutes). Some data suggest 750 hours mandated each year in primary-school classrooms (Amadio
1998; UNESCO-IBE 2000). during the first 2 years of primary education countries mandate on average
710-740 hours of instructional time per year. Intended instructional time increases in each subsequent grade
level and reaches approximately 900 annual hours in grade 8. This pattern translates into an average
supplement of about 25 annual instruction hours per grade level, although these increases are not linear.
There are significant jumps during grades 3-5, and then again between grades 6 and 7, when the transition
between primary and lower secondary education typically occurs.
14. Benavot and Amadio 2004.
15.Benavot 2002
 6

Figure 2-1. Instructional Time Indicators in Basic

Education of Some Middle Eastern Countries

Sources: Amadio 1997, Millot and Lane 2002.

How efficiently can instructional time realistically be utilized? Curricular objectives are
prescribed for coverage within specific timeframes, so the implication is that students
should be engaged in learning 100% of the time. In addition, homework would
supplement classroom time. In reality, it is impossible to use 100% in instruction;
children and teachers will interact socially some of the time. Losses are limited in
industrialized countries because they have reasonably effective administrative
mechanisms to regulate school time and monitor teacher compliance. In these countries,
the school year is implemented as planned, substitute teachers are called for absent
teachers, and parents are attentive to students’ progress. Thus, the school-level sources of
time loss are near zero, and the challenge is to maximize time on task. For this reason,
instructional studies in the US focus on how classroom time is used. Measures have varied
broadly, with studies in the 1970s and 1980s showing on-task time from 38% to 96% of
available time.16 Time use has progressively increased. A longitudinal study of eight
elementary schools in Chicago, for example, found that 85% of the daily allocated time was
dedicated to instruction.17 It may be difficult to exceed this efficiency rate, and this may be
one reason why U.S. educators have lost interest in this concept since the 1990s.

Outside the industrialized countries, however, concerns have grown that curricula cannot
possibly be covered in the amount of time students spend engaged in learning at school or
even with private tutoring at home. Sources of ‘leakages’ may be multiple, and voluntary or
imposed. Financial, cultural, and political events18 reduce schooling time in ways that are
rarely seen in industrialized countries. How much deviation from the 100% standard would
still enable curricular objectives to be reasonably well achieved in various countries? No
research has been found on this topic.

16. Smyth 1985; Anderson, Ryan and Shapiro 1989; Fisher et al. (1978) estimated that academic learning
time in the US amounted to about two-thirds of total engaged time.
17. Smith 2000. Other studies of the 1990s are local and may not be published in peer-reviewed journals.
18. Millot, 1994, Millot and Lane 2002.
 7

Though statistics may not be directly comparable, systematic surveys as well as ethnographic
studies have shown surprising losses and highlight the issue of service quality to the poor.
Some examples are:

Senegal. A study undertaken by the inspectors of the Académie de Diourbel found a teacher
absenteeism rate of about 30%, with strikes initiated by both students and teachers. 19 In the
final secondary grade and before the final examinations, almost five weeks of class time were
lost between October and March. With a teacher absenteeism rate of 32%, the deficit
amounted to 112 hours or about 14 weeks. In the philosophy course only 4 of the 23
chapters were covered, in physics 7 of the 17 chapters, and in chemistry 7 of the 11
chapters were covered.

Bangladesh. A 2003 ethnographic study of time use in eight low-income schools found
that schools operated 19% to 55% fewer days than scheduled in the school calendar. The
losses had multiple sources. One month of contact time with students at the beginning
and at the end of the school year were sacrificed to administrative and non-teaching
activities. Rural schools seldom opened at the expected time. They allowed travel time
for non-resident teachers and gave a 1.5-hour break to students attending Koranic school,
but did not stay open later. Inside the classroom, time was not used well, either. Teaching
occupied on average 63% of the class time in the classes observed. (Lecturing occupied
about 83% of that time.) Students also showed high absenteeism; only 43 to 67% of
them were in attendance on the days of surprise visits. Teachers estimated that only about
50% of the children were very regular in their attendance. The study found that the
problems were more prominent in isolated areas. However, schools were more likely to
comply with schedules if they received regular visits by authorities. 20

Dominican Republic. A USAID-financed team visited schools on three different

occasions. It found that on average students were receiving instruction 77% of the time.
Main reasons for time loss were various meetings of teachers with parents, district and
other officials, and training (about 42%). Strikes accounted for 19% and teacher
absenteeism accounted for 8% of the lack of instruction. Overall, schools used 65% of
the intended time in teaching. Afternoon shifts spent less time in teaching (58%) than
morning shirts (73%), a trend that was also shown in the Brazilian state of Pernambuco
during the World Bank study (See Chapter 3).21

Mexico. A study from the 1970s showed that a typical teacher used about 50% of the
scheduled school week on actual instruction. The rest of the time was lost to absences,
leaving early, administrative interruptions, and classroom disturbances. 22

19. Diouf 2005

20. Tietjen et al. 2004
21. EDUCA 2005
22. Munoz Izquierdo 1979. Other estimates with limited data include: In Colombia, off task time was
estimated at about 40 percent (Arancibia 1987). One third of instructional hours might potentially lost
because the teacher’s voice is drowned out by rain on metal roofs in Malawi and less than half the
prescribed time for the early grades may be taught in Indonesia (Lockheed and Verspoor 1992).
 8

TIME LOSS AT THE SCHOOL LEVEL

In all countries, the school calendar consists of class days and scheduled holidays.
Inevitably, some days will be lost due to unforeseen reasons, such as weather. In higher-
income countries, schools are prepared for such events; for example, school districts in
the U.S. may build into the school year four days due to snow buildup and extend the
school year if snow lasts longer. Aside from using substitute teachers when needed,
provisions are made to keep students occupied during periods of in-service training that is
conducted during school hours. In lower-income countries, however, money for
substitute teachers is rarely available, and school operation may be overridden at the local
level. As a result, schools may operate fewer days than expected. Reasons include long
matriculation periods, in-service teacher training, climatic conditions, and infrastructure
in poor condition (e.g., severely leaking roofs or damaged walls that may make some
classrooms unusable). In Mali, for instance, schools were found to function 70% of the
official time. 23 In Honduras, schools were open 114 days of the official 200 in 2001. 24
Surveys in Nepal suggest that schools operate on average for three hours per day, a fact
that halves the teaching time available from over 1,000 hours to just 540 hours. 25 In
Burkina Faso, a minimum of 16% of the official allocated time was lost due to breaks and
exam periods. 26

In poorer countries, instructional time

Figure 2-2. Argentine students carrying
textbooks from the library during class time may also be lost also due to the
inefficiencies involved in large classes
that may exceed 100 students. To deal
with large classes, countries like
Senegal, Guinea, Bangladesh, Ghana
or Niger split a class into two and
divide the time. Thus, a school day of
five periods may be reduced to three,
and the number of instructional hours
provided to students is curtailed. This
is known as split-shift and results in a
32-42% loss of time (Table 2-1). 27
Some studies have reported that this
policy did not significantly impact
Source: Author; Iaccarini agricultural school, Mendoza,
Argentina student performance, 28 but this finding
may be due to an already low quality
of education. 29 Evidence from Guinea and Burkina Faso suggests that the split-shift
arrangement reduces time on task, particularly in the afternoon shift, and ought to have a

23. Kim 1999

24. OED 2004
25. Watkins 2000, p. 112
26. Dia 2003
27. Split-shifting is sometimes called double-shifting, but the interpretation is erroneous. Schools may
operate double and even triple shifts and offer students all the expected instructional hours.
28. Bray 2000, World Bank and IADB 2000, p. 46
29. Linden 2001
 9

negative impact on achievement. 30 For example, in Guinea, students from split-shift

classes scored 3.6 percentage points lower in French and 5.6 points lower in math. 31 An
additional consequence may be higher teacher absenteeism, since this scheme is
demanding and repeating the same information to multiple sections may be tiring. 32 The
original plans in some countries called for the “off-shift” students to be occupied with
schoolwork outside the class. However, it has often been impossible to organize teaching
during “off-shift” hours, and children typically just go home.

Multigrade teaching in low-income countries may also result in reduced instructional time.33
In multigrade classes that are common in small rural schools, students are in school but get
only a fraction of the teacher’s time; if classes have no textbooks or students cannot read
them, they cannot use their time productively when the teacher works with another group.
Strategies to cover more students with fewer teachers or classrooms risk reducing the
quantity and quality of instructional time available for pupils.

Table 2-1. Instructional Hours in African Countries Using “Split-Shift” Schooling

Type/ No. of hours Cote Burkina
Mali Guinea Senegal Average
annually d’Ivoire Faso
Standard Classes 888 747 675 754 858 784.4
Split- Shift Classes 645 585 547 580 603 592
% Difference 37.7% 27.7% 23.4% 30% 42.3% 32.2%
Source: Kim 1999. Amounts average for all grades

How does class size relate to instructional efficiency? The relationship between class size
and student performance is complex.34 Combined with meager school resources large classes
adversely affect teacher recruitment, especially to posts in rural areas and create a vicious
circle.35 But even without splitting shifts, lowering mean class size has substantial costs and
is not necessarily the most efficient strategy for increasing achievement. For example, an
early study conducted in Malaysia revealed that lowering class size would raise student
achievement by 1% but would cost US$50 per student.36 In a study of primary schools in
Israel it was found that increasing instruction hours and reducing class size had significant
effects on students' achievements. It seemed more cost effective to increase use of available
instructional time than to decrease class size.37

TEACHER ABSENTEEISM

Teacher absenteeism has been documented extensively.38 Higher absence rates are predicted
by factors at the community level (remoteness, parents’ education level), teacher level
(teacher’s professional or age-related seniority), and management level (physical

30. Dia 2003

31. Barrier et al. 1998
32. Linden 2001; Suryadarma et al. 2004
33. Attar 2001, Njie 2001, EARC 2003
34. See Abadzi 2006 for a review
35. Onwu 1999
36. Beebout 1972, in Fuller 1987
37. Lavy 2001
38. Chaudhury et al. 2004b, 2004c; Jacobson 1991
 10

infrastructure, multigrade teaching, inspection frequency). Absence rates decline as national

per-capita incomes rise.

As Table 2-2 summarizes, absenteeism in primary schools ranges from 11% in Peru to 27%
in Uganda. Losses can be large. For example, a study of schools in one region of Kenya
found that teachers were absent from school 30% of the time.39 An unpublished study
conducted in a sample of secondary schools in Senegal reported that teachers were absent
approximately 30% of the time.40 An observational study of 120 schools in rural Ghana,
reported that, on average, teachers attend schools four days a week, implying that 5.5
instructional hours are lost each week. On the day that researchers visited schools, almost
one-fifth (19.4%) of the teachers were absent.41 A recent survey of absenteeism in Indonesia
found that out of those reported to be full-time teachers 21% were absent from school; 27%
out of those who were not absent at the time of the interview were out of their class and only
53% of them were engaged in teaching activity.42

Teacher absenteeism studies in South Asia show similar results. A nationally

representative survey of 3750 schools in India showed that 25% of teachers in government
primary schools are absent at a typical point during a school day. 43 Absence rates ranged
from 15% in Maharashtra to 38% in Bihar and 42% in Jharkhand. Another survey in
Udaipur focused on 60 non-formal education centers run by a non-governmental
organization and found teachers’ absence rates of 36%. 44 The PROBE study in India,
which was carried out in 234 randomly selected villages of Bihar, Madhya Pradesh,
Rajasthan, Uttar Pradesh, and Himachal Pradesh in late 1996, found that one-third of the
teachers were absent during research staff school visits. 45 Other surveys of primary schools
in the Indian states of Uttar Pradesh and Madhya Pradesh found that 17% of teachers were
absent from school and additional 20% were in school but absent from class. 46
Absenteeism among primary school teachers in West Bengal was 20%, 47 in Bangladesh it
was 15.5%, 48 while in Pakistan, about 18% of public and private school teachers were, on
average, absent from school. 49 Cross-sectional averages may mask the extent of this
problem because in Pakistan the researchers found that 23.5% of primary school teachers
were absent during at least one of the two visits.

39. Vermeersch and Kremer 2004

40. Diouf 2005
41. EARC 2003 (Educational Assessment and Research Centre)
42. Suryadarma et al 2004
43. Chaudhury et al. 2004b
44. Duflo and Hanna 2005
45. Public Report On Basic Education; De and Dreze 1999
46. Rao 1999; World Bank 2001
47. Sen 2002
48. Chaudhury et al. 2004a
49. Ali and Reed 1994. Subsequent research in rural Northwest Frontier and Punjab showed 14%
absenteeism in public schools; in boys’ schools teacher absenteeism was only 11% and in private schools it
was 9% (Sathar et al. 2005).
 11

Figure 2-3. Not all absenteeism is illegal; teachers may show up but not teach
when expected
Training and maternity leave Teacher off task

Source: author • Source: Barbara Bruns

Why are teachers absent? Prominent stated reasons include teacher participation in
educational workshops, in-service training sometimes lasting for several weeks,50 personal
problems, traveling to administrative centers for salary compensation, and casual leave.51
Pregnancy and housework may affect female teachers.52 Illness in general and HIV/AIDS in
particular have become major causes of absenteeism in certain countries.53 Some research
links poor physical school conditions to absenteeism.54 Another aspect of absenteeism is
refusal to take up postings in desirable locations. For example, in Ghana, a significant
number of newly trained teachers every year refuse posting to rural schools, and many of
those who accept eventually will ask for a transfer to urban schools or quit teaching.55 In
some countries teacher appointments are given as political favors, and appointees may not be
expected to teach.

However, the stated causes for teacher absenteeism often make the fact that the institutional
environment does not really require teachers’ constant presence.56 At the root are limited
authority by schools, lack of interest in enforcing sanctions, and the lack of parental
involvement.57 Absenteeism is often lower among female teachers, among teachers born
in the district where the school is located, and among teachers of children whose parents
are more literate. 58 However, few studies have explicitly explored the linkages between
accountability and absenteeism. Longitudinal research is needed to establish cause-effect
relationships and ferret out the conditions and mechanisms that inhibit or promote teacher
absenteeism.

50. Fairhurst et al. 1999; Dia 2003

51. EARC 2003
52. El-Sanabiy 1989
53. World Bank 2002
54. OED 2004.
55. EARC 2003; the study reported teacher attrition, 12.5% of the overall school staff.
56. Rogers et al. 2004
57. King and Ozler 2001; Lockheed and Verspoor 1992
58. Chaudhury et al. 2005, OED 2004; Lewis and Lockheed 2006.
 12

Figure 2-4. Call for a strike of indefinite duration

Teacher unions and the strikes they
in Argentina proclaim account for various amounts
of time loss worldwide. The conflicts
between governments and the teachers
unions may affect student
performance. In Mexico it was found
that union power is associated with
test outcomes; while low and high
influence of unions was not
significant, medium power was found
to be significant and had a relatively
large negative effect.59 In Argentina,
adversarial political alignments were
associated with a decrease in effective
numbers of class days, with an indirect
negative effect on student
performance in Argentina.60

Private schools for the poor. The

Source: Author, poster in Rio Negro department of education
threat of prolonged teacher strikes and
other forms of absenteeism drives parents to send children to private schools. (Many
teachers also send their own children to private schools.) This is one reason why private
schools for the poor have sprung up in the low-income areas of countries such as India. They
often operate in squalid buildings and charge very little. Since they often operate without
a license, authorities may not know about them. But visitors have found teachers present
and students who seem engaged. This phenomenon has received attention and
engendered calls to promote private schools and support them. Clearly, accountability to
paying parents encourages the school owners to maximize instructional time, and thus
improve students’ performance.
Table 2-2. Provider Absenteeism Rates by Country and Sector
Absence Rates (%) in:
Primary Schools Primary Health Centers
Bangladesh 16 35
Ecuador 14 --
India 25 40
Indonesia 19 40
Papua New Guinea 15 --
Peru 11 23
Uganda 27 37
Zambia 17 --
Source: Chaudhury et al. 2004c; calculations are from facility surveys, except for Papua New Guinea (NRI and World
Bank 2003) and Zambia (Das, Dercon, Habyarimana, and Krishnan 2005).

59. Alvarez et al. 2006

60. Murillo et al. 2002
 13

Teacher tardiness. Data on this issue are limited. For example, a survey of Ghanaian primary
schools reported that in addition to the 20% of teachers who were absent from school on the
day of the survey, another 29% were late in arriving at school.61 The report concluded that as
a result of teacher absenteeism and tardiness, actual pupil-teacher interaction time was
estimated to be only 70% of available instructional time. Reasons for tardiness are unclear
and need to be better understood.

Absenteeism effects on student performance. Teacher absenteeism is particularly costly to the

poor, since they cannot afford private tutors to cover the curriculum and pass high-stakes
examinations. Thus, it has been related to lower student test scores in primary schools.62 One
study showed that a 5% increase in the absenteeism rate of teachers who stayed with the
same class for two years reduced student gains by 4-8% during the year; the size and
precision of these estimates was the same for both math and English.63 In an Indonesian
study, higher teacher absenteeism was related to lower fourth-grade student achievement on
math (but not dictation) after controlling for household characteristics, teacher quality, and
school conditions).64 Absenteeism among teachers also encourages similar behavior among
students, notably in countries such as Mali and Somalia.65

But the relationship between teacher absenteeism and student performance is not always
clear. Some studies find either a weak effect or no effect between teacher absence and student
attendance and test scores.66 A study of primary schools in Pakistan found a 10%
absenteeism rate but no correlation between teacher absence and achievement levels.67
Another study in the Northwest Frontier Province of Pakistan68 found that higher rates of
teacher absenteeism increase student promotion rates for a given level of test scores but
reduce student continuation rates. The explanation is that assessments of pupils’ progress by
absent teachers are less accurate, but students may not know enough material to continue,
even if promoted. However, many studies rely on self-reports and are descriptive. The effect
of teacher absenteeism on student achievement needs to be clarified with studies that have
robust experimental designs.

USE OF CLASSROOM TIME (TIME ON TASK)

Ideally, students should be engaged in learning during the entire time they are class, but this
is often not the case. School-based surveys of teacher activity carried out by PROBE (Public
Report on Basic Education69 ) researchers in four Indian states found that in only 53% of the
schools visited by the research staff were teachers actually teaching in their classrooms. The
survey found that in 21% of the surveyed schools teachers were just “minding the class”; in
18% of the schools they were talking with other teachers; in 11% they were sitting/standing
outside the room; in 7% they were in the head-teacher’s room; and in 23% of the schools
teachers were observed in other non-teaching activities. Similarly, an evaluation by the

61. EARC 2003

62. Suryadarma et al 2004, Chaudhury et al 2004a
63. Das et al. 2005
64. Study cited in Lewis and Lockheed 2006, p. 67.
65. Lockheed and Verspoor, 1992; EARC, 2003
66. Ehrenberg et al. 1991, Chaudhury et al 2004b
67. Reimers 1993
68. King, Orazem, and Paterno 1999
69. De and Dreze 1999
 14

Independent Evaluation Group (IEG) of the World Bank in Brazil found fewer instructional
activities carried out during school visits in lower-income states (Table 2-3). About 92% of
the activities in the higher-income state of Goias focused on basic skills acquisition in
comparison to only 35% activities in Ceará.

Table 2-3. Classroom Activities during an IEG Mission Visit in Brazil (2002)
State Ceará Rio Grande de Norte Alagoas Goias São Paulo

Copying 6 6 2 8 (2 secondary)
40% 33% 22% 17%
Text production 1 3 3 4 2
7% 17% 33% 28% 4%
Reading lesson in class 2 1 1 1
13% 6% 11% 2%
No activity 1 5 (2 waiting for bell) 1 8
7% 28% 11% 17% (some teachers
absent)
Teacher busy with 2-3 students 1 1 4
only 7% 11% 8%
Teacher teaching whole class 2 1 6 10
13% 11% 43% 22%
Group practice 1 2 3 4
7% 11% 21% 8%
Art and Play 1(scheduled activity) 1 (play with letters) 1 (play with letters) 1 (drawing) 8 (during subject
7% 6% 11% 7% matter)
17%

No. classes 15 18 9 14 45

No schools 8 6 7 5 8

Apparently instructional
34% 45% 48% 92% 40%
activities (some suboptimal)
Source: OED 2002

The scarcity of school resources, such as textbooks, and the presence of unqualified
teachers (especially in rural schools) often result in reductions in classroom time. 70 A
comparison of classroom instructional time in three Latin American countries found that
Brazilian students spent significantly more time copying instructions than Chileans and
Cubans. Few Brazilian schools used prepared activities, instead requiring students to
copy math problems from the board before beginning work; test scores tended to reflect
these differences. 71 In the Gambia and Burkina Faso, a large amount of time was lost
writing lessons and problems on the board, because students lacked access to textbooks.72
However, students must know how to read and understand the textbooks in order to learn. A
study on a Kenyan program where an NGO provided textbooks to all pupils found
improved used instructional time in classrooms, but test scores did not improve. 73

Other things being equal, time on task distinguishes between students who learn a lot
about a subject and students whose learning gains are modest. Some teachers are willing

70. Attar 2001; Njie 2001

71. Carnoy, Gove, and Marshall 2004
72. Dia 2003
73. Glewwe et al. 1999
 15

to spend more time with students, and differences across classrooms may be striking,
particularly for children with low initial reading scores. For example, in OECD countries,
some first-grade teachers spent as little as 43 minutes on language arts while others
devoted over 104 minutes per day. Extrapolating to an academic year, some children
received as much as 180 more hours of instruction in language arts compared to other
children. 74 Similar results were seen in a Peruvian study, where some low-income
schools taught much more reading than others. 75 Clearly, engaging students as much as
possible on learning tasks is a hallmark of teacher quality. And engaging students in
activities that are more conducive to subsequent recall and to the formation of linkages
among pieces of information is another important feature of quality.

U.S. research in the 1980s found that corrective and immediate feedback, attention to and
transfer of prior learning, and the teacher interaction with students were important factors
affecting achievement. 76 Specifically, the following relationships were found between
time use and classroom activities: 77

• more time spent on discussion and review (r = +0.40);

• more time spent reading aloud (r = +0.59); and
• more supportive, corrective feedback from the teacher (r = +0.50).

Activities found to be negatively associated with reading gains were:

• more time spent in silent reading (r = -0.23);
• more time spent in sustained silent reading (r = -0.52);
• more time on classroom management (r = -0.24);
• more time spent by the teacher on organizing activities (r = -0.52);
• more time spent by the teacher on social interactions (r = -0.52); and
• more time spent on negative interactions (r = -0.29).

Time use in class is partly related to subject matter and need to pay attention. Yair
(2000) gave students in the US watches that beeped at random times, during which
students reported what they were doing. He found that the gap between allocated and
productive instructional time is significantly larger among minority students. Students
reported engagement in lessons only 55.4% of the time; 62% in grade 6, and 47% in
grade 12. Disengagement increased with age and absorption in other activities. Math
held their attention 63% of the time, reading 55%, lab work 74% of the time, group work
70%. Boys were engaged 21% more than girls and blacks were engaged 29% less.
Relevant and challenging materials held the attention more than boring materials.
Teacher-centered methods that demand greater 'pan-optical' control got less engagement.
Thus, research suggests that students may be more engaged in hierarchically oriented
courses (such as science) that demand prior knowledge than in courses where prior
knowledge is not needed. Thus, the preponderance of continuous lecturing that requires
little response from students may result in low rates of instructional time.

74. Kelly 2003. OECD is the Organization for Economic Cooperation and Development.
75. Crouch et al. 2005
76. Quartarola 1984
77. Stallings 1980
 16

Greek research suggests that the weaker students are off task more often than the better
students. The better students in a class were engaged in learning 76% of the time,
average students 73% of the time, and weaker students 57% of the time. Off task time
(absentmindedness, social interaction) amounted to 9.7 minutes for the better students,
10.8 minutes for the average students, and 17.4 minutes for the weaker students. 78 This
research suggests that the students who are more knowledgeable can construct knowledge
networks more efficiently.

TIME SPENT LEARNING PRESCRIBED CURRICULA (ACADEMIC LEARNING TIME)

Ideally students should not just be engaged in any learning activity, but should spend
their time in activities that teach the prescribed curriculum. However, the degree to
which this happens in lower-income countries is uncertain. Little systematic information
exists regarding the amount of time schools actually spent presenting new material and
progressing with the specified curriculum. Empirical information suggests considerable
deviation. For example, in Ghana, a large portion of rural schoolteachers did not follow
the prescribed weekly timetable. 79 This may happen because the teacher is not keeping a
good track of the time that should be spent in various topics throughout the year or does
not know certain topics well, such as math. Also teachers teaching multiple sections may
be confused and present topics that have already been covered.

Teachers in low-income areas often do not know how much time to devote to certain
topics or how to budget their time throughout the year. Teacher training and supervision
rarely focuses on the use of allocated time or on planning to ensure that curricula are
covered during the year. One means for doing so is giving teachers curricular calendars to
help them easily visualize where they should be and to prevent classes from falling
behind.

Students who learn the prescribed curriculum are most likely to score well in
achievement tests, so academic learning time may be a more useful predictor of learning
outcomes than mere engagement in learning activities. 80 However, measuring the time
spent teaching the relevant curricula may be most advantageous to middle-income
countries that already use time well in class and need feedback for fine-tuning. In lower-
income countries or areas, just increasing the instructional time devoted to the curricula
of any previous grade might improve learning outcomes.

RELATIONSHIPS BETWEEN INSTRUCTIONAL TIME, POVERTY, AND ACHIEVEMENT

Due to the issues discussed above, the schools of the poor may have less time for their
students. US studies have found socioeconomic differences in the amount of time
students spend learning to read, with classroom interruptions and disruptions a salient
problem in schools attended by pupils from low-income families. 81 One longitudinal
study found that high socioeconomic status students engaged in writing, reading and

78. Matsagouras 1987.

79. EARC 2003
80. http://education.calumet.purdue.edu/vockell/edpsybook/Edpsy2/edpsy2_intro
81. Stevens 1993
 17

academic discussion 5% more time per day than poorer students. 82 The study estimated
that students of low socioeconomic levels would need to attend school 1.5 months during
the summer break in order to attain an equivalent amount of engaged learning time.
Wastage adds up over the years and creates risks for poorer students.

Although research has been ex post facto rather than experimental, the relationship
between instructional time and student achievement has been consistently positive. 83 For
example, an extensive review of U.S. studies concluded that other things being equal, the
amount learned is generally proportional to the time spent in learning. 84 Variables
measuring curricular exposure are strong predictors of test scores 85 and correlations
between content exposure and learning are typically higher than correlations between
specific teacher behaviors and learning. 86 These positive relationships have been fairly
consistent in studies employing different analytical perspectives, measurement strategies
and units of analysis. 87 Two comprehensive reviews of research on learning effects
demonstrated the positive influence of time on learning and the increasing precision in
defining it: in an earlier review of 35 studies positive effects were found in 30 (86%) of
the cases; in a later review of more than 100 studies positive influences were seen in 88%
of the studies. 88 The association between time and learning was weakened only when
other classroom factors were controlled, such as student differences in aptitude and prior
knowledge, differences in the quality of classroom instruction and morale, or properties
related to pupils' socio-economic environments. 89

Research outside the U.S. has been sparse. More instructional time spent on general
science was associated with higher academic achievement in Iran, India and Thailand. 90
Increased pupil reading time had a positive effect on pupil achievement in Chile and
India. School-based instructional time was found to be especially significant for children
who got little school academic engagement after school hours. 91 Instructional time was
found to be one of three major areas (in addition to teacher quality and textbook
availability) in which consistent achievement effects were obtained. 92 In a review of 14
studies involving instructional time in developing countries, 12 studies showed a positive
relationship between instructional time and student achievement. Another study
examined the influence of a variety of active learning methods and found instructional

82. Greenwood 1991

83. Bloom 1971; Fisher and Berliner 1985; Gettinger 1984, 1989; Rosenshine 1979; Walberg 1988; Wang
1998; Wiley and Harnishfeger 1974
84. Frederick and Walberg 1980,
85. Content exposure was the most significant predictor of written test scores after everything else was
controlled, while quality of instructional delivery was the most significant predictor of hands-on test scores.
Further, students' attendance rate, content coverage, content exposure, and quality of instructional delivery
are significant predictors of students' achievement test scores. (Wang 1998, p. 150)
86. Rosenshine 1979
87. Huyvaert 1998
88. Walberg and Fredrick 1991
89. Walberg 1988; also Worthen et al. (1994) found that the learning effects of increased time may be
overstated.
90. Heyneman and Loxley 1983. In rural India student achievement was higher in schools with more
instructional time; schools teaching the highest number of hours reported 66 more hours per year than
schools with lower achievement (World Bank 1997).
91. See Suryadi, Green and Windman (1981) in Lockheed and Verspoor 1992.
92. Fuller and Clarke 1994
 18

time to be a consistent predictor of pupil performance. 93 In the Philippines and Ethiopia

more pupil-oriented teacher behavior led to improved use of time devoted to learning and
had a significant impact on learning processes and resulted in higher achievement
levels. 94 But studies have not captured all of the important dimensions of time loss and
gain in lower-income countries, so the magnitude and significance of the relationship
between instructional time and educational achievement may seem uncertain. 95 For
example, in Pakistan it was found that teaching time by itself was a poor predictor of
student achievement, and that the effective use of time was a more accurate predictor. 96

Reduced instructional time may be related to higher dropout rates. For example, in Egypt
girls attending multiple-shift schools with fewer instructional hours were five to six times
more likely to drop out before completing lower secondary education than girls attending
a single-shift school. 97

Figure 2-5. Time lost in book distribution in Honduras

Tutoring. To make up for
instructional time lost to strikes,
absenteeism, and lack of feedback
parents have resorted to private
tutoring. This is a major
phenomenon worldwide that has
grown dramatically in recent
times and has affected the
priorities of those who tutor.
Families with the necessary
resources are able to secure not
only greater quantities but also
better quality of private tutoring.
Source: author, rural school 2003
Children receiving such tutoring
are then able to perform better in school, and in the long run to improve their lifetime
earnings. However, children of low-income families often do not receive such benefits
and may drop out of school at an earlier age. 98 This is one more way that the better off
get more instructional time than the poorer students.

The association between instructional time and achievement lead some educators to argue
in favor additional hours and a longer school year. In higher-income countries, the length
of the year may not greatly affect performance (for example, Japan has 210 school days
and Taiwan only 180). But poor students with language delays may benefit from longer
instructional days when used well. Some countries (notably Uruguay) have “full-time”
schools that keep students for about eight hours and offer many enrichment activities.
Low-income students in Uruguayan full-time schools were 10% more likely to get

93. Armitage et al. 1986, in Fuller 1987

94. Tan, Lane and Coustère 1997; Verwimp 1999
95. Anderson 1976; Karweit 1976, 1983
96. Reimers 1993
97. Lloyd et al. 2003
98. Bray 2006
 19

passing scores in grade 3 over those attending regular schools.99 However, the effect is
modest and suggests that merely increasing the time students are in school may have
limited effects 100 without an increase in the time students are actually engaged in
learning.

Overall, the instructional time received by students in the developing world is reduced
appreciably due to various conditions and forces. While some of these factors are
episodic in nature, others are structural with long-lasting effects. 101 Other factors come
to play besides stark time use, such as the elaboration of concepts by students and the
memorability of activities undertaken during the learning sessions. Nevertheless,
cognitive research would predict that students elaborating concepts for a longer period
would remember it better than students covering more concepts for shorter periods. Thus
findings underscore the importance of effective time use in lower-income classrooms,
especially through introducing more interactive instructional methods that enable
students to “contemplate”, analyze or synthesize the subject matter. It is important to
obtain detailed data about the incidence of all these factors and focus country dialogue
maximizing the use of effective time.

99. Cerdan-Infantes and Vermeersch 2007). Full-time schools in Uruguay cost 60% more than regular
schools and offer students extracurricular activities rather than extra instruction (OED 2005a).
100. Karweit 1985
101. Millot and Lane 2002
 21

3. INSTRUCTIONAL TIME DATA COLLECTION AND

FINDINGS
A grant from the Bank-Netherlands Partnership Program for Global and Regional Initiatives
(BNPP) offered the opportunity to compare time use in countries with different
socioeconomic indicators in in 2004-2006. The first step in conducting an instructional time
study was to search for innovative assessment methods and for experts knowledgeable about
classroom observations and time use.102 A research design was developed with the help of the
UNESCO International Bureau of Education (IBE), an institution that has curricular expertise
and keeps a database of the hours countries spend in teaching various subjects. Then
countries where IEG evaluations had been recently undertaken were contacted through
World Bank task managers and were invited to participate in the study. Also some countries
were invited where a new project was being appraised in hopes that baseline data would be
collected. 103

The countries that showed an interest in carrying out studies were Tunisia, Morocco, Ghana,
and the Brazilian state of Pernambuco. To build capacity, and test the feasibility of the
methods, the surveys were not administered by private companies or consultants. Instead,
Ministries of Education in each country agreed to train and use local government staff and
carry out interviews and observations on a representative sample of schools. The Tunisian
Centre National de l’Instruction et des Recherches Educatives (CNIPRE), and the Moroccan
Direction de l’Evaluation, de l’Organisation de la Vie Scolaire et des Formations
Interacadémiques extensively collaborated with the Bank (See Annex A for methodological
details).

Methodology. Efforts were made to develop a measurement method that would cost
relatively little, be administered fairly quickly, and capture time loss from multiple sources.
Studies on teacher absenteeism have relied on multiple visits to samples of schools. To keep
costs low and implementation simple, the study made a single unannounced visit to the
sampled schools and compared information on time loss parameters from multiple sources:
(a) Recall and self-reports by principals, teachers, students, and parents, (b) school-level
observations, and (c) classroom measurements of time use. Surveys were developed and
piloted to obtain self-reports and school-level observations. Classroom-level observations
were obtained through a quantitative instrument that registered the amount of time spent in
various common classroom activities. It also recorded the percentage of students who
appeared not to be paying attention to the instruction (off-task rate). This measure estimates

102. The team sought the advice of Jane Stallings (retired dean of education at Texas A&M and a pioneer
in instructional time measurement), Stephanie Knight, Eugene Schaffer (professors and Texas A&M and
the University of Maryland in Baltimore county respectively), David Markham (retired systems engineer),
the members of an International System for Teacher Observation (ISTOF), and the World Bank
Development Economics Department (DEC) that has carried out teacher absenteeism studies.
103. The Independent Evaluation Group conducts qualitative evaluations on a sample of completed projects
that are called Project Performance Assessment Reviews (PPARs). Initial plans included Guinée and
Cambodia. Staff from Cambodia received training in instructional time assessment for use at a later time.
 22

the number of person-minutes wasted in a class (e.g. 50 students unengaged for 20

minutes).104

Observed time loss was projected through the entire school year, taking school days and
holidays into account. Repeated observations in a random sub-sample of schools of
Pernambuco about 1.5 year after the first observation showed that overall time loss
parameters were stable. (See procedures and analysis details in Annex A.) Overall,
information from the different informants in schools was consistent, and time loss estimates
could be averaged during analysis.

HOW MUCH OF THE INTENDED TIME WAS SPENT ON LEARNING ACTIVITIES?

Instructional time losses varied by country. In Tunisia, the country with the most efficient use
of time, students were engaged in learning 77.9% of the available time. In Ghana, however,
students were engaged in learning only 38.7% of the time, and in Pernambuco 63% of the
time (Table 3-11, Figure 3-1). Morocco had indicators somewhat lower than those of Tunisia
(71.1% use of time for learning), but school closures and teacher delays were apparently
underreported. Translated into number of days effectively available for learning, losses are
palpable. For example, only 76.3 days were devoted to learning tasks of the 197 officially
available to Ghanaian students, while for Tunisian students 148.1 of the 190 days officially
available were devoted to learning tasks. In effect, Tunisian students get twice as much of
the intended classroom time as Ghanaian students.

Figure 3-1. Percentage of time lost at the country level

100%
87.4% 89.9%
90% 86.1%
77.9%
80% 71.1%
70% 63%
60% 55.1%
50%
38.7%
40%
30%
20%
10%
0%

Pernambuco Ghana Morocco Tunisia

(Brazil)

% year schools and teachers available % time spent in learning tasks

Time loss varied among schools within countries. In Tunisia, Morocco, and Pernambuco,
most of the schools sampled seem to have experienced relatively few school closures, most
teachers seem to arrive at school on time, and they are rarely if ever absent or late. Reports of
early departures were rare. However a few schools reported high closures for various reasons
(climate, structural issues, or teacher strikes) and skewed the averages, while some teachers

104. This has been shown to be related to learning outcomes and has been sometimes used as a proxy
variable in studies (e.g. Stallings et al. 1979).
 23

had inordinately high rates of absenteeism. For example, about 85% of Tunisian teachers
report few if any absences, but about 15% have missed much of the school year (reportedly
due to illnesses and pregnancies.) Data from Pernambuco showed different absence rates
according to the day of the week. There was 55% more absenteeism on Monday (4.79% of
teachers) than on Thursday (2.69% of teachers). Also afternoon and evening shifts showed
greater losses than the morning shift.

Table 3-1. Instructional Time Use in the Four Countries of the Study
Pernambuco Ghana Morocco Tunisia
(Brazil)
School Year (Days) 200 197 204 190
School Closures (Days) 4.79 3.17 1.38 5.15

Days after Closures 195.21 193.83 202.62 184.85

Teacher Absence (Days) 12.76 43.01 13.36 11.55

Teacher Delays 5.50 39.75 6.94 1.27
Early Class Dismissals 2.31 2.43 6.68 1.22

No. days schools operated 174.65 108.6 175.6 170.8

% Year available for teaching 87.3% 55.1% 86.1% 89.9%

Engagement Rate in Interactive

or Passive Classroom Tasks 72.1% 70.2% 82.6% 86.7%

School Days Devoted to Learning 125.9 76.3 145.1 148.1

School Year % Spent Engaged
in Learning Tasks 63.0% 38.7% 71.1% 77.9%

Student off Task Rate 19.3% 21.1% 9.2% 9.9%

Instructional Efficiency Given Off-Task Rate 50.8% 30.5% 64.6% 70.2%

Student Absence (Days) 7.82 9.04 4.30 3.35

Student Delays (Number of Times) 5.64 10.61 5.187 2.63

How satisfactory are these results? Comparable indicators for this level of detail are
almost nonexistent. US indicators ranged from 39% to 96% in the 1980s and to 85% in
2003 (Chapter 2). However, they mainly measure time use in the classroom and are not
really comparable. As more data are collected in other countries, the range of the time
loss parameters will be understood better.

HOW OFTEN WERE TEACHERS ABSENT OR LATE?

Teacher delays and absences (legitimate or otherwise) resulted in about 11.5-43 days lost
during the school year. The number of days schools were closed for various reasons
during the school year was reported as low, from 1.4 day in Morocco to 4.8 days in
Pernambuco. But no details were asked about the timing of closures or about partial
cancellations of classes, so the figures may be underestimates.
 24

Teacher delays. It is not sufficient to show up at school, teachers must also teach. About
15.4% of Moroccan teachers were observed to delay entering their class by about 8
minutes. If this delay is consistent across classes, it constitutes 2.4% of the time broadly
lost or 4.91 days of the 204 in the Moroccan school calendar. In Pernambuco, observers
also found 12% of the teachers (5% in the 40-school sample) outside the classes during
class time (half were chatting). Tunisia and Ghana did not explore whether teachers were
inside classrooms.

Present but not teaching... The Moroccan enumerators observed that 15.4% of teachers
were late by 7.8 minutes in entering the class. If this delay is consistent across classes, it
constitutes 2.4% of the time broadly lost or 4.91 days of the 204. In Pernambuco,
observers of the Pernambuco survey found 12% of the teachers (5% in the 40-school
sample) outside the classes during class time (half were chatting). Other countries did not
ask this question, which has generated in other countries concerns about teaching
avoidance. However, a 2003 USAID study in Ghana found that on the day of school
visit, 25% of the teachers were present but not teaching. After missing school for about
20% of the time, 9% of the time left is lost to student abandonment. 105 The study
estimated 8.25 hours lost per week in the mainly rural and periurban communities of the
sample.

Assigned to a school but not teaching…. In several countries teachers may be assigned to
schools where they do not teach, often as a political favor (Chapter 2). This survey did
not ask the schools how many teachers were assigned to them vs. how many teachers
were actually expected to show up every day. However, the Superintendency of
Pernambuco found at the time of the survey large numbers of teachers with fewer
assigned hours of classes. There were also teachers who were on various types of leave,
others who had been seconded to schools of their liking, and some who had effectively
abandoned the public schools. For example, the district of Barrentos showed that 175 of
the 459 teachers (38%) missing, while Sertão de San Francisco had 40% of the teachers
missing. In addition some “substitutes” were found who had been illegally contracted by
teachers to teach in their stead. Principals were often cognizant of the problem and
assented. In addition to the salary wastage, training becomes impossible since it is
unclear who teaches in various schools. Teacher unions complain when actions are taken
against absent teachers, and conflicts are frequent.

Poor record-keeping and potential under-scheduling. A detailed examination was

available only from a sample of 40 schools in Pernambuco. Enumerators found absences
in 88% of the school registries. They also found that 74% of the teachers’ log books
(livros de ponto) were either missing or were not filled out (Figure 3-2). One reason is
the limited interest in these log books. Teachers reported that in 62% the registries are
reviewed by the principal and 32% just by the secretary. Principals interviewed
mentioned that they are not really responsible for teacher absenteeism, they just report it.

In the 26% of the registries that were filled, the number of student hours that should be
taught was lower than expected, particularly in the Portuguese language classes where
more than half the times the amount of time students should have studied was missing; of

105. EARC 2004.

 25

23.1 hours that teachers ought to be teaching in grades 1-6 in Pernambuco, they were
registered to teach only 17 hours. It was not possible to ascertain whether empty registry
entries signified failure to document or failure to teach. But some of these teachers may
have been scheduled for fewer hours of instruction, as found out by the Pernambuco
Superintendency office. This tendency is a pattern also found in other countries, such as
Indonesia. 106

Figure 3-2. Expected and registered number of hours

taught in Pernambuco

Hours Montly hours by grade

35
30
25
20 Expected
15 Registered
10
5
0 Grade
1 2 3 4 5 6 7 8

Ghana did not provide data on the use of class and school registers. However, the
USAID study in 2003 reported that 97-99% of the classes had registers and several were
using them effectively.

Roll call. In all classes, a significant amount (maybe up to 10 minutes) can be taken up by
calling the names of students to check presence. If this is done every day, then up to 8%
of class time (about 16 days per year) may be spent in this task.

WAS THERE RECUPERATION OF CANCELED CLASSES?

When asked how often remediation takes place or what activities are done when teachers
are absent, responses are contradictory. Teachers in Pernambuco mentioned that in 80%
of the days there are substitute activities, but only 60% of the principals mentioned this
option. The principal or another teacher may take a class, but substitute teachers in
Pernambuco seem to be used only in about 17% of cases. Fewer than 50% of classes lost
are recuperated, and the number is uncertain.

In Tunisia findings were similar. Although principals stated that they usually have
substitute teachers (Table A-12), teachers stated that students generally go home (Table
A-13).

106. also see Millot and Lane 2002

 26

Table 3-2. Solutions adopted by Tunisian and

Moroccan principals during teacher absence
Tunisia Morocco
Use a substitute teacher 74.5% 58%

Send the children home 52.9% Not asked

Distribute them to other classrooms 29.4% 72%

Keep them in their classroom 21.6% Not asked

Principal teaches the class 17.6% Not asked

Other arrangement 3%

WHAT POLICIES WOULD INCREASE TEACHER ATTENDANCE AND PUNCTUALITY?

The reasons for teacher absences, self-reported or reported by principals typically

involved illness, pregnancies, family events, and training. There was little difference
between men and women in this respect. Ghanaian teachers often reported distance and
the need to collect salaries from town as a reason for missing classes.

Teachers were asked for suggestions to decrease absenteeism and delays. They proposed
salary increases and government compliance with its own rules about absenteeism. For
example, Brazilian principals and teachers suggested that absenteeism would be reduced
through: (a) Salary increases (over 70% of teachers), (b) sanctions or punishment of
absent or late teachers, (c) assigning teachers to schools near their homes and (d) help for
new teachers. Principals thought that better communication with district offices (60%)
might help.

In Ghana, existing measures already include salary deduction of the hours absent, regular
and effective monitoring by head teachers and circuit supervisors, district disciplinary
committee in the enforcement of teachers’ code of conduct. However, these sanctions
tend not to be used in Ghana or elsewhere. The reasons in each country are unclear, but
the political power of teacher unions often ensures impunity for all but the most serious
violations.

HOW OFTEN WERE STUDENTS ABSENT?

Projected through the year, students missed relatively few days: from 3.35 days in
Tunisia to 9.81 in Ghana and reported being late from 5.46 days of classes in Brazil to
10.7 in Ghana. These percentages are roughly similar to those reported by teachers and
principals. In Ghana principals reported student attendance rates as 85% and teachers as
90%, while in Brazil, 98% of boys and 99% of girls were reported as attending regularly.
There is a tendency for student absences to follow teacher absences, and the lower
attendance in Ghana may be related to the high teacher absenteeism in that country.
Students missed classes mainly due to personal or parental illness, family obligations.
Reasons tend to be similar across countries, with the exception of Tunisia, where some
students refused to go to school if they had not completed their homework, and Ghana
where some students did not attend because they could not pay tuition.
 27

In the countries where the research took place, gender was not an important determinant
of absenteeism, so responses usually did not differ by gender. The survey did not explore
whether students skipped classes and left early, as was often found to be in a 2004 study
of rural schools in Bangladesh (Chapter 2).

It would be desirable to estimate the amount of learning time available after accounting
for student absenteeism, but the latter is not a subset of teacher absenteeism. Students
may go to school and find teachers absent, or they may be absent on days that teachers
are present. Observations also suggest that two types of students exist: those who are up
to date with the classroom material and attend regularly and those who know too little to
keep up and attend sporadically. Without more detailed data, the amount of total learning
time given student given absenteeism could not be calculated.

HOW MUCH OF THE CLASSROOM TIME WAS USED FOR LEARNING ACTIVITIES?

The Stallings Classroom Snapshot, a well-known U.S. instrument, was adapted for
international use. The system helps record and classify activities as interactive
(instruction, discussion, questions), passive (copying and seatwork), and management
(various activities peripheral to instruction). It also records a rough percentage of
students who appear to be off-task (Table 3-3). Classrooms engaging students
interactively are more efficient and may be more likely to foster memory consolidation
than classrooms where students spend more time on non-instructional activities or are
off-task.

Table 3-3. Percentage of Time Spent in Instructional Tasks

Instructional Strategy U.S Classroom Pernabuco Ghana Morocco Tunisia
% of time Criteria (Brazil)
Interactive instruction 50% or more 52.4 59.9 62.8 61.2
Oral Reading 6.7 8.7 15.7 15.3
Teaching, Explanation 32.8 19.9 26.7 27.9
Discussion 6.3 24.1 6.6 6.2
Practice Drill 1.4 6.5 12.3 11.3
Passive instruction 35% or less 19.6 10.3 19.9 25.6
Seatwork 16.3 7.4 14.8 22.9
Copy 3.0 2.9 5.0 2.7
Total Instructional Time 72.1 70.2 82.6 86.7
Organizing/Management 15% or less 27.9 28.0 17.8 13.3
Student Off Task Rate 6% or less 19.3 21.1 9.2 9.9

Observations showed that the most important block of time lost tended to be inside the
classroom. Ghanaian teachers spent 70.2% of the time engaging students in learning, while
Tunisian teachers spent 86.7% of the time. Overall the largest chunk of time is spent on
instruction, either lecturing or explanations. This ranged from nearly 33% in Brazil to 20%
in Ghana.

Teachers in all four countries addressed the entire classroom 70-90% of the time (Figure 3-
4). They mainly just talked and used the blackboard (“chalk and talk”), while students wrote
 28

on notebooks. This activity delivers material efficiently from the teacher’s viewpoint, but if it
prolonged it can result in limited student attention and subsequent recall. (In fact, lecturing
can be considered active engagement for teachers and passive for students.) One might
expect limited interactions to happen mainly in large classes, but it happened everywhere.
Average class sizes were reasonable, (ranging from 25 students in Pernambuco to 31 in
Morocco), but class size was correlated only to a small extent with the likelihood of a teacher
addressing the entire class (r=0.13), with the likelihood of addressing a single student (r=-
0.12), and the tendency to lecture (r=-0.28). Predictably, the time allotted to discussion and
oral practice drill tended to be limited (Table 3-3), although students could be repeating in
unison during some of these activities.

Figure 3-3. Interactive teaching: brief lecture, practice, discussion

Lecture Practice drill, discussion

Source: Barnette and Benavot 2005. Source: Barnette and Benavot 2005

Less effective learning delivered. Cognitive research suggests that students are more likely to
retain material that they have had the chance to contemplate and to encode in multiple
cognitive networks (visual, auditory, etc.).107 Though engagement rates seem high, the
activities observed overall gave few opportunities to contemplate the subject matter. It was
unknown whether appropriate teaching aids were available in the schools surveyed, but
manipulables (such as geometry tools) and audiovisual equipment were used only about 3%
of the time. Textbook use ranged from 23% of the time in Pernambuco to 7% of the time in
Ghana, where few textbooks exist. Little time was spent reading aloud in Pernambuco and
Ghana (6.7-8.7% of the time) but much more was spent in Morocco and Tunisia (about 15%
of the time), where koranic reading is also taught (Figure 3-5). In all four countries, less than
5% of the time was spent in group activities, that might help contemplate the concepts taught.

107. See Abadzi 2006, chapters 4 and 8

 29

Figure 3-4. Whom were teachers addressing Figure 3-5, Use of Instructional Materials by
in class? Country
100 45

90 40

80 35

70 30

60 25

%
20
50
15
40
10
30
5
20
0
10

ng
ng

es

rs
d

l
s

ua
oo
0

ok

to

ni
oa

iv
hi

is
eb
Bo

la
at

ar
ot

kb

ov

cu
ul

Le
ot
N
Pernambuco Morocco Ghana Tunisia

ac

di
ip
N

al
Au

e
an
Bl

iv
M

rat
pe
oo
1 student small group everyone

C
Pernambuco Morocco Ghana Tunisia

The non-instructional time was taken up with organization/management activities, some of

which are necessary, but also included socialization and the teacher being out of the room. In
Ghana and Pernambuco, 28% of classroom time was taken up by such activities. Possibly
due to poorly organized learning activities, 19.1% students in Ghana and 21.3% in
Pernambuco were on average “off task,” uninvolved with class activities. By comparison,
only about 9.9% of Tunisian students and 9.2% of Moroccan students were “off task” during
class (Table 3-4). This observational inference suggests that students might not be learning
when distracted. As a result, Tunisia and Morocco offered not only more class time in
learning tasks, they also engaged a higher percentage of students in learning than Brazil and
Ghana. It can be argued that the former countries teach their students more efficiently than
the latter.

Grade-wise changes. The frequency of various classroom activities predictably changes

with age. Teachers of lower grades spent more time on organization, and younger children
had higher off-task rates, given a briefer attention span. With the passage of years, reading
aloud and drill tend to diminish while lecturing and discussion increases. For example, in
Morocco reading activities took up about 17% of the time in grades 1 and 4, but only 2.5%
of the time in grade 5 (and 15.1% in grade 6, as students perhaps prepared for
examinations). In Brazil 8th graders spend proportionately more time copying because they
do not have textbooks at the secondary level. Tunisian 6th graders and classes with a larger
percentage of male students were slightly more likely to be off task and less likely to be
engaged in interactive instruction. Gradewise time on task indicators can be compared in
grade 4, which was measured for all countries (Table 3-5). 108

108. A large-scale US study found that fifth graders spent most of their time (91.2%) working in whole-
group or individual-seatwork settings and only 7% of the time in small-group instruction (two to five
students). In the same grade, 37% of instruction was in literacy and 25% was in math, while in first and
third grades, more than 50% of instruction was in literacy and less than 10% was in math. Science and
social studies activities occurred in 11 and 13% of intervals in fifth grade, respectively. In the fifth grade
spent 17%of their time instructing students on managing materials or time (Pianta et al. 2007)
 30

Table 3-4. Average percentages of Time Use in 4th Grade (common for all countries)

Student off Interactive Organization-

Discipline Task learning Passive learning management
Country % % % % %
Pernambuco 1.75 18.21 52.89 18.61 28.50
Ghana 1.44 25.92 52.50 12.50 35.00
Morocco 4.20 9.37 62.85 20.03 17.78
Tunisia 0.88 9.64 61.70 26.32 11.98

Parents may not always be aware that time in their children’s classrooms could be used
better. Most parents in Pernambuco (77%) do not complain about absences, and 87% of
them expressed satisfaction with students’ teaching and learning processes.

Engaging students in learning relevant to the curriculum. The study measured

engagement in any kind of learning activity that the teacher organized. It was not
possible to measure time spent on the last parameter of the instructional time model,
which is time spent just on relevant curricula. To do so, more complex instruments
would be needed, such as student artifact studies, video studies, observations by
supervisors knowledgeable in each subject, or interviews with evidence regarding how
far students are behind given curricular specifications. Such a study would have taken
more time and funding than was available. However, informal observations in lower-
income schools suggest that some learning activities were appropriate for lower grades
but had not been mastered earlier.

HOW MUCH HOMEWORK DID STUDENTS COMPLETE?

Homework increases instructional time, and all countries asked survey questions about it.
Parents, students, and teachers generally agreed about homework frequency and effort.
Teachers estimated that they gave 1 to 3 hours of homework, while most students
reported spending at least 1 hour or more on their homework (Table 3-6). Nevertheless,
in many areas individual students lack textbooks, and it is uncertain how much
homework can be done under these circumstances and what its value is. Therefore this
variable was not added in the instructional time model.

Table 3-5. Homework Efforts Reported by Students and Parents

Homework Effort Pernambuco Ghana1 Morocco Tunisia
% % % %
2
Respondents Parents Students Parents Students Parents Students Parents Students
Less Than One Hour 24 27 84 54 54 58 31 54
1 to 2 Hours 56 44 14 6 35 16 40 15
2 to 3 Hours 14 13 4 4 12 3
3 to 4 Hours 3 8 1 1 7 1
4 to 5 Hours 0 5 1 1
More than 5 Hours 2 2
1
Ghana’s questionnaire did not permit responses beyond more than one hour of homework.
 31

TRENDS IN LINKAGES BETWEEN LEARNING OUTCOMES AND POVERTY STATUS

This study could obtain only preliminary evidence on the linkages between time loss and
student performance. Achievement test data were available only in Pernambuco
(Education Evaluation System of the State of Pemambuco-SAEPE). Due to various data
and computational issues, measurement precision was modest. Also, for logistics
reasons, multiple regressions were computed rather than hierarchical linear modeling, a
process that may bias some standard errors and effect sizes. 109 Nevertheless, low but
significant relationships were found between achievement and various time indicators.
After controlling for socioeconomic indicators, regressions showed that significant
amounts of test score variance were predicted by: (a) the number of days Brazilian
students missed, (b) the number of teachers absent on the day of the survey, (c) amount of
passive instruction (often related to lower test scores), and (d) management activities.
Approaching statistical significance (p<0.07) was the number of times a teacher left early
the previous month.

Area-level socioeconomic data were available only for Pernambuco and Morocco. In
both countries time parameters showed low but statistically significant relationships
between instructional time use and socioeconomic indicators, such as area life
expectancy, income level, and literacy rate. Particularly in Morocco, poverty rate and
rural school locations were positively correlated with school closures, interactive and
passive instruction, time spent in management, and off task behaviors. In general,
interactive learning time was negatively correlated with poverty and urban location.
Though this finding is preliminary, it is consistent with earlier research in the US which
showed that in poor areas students have less instructional time (Chapter 2).

METHODOLOGICAL LESSONS AND DATA LIMITATIONS

The study mainly sought to develop a feasible methodology for measuring instructional
time use across countries. It was found that it is possible for governments to do so in a
reasonable time period, and with a reasonable cost. Three of the four countries were in
the middle-income range, but the Ghana Ministry of Education also successfully carried
out the survey and processed the data.

However, instructional time concepts are not widely known. Thus, the study faced
several methodological challenges that included lengthy implementation delays, an
insufficiently detailed assessment of school closures, and surveys modified by different
countries beyond comparability. One consequence of collaborating with Ministries of
Education rather than hiring consultants was that each country administered slightly
different versions of the questionnaires. Some government staff did not sufficiently
understand the purpose of various survey questions and the importance of obtaining clear
numerical answers that would help estimate the number of days lost. The data collected
by the study were treated statistically to extract similarities and equate to the extent
possible the time losses reported in each country in order to compare them. As a result of
these problems, the datasets included measurement error that reduced the potential
strength of relationships among various variables (Annex A).

109. Osborne 2000

 32

This study was only the first phase of this research topic and it concentrated on assessing
the number instructional hours of schools. It did not deal with the amount of time and
number of breaks, lunchtime, or other issues involving the distribution of time in a school
day. Also it did not delve into issues of ‘classroom climate’ and emotional support, which
are important but not easily measured. Given the limited timeframe and choice of
countries it was not possible to get more achievement test score data. In future work,
efforts should be made to get national, international, or regional test scores for schools
(e.g. TIMSS, PIRLS, SACMEQ, PASEC).

Based on the results of this study, the instructional time instruments have been revised. It
is likely that greater measurement precision will show greater time loss, although
measuring ever smaller segments of time loss may reach a point of diminishing returns.
Nevertheless, the lessons learned thus far permit the development of some hypotheses
that can be tested in subsequent research. These are presented in the next chapter.
 33

4. IMPROVING INSTRUCTIONAL TIME USE: POLICY

IMPLICATIONS
Classroom activity represents the moment when the education financing gets converted
into knowledge. For decades it has been uncertain which variables are most important for
this process. Classroom events have often been considered a ‘black box’, involving
details into which donors and some governments have traditionally hesitated to delve.
Without this information, the donor community finances educational inputs (salaries,
buildings, materials) without a clear hypothesis of how these will lead to outcomes. 110
Not surprisingly, learning outcomes are often limited (Figure 3-1). This study offers a
glimpse of “black box” variables that have been outside the radar screen of the education
sector. For poorer areas and countries, these “details” are too important to overlook.

Figure 4-1. Classroom instruction as a black box mediating input and outcome variables
Project inputs Impact evidence

Buildings Enrollments
Textbooks Dropout
Salaries Repetition
Training Learning outcomes
School grants
Other inputs

Instructional time is the mediator variable that has escaped scrutiny and measurement
thus far. Learning outcomes should not be expected without sufficient teaching and
practice to master the instructional objectives of specific subjects. In very gross terms, if
schools use 39% of the available time, they may teach only 39% of the curricular objectives.
Since knowledge is cumulative, performance losses should exceed 39% over time. Curricula
worldwide are overloaded, so even modest time wastage may result in significant student
losses. To make up for absenteeism, teachers may just lecture in a hurry rather than and
analyze the content and use the teaching aids provided to schools, or they may omit parts
of the curricula. So, the average Ghanaian public-school students cannot master the
curricula just by going to school; they need private tutoring to pass exams.

Discussions of educational expansion in low-income countries often bring up a tradeoff

between quantity and quality. But aside from the instances where instructional time is
cut purposefully to fit more students in multiple shifts, the tradeoff may not exist. In
effect the issue is quantity of teaching, sufficient to master the prescribed curriculum.
Good time use is not sufficient to produce quality teaching, but it is necessary. For
greater efficiency, activities must be used that help students form linkages among pieces
of information and retain the information for the long term. 111 The real tradeoff may be
the lack of supervision and interest in whether teachers attend to low-income students.

110. IEG 2006b.

111. High-quality teaching also means offering immediate and corrective feedback and creating “useful
difficulties” that have may enhance memory, such as challenging students to invent processes for dealing
with problems (e.g. Bjork and Bjork 1992).
 34

The current study has shown that even in the better-off countries instructional time may
be wasted or used suboptimally. This loss may account for some puzzling performance
problems. For example, Argentina has made much investment on teacher training and
instructional materials, but no change is evident in test scores. 112 Yet the sophisticated
Argentine school system largely leaves curricular decisions up to the teachers, relies on
copying, uses time to distribute and collect textbooks several times a day, and thus spends
much expensive time in non-instructional “management” Giving students textbooks and
reducing these “passive” activities in favor of interactive time and feedback might
increase student achievement.

The more this topic is understood and discussed, the likelier it is that some measures can
be taken. Some options and considerations derived from the findings and from the
literature review (Chapter 2) are mentioned below. These would be hypotheses to test in
subsequent research.

ECONOMIC IMPLICATIONS OF INSTRUCTIONAL TIME USE

Even as a methodological pilot, the study confirmed that under some circumstances the time
and opportunity to learn are very limited. Time is a mediator variable that has escaped
scrutiny and measurement thus far. It is not enough to provide the ingredients of instruction
and assume that they will be used in class. Students must get sufficient time to master the
instructional objectives intended in specific subjects. Furthermore, inputs like teaching aids
must be employed within the timeframe available to students, or they may not promote
student learning. The time devoted to learning the material prescribed by the curriculum may
be the crux of educational “quality”. The quantity-quality tradeoff that is often mentioned in
relationship to rapid expansion of education in low-income countries may be mitigated if
measures are taken to give students the instructional time they need, even as class sizes
increase.

The study did not include an economic analysis, but the findings suggest that instructional
time loss has significant economic implications. Government revenues pay for teachers’
salaries, buildings, teacher training, and materials, and it is expected that 100% of this
investment will be used for student learning. In fact, an hour of class in a particular
school is a budgetary fraction corresponding to the amount of time schools officially
operate (about 180 days, 4-5 hours per day at the primary level). It is possible to cost
time wastage down to the minute. Probably no schools use 100% of time well, but losses
of the magnitude shown in this study suggest that schooling costs more than it ought to or
achieves less for what it costs. Some of the implications are:

Misleading education expenditures. Education expenditures, measured as a share of GDP

or of total public expenditures, are used to estimate financing needs for specific countries,
and economists often find that expenditures must increase. However, increasing public
expenditures for any level of education when time use is poor increases inefficiency and
wastage. If a country doubles its education budget but still continues to use only half of
the available instructional time, the extra financing will have a much smaller impact.

112. IEG 2006b.

 35

Time use should be estimated and taken into consideration before such decisions are
made.

Invalid cross-country comparisons. Education expenditures are often compared among

various countries and regions. Some countries may devote a smaller percentage of their
GDP to education but use time better and thus use funds more efficiently than others
which spend more and waste more. Comparisons may be misleading if the percentage of
instructional time used in various countries varies dramatically.

Underestimated teacher salaries. If wages are calculated based on the number of hours
staff really work, teaching in some countries may really be a part-time job, whose
earnings are considerably higher than formally calculated. Low salaries may force some
teachers to work elsewhere and drive the more efficiently working people out of the
profession. But before efforts are made to increase salaries, it is important to show how
much teachers are paid for the work they do and the likely financial loss that will result if
their salaries are increased without improved instructional time.

Distorted rates of return. Rates of return are calculated with the implicit expectation that
students will be taught and will actually learn basic skills. Projections such as the amount
of marginal earnings of an additional year of schooling may be unrealistic if students are
served for only half the year or if students are illiterate and cannot benefit from the
instruction. In particular, assumptions about the pro-poor poverty alleviation effect of
education may be unrealistic, given that the schools of the poor make less effective use of
instructional time.

Hidden social inequity. Uneven time loss could also affect the benefit-incidence analyses
and Lorenz curves. Primary education is generally considered to be more pro-poor than
other levels of education, so it has more low-income schools. If, however, time is used
less well in the schools of the poor, the equity effect is reduced. More public investment
at that level will not mitigate poverty, unless it increases instruction.

Confounded internal efficiency indicators. Data linking internal efficiency measures to

time loss are sparse, but unless students have access to private tutoring, learning
outcomes are affected by the amount of instruction students get. Other things being
equal, repetition and dropout are likely to be higher in a country which uses about 40% of
instructional time for learning (e.g. Ghana) than in another which uses about 80% (e.g.
Tunisia). Furthermore, when these indicators change over time, the reasons are unclear.
Dropout could be due to external factors (e.g. being AIDS orphans or getting better job
opportunities) or to limited instruction.

Underestimated unit costs. When time is wasted, governments assume that students get
services that are not in fact provided, so unit costs per student would be distorted.
However, unit cost per successful graduate would more accurately reflect the real cost of
providing services to students.

An important task is to inform governments of this little-understood issue and engage

them in policy dialogue to modify the policies that affect instructional time. For example,
failure of schools to start on an appointed day is often due to late teacher assignments,
 36

ongoing registration, and school repair issues. In some countries (though not in the ones
included in the study) rural teachers must travel to nearby towns to get paid, closing the
schools. Local-level flexibility regarding the dates schools start and end is desirable but
may result in abuses, since monitoring becomes difficult. In countries where these issues
are a problem, policy discussions must focus on them and measures must be taken to
eliminate the root problems.

There is some evidence across countries that teacher absenteeism is related to

accessibility. Investments in rural roads and transportation might be studied for an
impact on improved teacher attendance. These may be important in helping teachers who
live at some distance come to the school every day rather than take off Mondays, Fridays,
and the holidays. Similarly evidence might be sought regarding the effects of improved
working conditions. More attractive schools have been linked to the likelihood that
students and teachers will stay on the premises. 113 The costs and benefits may be
examined of building more attractive schools at a slightly higher cost or of repairing
schools to improve appearance and functionality.

TEACHER ATTENDANCE AND ACCOUNTABILITY

Wastage of instructional time to some extent may be “normal.” It may be due to the
makeup of the human reward system of the brain that is set up to function on a short-term
frame, and the rewards of the teaching profession for teachers are not immediate (if they
can be obtained at all). 114 Unless teachers actually enjoy interacting with children and are
reinforced by small improvements in day-to-day learning, a bit of daily time loss seems
inconsequential. Children usually prefer not to study, anyway. The stresses of large
classes, unresponsive students, and insufficient knowledge to deliver demanding
curricula may make teachers avoid the tasks that they consider tedious. The solitary
nature of the job makes it possible for teachers to get away with little work, particularly
when school management is ineffective and the parents are poor. 115

Poor service provision to low-income populations is a problem in education, but similar

trends are found in health studies (Table 2-2). 116 What incentives have been more
effective in getting teachers to show up? Effective options thus far have been limited.

Educators consulted during the research overwhelmingly suggested salary increases.

However, these increases are not usually linked to improved attendance. 117 Nevertheless,
teachers in some countries are so poorly paid that few can afford to live just from
teaching. In some countries it is appropriate to question whether there should be many
poorly paid teachers who work part-time or fewer and better paid teachers working full-

113. OED 2004.

114. For a detailed explanation of the neuropsychological research available see Abadzi 2006, p. 120.
115. Bridges and Hallinan 1978.
116. In addition to high absenteeism, health care providers often withheld available care from the poor, did
not follow protocols, or did not give the poor the best advice. (Das and Hammer, 2004, 2006; Leonard et al.
2005)
117. Chaudhury et al. 2004a, 2004b, 2004c
 37

time. Efforts might be made to create wage incentives in poorer areas, though incentives
are difficult to manage, and their impact is uncertain. 118

Study participants also suggested enforcement of rules that already exist. However, many
governments are unwilling or unable to control teachers and unions. In some countries,
teachers, rather than parents of poor students, may be the governments’ clients. They are
a large electoral block that is capable of much mobilization, long strikes, and also
allegiance to favorable governments. Unions may thus be able to hold frequent and
prolonged strikes with impunity and with pay. Also, unions may secure important
concessions that include 1-2 days of personal leave per month and thus making
absenteeism legal. Medical leave may be abused by teachers, who may abandon classes
for months at a time despite a lack of substitute teachers. However, few countries dare to
fire non-compliant teachers or even hold them accountable. To deal with teachers’
interaction stress and the low pay that often comes with large numbers of employees,
teachers may be programmed for fewer hours than expected to teach by law (only 15-18
hours in many Indonesian schools and in the Middle East; Chapter 2). 119

It is important to communicate effectively to teacher unions the specific effects of time

loss, possibly through public awareness and pressure. There is also a need to get more
buy-in from teachers on time use. Many may not be well aware of how their habits add
up to disadvantage the poor, who cannot afford private coaching classes. The World
Bank has made various attempts at policy dialogue, particularly in the LAC region. The
donor community can prioritize dialogue on this and on related issues shown below.

EFFECTIVE SUPERVISION - STRENGTHENING THE SUPERVISORY CHAIN

Sometimes a low-level equilibrium of instructional time seems to exist that meets

minimum requirements for employment and community complacency. This level may
become the norm in a certain locality, and families may be unaware of the discrepancy
between intended and actual instructional time. Education officials may also implicitly
accept the gap. 120 Lack of accountability means that there is no pressure to change this
equilibrium.

Typically teachers have the obligation just to deliver the lessons, rather than ensure their
mastery, and mere delivery can be done in relatively little time. Nominally, school
administrators should be doing quality control. However, in areas such as Pernambuco,
principals stated that they were merely responsible for recording absenteeism, not for
doing something about it. If administrators do not show interest into how time is used
and do not urge teachers to use it well and follow up with them, teachers receive a clear
signal that attendance is left up to their conscience.

The lack of interest from superior to subaltern suggests a breakdown in the supervisory
chain from the central, state or municipal authorities down to schools. Governments
should take actions to strengthen this chain, providing training and feedback to teachers
and principals on use of time, and empowering communities to monitor teacher activities

118. Vegas 2006

119. IEG 2006a.
120. EARC 2003
 38

(as has been the case in Honduras and El Salvador). 121 If superiors demand that time be
used better and that teachers be evaluated based on time use, their subalterns are more
likely to ensure that this happens. This means that governments must be convinced of the
need to take action at the district and local levels, and the public must understand this
issue better.

Increasing the frequency of inspections, particularly in rural schools. In different sector-

country combinations, lower absence rates were associated with either the frequency of
visits in the facility’s district or the facility’s proximity to a supervising ministry
office. 122 Since inspections and supervision are often problematic, focusing school-based
management on instructional time issues may be effective. School systems may also
offer intrinsic rewards to teachers who are rarely if ever absent. Methods such as bonuses
for evidence of attendance (e.g. a snapshot as used by a Rajasthani NGO) may succeed in
improving attendance. Workshops and in-service should not be organized during teaching
hours unless the school year is lengthened by a commensurate number of days. It may be
cost effective to pay teachers extra for in-service training.

Making supervision effective. Often supervisors really are not sure what to look for. One
reason for limited or non-existing supervision is the vagueness of observation criteria and
the large amount of time that teacher observation often requires. The guidelines and
requirements needed to focus local authorities on how well students are taught can only
take up a few minutes. The composite variables of the classroom Snapshot (interactive
and passive instruction, organization, no instruction, and percentage of students off task)
constitute a supervision framework and can furthermore be observed within a few
minutes. Other techniques of 5-minute supervision also exist and may be easier to use
than lengthier procedures often legislated for public schools. 123 In addition, showing
interest in teachers’ timeliness and praising them when they come on time and use time
well may achieve a lot.

Figure 4-2 Student groups engaged and Figure 4-3. Groups unsupervised,
supervised (Indonesia) engagement uncertain (Brazil)

Source: author

121. OED 2004a

122. Teachers in Bangladesh were 10% more likely to be absent in secondary schools that had never been
visited by education officials. Secondary school teachers were 68% less likely to be absent in schools
attended by pupils with better educated mothers (Chaudhury et al. 2004a).
123. E.g. Downey et al. 2004
 39

IMPROVING LOCAL GOVERNANCE AND STAKEHOLDER INVOLVEMENT

Central governments aside, schools and local educational authorities have also paid little
attention to time loss and its costs. As the data from Pernambuco showed, many school
and class registers had scarcely been filled, a sign that local authorities may not search for
the evidence of instruction that registries could provide. One problem is that schools and
teachers often get little supervision from principals and supervisors, and that time use is
not a variable that is used frequently. Hopefully the data from this and other studies will
have some effect as findings are disseminated.

Weak incentives for performance, corruption, imperfect monitoring, and administrative

logjams are frequent culprits. Some countries have tried to address the problem by
involving the stakeholders in service delivery. Sometimes results have been impressing.
Giving parents a voice over their children’s education, patients a say over hospital
management, making agency budgets transparent-all contribute to improving outcomes in
human development. 124

However, low-income parents often do not know how to recognize the features of
effective schools. It is possible to raise consciousness among parents improve local
governance by disseminating information on the visual hallmarks of effective schools:
teacher attendance, textbook provision, time use in interactive instruction. A potential
outcome measure to be used is that all students should become fluent readers by the end
of grade 2. 125 Parents and village elders (even those who are illiterate) can perceive and
monitor reading fluency, because it resembles human speech. Similarly, simple goals
may be established for math. These characteristics can be taught through videos that
contrast acceptable and unacceptable features in schools as well as examples of children
who can read fluently.

TRAINING AND CONSCIOUSNESS RAISING ON THE CONSEQUENCES OF TIME LOSS

Good time use in many societies is not a priority, and this concept may be novel,
particularly for educators who have spent a lifetime moving at a certain pace and
watching their own teachers do so. Suitable and “memorable” training is needed to show
the effects of time loss at all levels of the educational chain and to promote culturally
viable strategies for doing so.

ƒ High-level officials may be engaged partly through WBI (and the Joint Africa
Institute). Decision makers may be asked to think about the points in their
systems where intervention may increase instructional time and discuss measures
that are likely to be politically acceptable. (See exercise in Annex C.);
ƒ The media may be used to raise awareness the public and to discuss the budgetary
and learning implications of time use. Audiovisual means (presented during
meetings or through TV programs) can show what occupied and idle classes look
like as well as how children should read at various ages when classroom time is
used well; and

124. World Development Report 2004.

125. Abadzi 2006
 40

ƒ Principals, supervisors and inspectors at the provincial and local levels could be
trained to increase instructional time but also to document time use in the school
as well as the classroom. A key issue is to demand visible student progress at brief
intervals, thus creating pressure on teachers to progress. Classroom events can be
evaluated through the variables outlined in the classroom Snapshot (interactive
and passive instruction, management, students’ off-task). Principals may guide
teachers to maximize interactive instruction and minimize organizational
activities in order to maintain student attention and minimize time off-task.
District directors might make a clear commitment that they will ask for reports of
instructional time use and follow up with their subalterns. Closer control, follow-
up, and social pressure may result in reducing time wastage by teachers.

Incentives aside, there is a need for teachers to learn how to engage all the students all of
the time and how to manage classroom events so as to minimize time loss. Training
teachers to use time better is particularly challenging in low-income countries, given low
levels of education and a long tradition of time wastage at all educational levels. Teachers
tend to be unaware of the degree to which their habits impact student learning, and data
from this study suggest that some spend more time in “classroom management” activities
than others. To decrease off task behavior, teachers need better classroom management
strategies to gain and retain attention of the students. Thus, time management should be
included in pre-service and in-service curricula. However this is a challenging task, given
that training programs often fail to modify behaviors. Training programs in many
countries are weak on classroom management strategies, and outside expertise is needed.

Many techniques have been developed to decrease off task behavior and retain attention
of the students in the classroom that are often not part of teacher training curricula (such
as proximal control, signal interference, hurdle help, buzz groups, key group monitoring).
Also, bibliography on time savers for teachers can be adapted to lower-income
schools. 126 Exercises with videotaped evidence can help them reflect on the effects of
poor time use on students, their habits and the habits of their own teachers whom they
role-modeled. They can also learn to plan and organize the material better so that the
class can spend its time working rather than waiting.

Systematic classroom observation techniques like the Snapshot can affect positive change
in the teacher behavior when feedback is given from the profiles generated from them. 127
Additionally, teachers may give more credence to recommended changes in their
instruction if supported by evidence collected through observation. 128 There are some
challenges with using systematic classroom observation to assess instructional time;
notably the presence of the observer in the classroom may lead to invalid conclusions
based on reactive effects such as evaluation apprehension and socially desirable
responding on the part of the teachers and students. There are also questions about the
amount of time that is needed to collect reliable and valid information about classroom
behaviors. 129 Another difficulty in carrying out time-based planning is the discrepancy
between how teachers and students experience time. There is a tendency for most

126. E.g. Wachter and Carhart 2003; Gore and Dowd 1999; Wood 1999
127. Stallings and Freiburg 1991
128. Anderson and Burns 1989
129. Kurz 2003
 41

observations to focus on teachers and their behaviors, as opposed to students and their
learning processes. 130 The impact of time on pupil learning may also vary in relation to
the time of the day in which classes are held: morning, afternoon and, in some cases,
evening hours. Many schools teach the more demanding and/or important school subjects
during the early hours of the day, a practice indirectly acknowledging this relationship.
However, US research suggests that with feedback teachers can be taught to use time
more efficiently in primary and in secondary school. 131 Grant funds were used to pilot in
Pernambuco a five-day training course that had been used with teachers in low-income
urban areas of the US, but teachers in low-income countries often have low levels of
education. Additional experimentation is needed with the prospective trainees to
understand better how to proceed.

Figure 4-4: Indonesia: “Benefit from Knowing exactly what to teach on a given
work time as much as possible” day and how to do it should result in better
time use. If teachers do not know the
material or cannot answer questions from
students (as it often happens with primary-
school math), they may avoid teaching and
find other occupations while in school,
including bureaucratic work. Ignorance of
the subject matter (such as math of the
upper primary grades) is a little-understood
phenomenon that may compromise the
efforts of governments to appoint qualified
teachers, particularly in rural areas.
Source: author Furthermore, teachers may waste time
because they do not prepare their classes in
advance. There is a need to transmit a repertory of routines on how to teach various
topics so as to maximize time, how to evaluate quickly whether students learned the
material. Also, as in Argentina, giving teachers textbooks rather than making them do the
job of putting them together may ease the burden placed on teachers’ administrative
work.

OPTIONS FOR INCREASING STUDENT ACHIEVEMENT GIVEN LIMITED TIME USE

Efforts by governments and donors may improve time use, but inefficiencies are likely to
continue. Taking the time-related variables into account is now simpler thanks to recent
cognitive research that helps explain the role of practice and feedback as means to learn
and recall complex knowledge structures. 132 With more applied research it may be
possible to focus on activities that maximize learning efficiency and encourage their use
over others.

How can the poor learn basic skills in conditions of reduced time? Some options are:

130. Connelly and Clandinin 1993

131. Stallings 1975, 1980; Stallings, Needles, and Stayrook, 1979; Stallings 1985, 1986, Stallings and
Freiburg 1991; Knight 2001.
132. See Abadzi 2006, Chapters 8 and 10 offer also learning implications for various interactive and
passive activities.
 42

y Instructional time should be reallocated towards basic skills (reading, language

and math), reducing many of the less important courses and requirements that fill
curricula in many countries.
y Textbooks for every student to take home and use in class reduce the time needed
to lecture in class, copy exercises on the board, or hand out photocopies.
Increased homework assignments constitute additional instructional time. Older
students could help after-hours to help younger students with their schoolwork.
y Methods that have been shown to be more efficient should be disseminated
through in-service training for certain subjects, such as phonics for reading.

In most cases lengthening the school day or year would merely perpetuate the
inefficiencies. 133 However, poorer students may benefit from longer instructional days
when used well.

Figure 4-5. How many students seem off task in this caricature?

Senegal – folk artist on glass

INSTRUCTIONAL TIME MEASUREMENT FOR EVALUATION AND RESEARCH

The parameters measured in this research could be useful as monitoring tools of

governance and systemic efficiency. Well-managed schools should stay open for the
number of days that governments specify. Researchers visiting schools should find
school registries with numbers of days and hours per subject matter approximating those
specified by the government. Low teacher absenteeism (e.g. below 10%) would
constitute evidence of effective school-based management. Finally, time on task in class
should constitute evidence of effective supervision by principals, inspectors, or
supervisors. Measurements in Tunisia through this study (78%) and the Chicago public

133. From “Time and Learning in CA Schools” 1984 (unpublished mimeographic document).
 43

schools (85%; Smith 2000) suggest that 80% use of time could be realistically attained.
This 80% figure could be used as a benchmark for lower-income countries to attain.

Project evaluation challenges. The World Bank and other agencies try to evaluate
project outcomes by comparing baseline to post-intervention data. Instructional time use
is rarely mentioned explicitly during various interventions, and baselines are typically not
available. How can a single measure at the end of a project help evaluate outcomes when
the counterfactual is not known?

If project interventions resulted in better management, the amount of time used for
instruction should increase over a baseline and approximate 100%, the amount that
governments expect schools will devote to teaching students. Without a baseline it is hard
to determine the amount of increase, but a criterion-referenced approach could be taken.
Curriculum developers in a country can be asked to estimate how many class hours
students would need to master objectives of specific subjects in various grades and also
estimate the number of objectives likely to be missed if the available time were reduced
by 20%, 40%, or 60%. In some subjects the loss would be limited to that subject (e.g.
social studies), but in others, like math, losses in grade 1 would impact learning in higher
grades and result in cumulative losses. It is possible to estimate and model these losses
for evaluation purposes.

There is a further need to improve time measurement not only for greater reliability but
also for ease of use and cost effectiveness. The concepts are complex, and different
studies have used different methods. This project has resulted in improved modular
surveys that obviate the need to craft new questionnaires every time a new study is
initiated. Currently, data that would give a sense of baselines, range, and likelihood of
improvement are limited, but as more are collected, these indicators will become more
meaningful.

Research prospects. Much needs to be learned about the effects of time use on test
scores and reading fluency acquisition in the early grades. It must be better understood
how the various systemic inefficiencies impact the amount of time ultimately available
for students in different countries. Clearly causality must be established to link time use
and learning outcomes in longitudinal studies and clarify the relative importance of inputs
such as textbooks and other instructional materials. The advantages and disadvantages of
group work in time use must be also studied. Instructional time use in private and public
schools within the same areas must be compared and patterns must be better understood.
Memory research issues, such the rate and efficiency of consolidating new information
need to be better understood vis-à-vis the amount and placement of instructional time as
well as the activities that maximize long-term recall.

The future of the Education for All initiative may depend on how seriously governments
and donors take instructional time wastage. Hopefully this research is only the
beginning.
 45 Annex A

Annex A. Study Activities, Methodology, and Detailed

Findings

As mentioned in the main text, the team decided to work directly with line ministries rather
than hire local consultants in order to develop implementation capacity and ownership. Thus,
much of the work was carried out by government agencies in collaboration with the task
managers. Overall, the study carried out the following activities.

y A survey of the research literature and development of a research design in

collaboration with the UNESCO International Bureau of Education (IBE; July –
December 2003; see Chapter 2);
y Development of questionnaires to assess instructional time use at different stages. In
collaboration with IBE, four questionnaires were developed aimed at (i) school
principals, (ii) teachers, (iii) students of later primary grades, and (iv) parents. These
inquired about various aspects of time loss, its reasons, and suggestions for
improvement, and essential demographic information. The questionnaires were
translated into French and Portuguese (September-December 2003); Tunisia and
Morocco subsequently translated them into Arabic. Counterpart staff in the four
countries of the study were asked to use them but adapt certain items if necessary for
local conditions. (See prototype in Annex B);
y Classroom observations for instructional time use. The Stallings Classroom Snapshot,
a well-known US instrument was adapted for international use. The instrument allows
for specific, behaviorally oriented feedback to teachers and is useful for training
purposes in addition to assessment. Counterpart staff in participating countries
(students and professors in Pernambuco, pedagogical counselors and inspectors
elsewhere) were trained in its use in five-day sessions. To ensure the stability of
observational data, trainees passed examinations that included attaining acceptable
inter-rater reliability and were certified as competent to use the instrument consistently;
y Project launch workshop and training in Tunis (January 2004), where the first batch of
supervisors was trained and methodology was extensively discussed. Subsequent
training took place in Rabat (February 2004), Recife (June 2004) and Accra (October
2004);
y Data collection: School visits, administration of the questionnaires and the Snapshot to
a sample of schools in each country, cleanup and entry of data for later analyses (March
2004-July 2005 in various countries, see Figure A-2);
y Socioeconomic and achievement data collection for the schools included in the
samples. Government offices were asked to provide any available data. Test scores
were only available for Pernambuco, while socioeconomic indicators were available at
the district level in Pernambuco as well as in Morocco;134
y Development and piloting a training module to provide feedback to teachers on the
basis of the Snapshot results. Training of 28 district supervisors took place in Recife,
Pernambuco (October 2005) and was videotaped for future use and improvement. The
trainees were enthusiastic and learned the meaning of the codes and how to increase

134. Haut Commissariat du Plan 2004. Moroccan student grades were also made available, but these could
not be equated and used.
 46

active instruction and accomplish classroom management activities efficiently. (At the
time that the report was written, follow-up had not been conducted to find out whether
trainers had used these skills.);
y Development of new software for the scoring of the Classroom Snapshot to be used in
future assessments;
y Data analyses and report writing (January-August 2006). Data from the four countries
were analyzed separately as well as collectively. In Tunisia and Pernambuco, the data
were also analyzed locally, and reports were written that we taken into account in this
document;
y Dissemination. Instructional time policies would be presented to World Bank staff in
October-November 2006. In collaboration with IBE an international conference has
been planned to share the results of the research, sensitize policy makers to the issue of
time, and plan for effective interventions; and
y Application of this work in India under different funding arrangements.

School Sampling Design

The study was conducted in Brazil, Ghana, Morocco and Tunisia from 2004 to 2005 (Tables
A-1 and A-2). Schools in each country were chosen through cluster sampling to represent
the country’s demographics, the regions of the country, and urban and rural populations.
Ultimately, the choice represented local realities and convenience. In Pernambuco, 180
schools were selected from areas relatively close to Recife (in 10 of the 17 administrative
districts), but also schools that had scored low, middle, and high in the state’s standardized
achievement test of 2002 (SAEPE). In Tunisia, the sample included 51 public and
community-managed schools chosen randomly in six districts (Tunis, Le Kef, Monastir,
Kairouan, Gabès, and Tozeur).

Table A 1: School location by country

Pernambuco Ghana Morocco Tunisia
Urban 174 45 87 16
Peri-Urban 0 34 NA NA
Rural 6 124 83 26
Total 180 203 172 51

Note: The location of 9 Tunisian schools was not specified

Within the schools, certain grades were chosen for observations through the Snapshot. In
Tunisia grades 1, 2, and 4 were observed (classes chosen at random through SPSS in case of
large schools), while in Brazil grades 2, 4, and 8 were selected. In Morocco, grades 1, 4, and
6 were selected and in Ghana, grades 1 through 6 were all observed, although grades 2 and 6
had a greater frequency of observations. Morocco and Tunisia carried out two observations
per classroom.

The plan was to collect the information in each country in the course of about 5 weeks
through 15 teams of 2-5 enumerators. Some areas took longer; in Pernambuco data were
collected in June-August 2004. Questionnaires and observations were administered
concurrently.
 47

Teams of trained observers visited schools, spending about three hours in each. They
administered questionnaires, ascertained how many teachers were absent on that day, and
observed a subset of classes through the Classroom Snapshot (grades 1-2, 4, and 6).
Interviewers were expected to verify teacher attendance in person, but in some areas the
teams concentrated on observing instructional time, and schools staff were asked to fill out
questionnaires and return them by mail. Thus teacher absences were not always directly
observed or verified through registries.

Table A 2: Respondents of Survey Questionnaires

Cross Country Comparisons
No. of Respondents Pernambuco Ghana Morocco Tunisia
Schools 180 203 172 51
Principals 180 214 172 51
Teachers 539 419 343 141
Students 178 1267 3916 1275
Parents 178 417 1960 1201

Methodological Challenges and Solutions

Obtaining instructional time loss estimates that could be compared across countries and doing
so with reasonable costs and logistics posed considerable logistical and analytical challenges.
The section below details the decisions and outcomes obtained. Future work should build on
these experiences.

The questionnaires developed jointly by the World Bank and the UNESCO-IBE were
translated into French and Portuguese and given to Ministry of Education staff during a visit
to the country. Since not all systems are the same, staff were asked to adapt some questions
as needed in the local context. However, the concept of measuring time loss was unfamiliar,
and some changes were excessive. Countries did not include all items, or they modified them
in ways that changed the information to be obtained. In some items wording became unclear
or incomprehensible; for example it was uncertain whether teachers in Tunisia reported only
unexcused absences and whether Moroccan participants understood the sophisticated
vocabulary used in the questions. Other items meant to elicit specific figures were replaced
with ranges (e.g. e.g. whether schools were closed for 1-3 days, 4-7 days, 8 days or more). In
Pernambuco, much of the questionnaire content was discarded and replaced with items that
invited vague responses (e.g. teachers were asked if they were late “often, sometimes,
never”). It became necessary to administer the questionnaire (approximately a year later, on
June 13-17 2005) to a subsample of 40 Brazilian schools to obtain more specific estimates of
time lost. Some differences were discovered during data analyses, and significant data
cleaning and recoding were required to make comparisons possible. (See results section.)

One-time visits and triangulation of opinions. In principle, it would be preferable to visit

schools repeatedly and register attendance, as teacher absenteeism studies have done. To
pilot rapid and inexpensive methods it was decided to carry out one visit and get information
from multiple sources, including comparisons with the previous year. The one-time survey
had disadvantages that included potential systemic biases regarding the day of the week and
season. Several assumptions were made for the purposes of data analysis (see sections
 48

below). Nevertheless, a single visit may reliably provide overall time loss estimates; a
Dominican study that involved three monitoring visits showed that the average time schools
dedicated to learning was overall the same (65% in the first two visits and 64% in the third;
EDUCA 2005.)

Estimation of time lost. Studies that involved repeated and lengthy observations (e.g. in
Bangladesh) sometimes found large amounts of lost time that the current study evidently
missed. These were due to inquiries about time losses that the current study did not
foresee or inquire about: frequent cancellations of the first one or two instructional hours
due to rain (in Pernambuco and also in a 2004 study on Bangladesh), impromptu strikes
and walkouts by students and teachers (according to a 2005 Senegalese study) or
preparation for athletic events (according to a 2003 Ghanaian study). Unfortunately the
BNPP study did not have the benefit of these insights and did not explore the loss of time
in sufficient detailed. Human memory tends to fail when asked about the frequency of
repetitive events, and school staff may have unwittingly or deliberately given low
estimates of school closures. Poor recall may have been more pronounced towards the
end of the school year. For example, Moroccan principals reported few if any days of
closure (mean 1.38 days) despite earlier informal observations suggesting frequent
cancellations (e.g. OED 2001). By comparison, Tunisian principals may have been more
accurate in reporting closures (projected at 5.15 days), and may thus appear to use time
less well. (Subsequent surveys must be more specific, graphically showing the calendar
days in the school year and inquiring about absences before and after specific holidays.)

School and classroom registries in principle should provide attendance data or evidence
of schooling. The team was warned that these are often incomplete or missing, and
indeed they proved unreliable. Though all countries were asked to find these documents
and register their contents, none of them did. (Tunisia reportedly has these data
computerized, but linkage to the database has been difficult.) The registries for grades 1-8
were specifically sought in a supplemental study of 40 Brazilian schools, and it was found
that 52% of them did not exist in class; of those that did, 74% of them were not filled.
Apparently there was no systemic need for this information.

Instruments of the Study and Their Use

(a) Surveys on instructional time use. The questionnaires querying principals, teachers,
students, and parents on time use were collected on the basis of the schedule shown on table
A-3. Other data on absences, the use of homework and demographic data were also
collected. (Morocco added an observer module.)

The questionnaires were administered in the various countries at different times during the
school year. It would have been desirable to coordinate administration and eliminate a
source of error related to requesting recall at different times of the year. With the passage of
time participants possibly lose track of the number of days schools were closed. For future
administrations a school calendar showing the days of operation should be obtained from
the government and shown to participants to refresh their memory.
 49

Table A 3: School Operation and Survey Dates

Questionnaire Data Approximate % of
Official
Initial and Final School Collection Period School Year Past at
school days
Year Dates (midpoint used for Midpoint of Survey
per year
analyses) Period
Pernambuco 200 First week of February July 2-August 31 About 102 days
(Brazil) – first week of (midpoint Aug. 1 2004) passed, or 51% of the
December time
[10-16 July holidays]
Ghana 197 September 16-August 5 October 25- November About 42 days passed,
30, 2004 22.8% of the time
(midpoint Nov. 13)
Morocco 204 September 8 -June 30 November 29- About 75 days passed,
(2nd Wednesday of December 7 2004 36.8%
September) (midpoint Dec. 3)
Tunisia 190* Sept. 15-June 30 May 24-29 2004 About 162 days
(midpoint May 26) passed, 87.4%

*Note: Tunisian schools are open 160 days for grades 1 and 2, but higher grades were observed, and teachers of lower grades must be
present in the entire school year. Of 190 days.

Stability of survey measurements. Surveys were conducted only once and at uncontrolled
periods of time. However, if the sources of error are random, findings may be consistent
across time. One opportunity for verification came with a 40-school sample in
Pernambuco almost a year after the original survey (see Table A-4). Although there was
some variance in the components, overall time loss was about the same as that of the 180-
school sample. The 2005 EDUCA survey in the Dominican Republic that involved three
monitoring visits showed that the time schools dedicated to learning was overall the same,
64-64%. Similarly the USAID-financed Ghana study found teacher absenteeism rates to be
about the same during repeat visits.

Table A 4: Instructional time loss estimates from observations in 40 Pernambuco schools

Teasons for Time Loss Projected to Year-End
Number of school days 200 Comments
Delayed start of school 0.58 18 classes did not start for 21.3 days
Classes canceled after start 1.83 mainly due to a strike
Teacher absenteeism 25.49 observed as 12.9%
Teachers late to class 1 average 5.8 minutes
Early departures, class ending 0.29 average 1 minutes

Days Lost 29.19

Percentage of Loss 14.60%
Days Remaining 170.81
Engagement Rate in Interactive Classroom Snapshot results
or Passive Classroom Tasks 72.10%
School Days Devoted to Learning 123.15

School Year % Spent Engaged 61.6%

in Learning Tasks
 50

(b) Stallings Classroom Snapshot. Use of time in the classroom was assessed through the
Stallings Classroom Snapshot. The same sampled schools were used. The observed
classrooms and teachers were drawn form a number of grade levels and a variety of
subjects (Table A-5).

There have been many instruments and attempts within the field of teacher observation.
Many of the observation instruments developed over the years are qualitative, require
personal interpretation, and often rate variables on Likert-type scales (e.g. classroom climate
is poor, moderate, excellent). There have been few examples of quantitative instruments or
attempts to compare time profiles 135 with normative, actual patterns of time allocation, and
the ability to give feedback to teachers and school authorities.

Table A 5: Sampled school locations by country – Snapshot in specific grades

Note: subjects observed through the Snapshot were: Tunisia - reading, language, math, science, social studies, other; Morocco -
reading, math, science, other; Ghana: math, English grammar, Ghanaian grammar, science, environmental studies, religion;
Pernambuco: Portuguese, math, science, geography, religion.

The Stallings Observation System, often called Classroom Snapshot is a “low inference”
observation system used in a number of teacher and school effectiveness studies. 136
Observers trained through a five-day detailed process take a “snapshot” of classroom
events every five minutes. They note codes related to various events on a one-page
matrix (Annex B-1). This rather complex instrument measures the incidence of about 52
variables that include activities, interactions, materials used, and instructional strategies
by teachers and students. Thus, in addition to documenting how time is spent in a class,
the instrument records events related to quality of education. Through extrapolation, it is
assumed that the same event that has been “photographed” will continue for the next four
minutes, although this does not necessarily happen. For the study, an observer stayed in
the classroom for an entire hour and produced 10 “snapshots” that sampled classroom
events.

The Stallings Observation System was initially developed to measure the level of
teachers’ implementation of twelve early childhood and elementary school Head Start
and Follow Through models, and was modified in 1977 for secondary school classrooms.
The system consists of two independent observation systems, the Five Minute Interaction
and the Snapshot, both of which use time sampling. The Five Minute Interaction
observations are coded in five minute periods while in the Snapshot, any interaction is

135. Karweit 1988

136. e.g., Anderson, Ryan, & Shapiro, 1989; Goodlad, 1980; Stallings & Kaskowitz, 1974; Teddlie, Kirby,
& Stringfield, 1989. “Low inference” means that the observations are recorded as seen, with little personal
judgment and interpretation necessary.
 51

coded at the moment, as if a picture were being taken. 137 To develop the instrument,
Stallings made detailed observations about the presence or absence of the desirable and
less desirable behaviors in a class, invited teachers to comment, and then observed them
in the fall, winter and spring semesters. Some teachers were placed in the workshop
group (experimental group) and others were not (control group). For the most part, the
treatment teachers changed their classroom behaviors and distributed time across the
recommended activities. Stallings reported that their students gained, on the average, six
months more in reading than did the students of the control group teachers. Further
observation indicated that treatment teachers maintained most of their behavior changes,
whereas control teachers' classes became more lax and less task-oriented. 138 Similarly
effective were efforts in secondary school. 139

Studies helped establish benchmarks for effective use of time in classrooms in the US. 140
Most teachers monitored students doing written work at their seats for 50% of the time, did
organizational activities for 38% of the time, and actively instructed students only 12% of
the time. On the other hand, effective teachers distributed their time so that 50% or more
class time was spent in active instruction, 35% in active monitoring, and 15% or less in
organizing and managing. In average classrooms, students were actively engaged in their
work only 15% of the time and were off task 85% of the time. But in the most effective
teachers’ classrooms the students were engaged 94% or more of the time and were only off
task 3% of the time. These data helped establish benchmarks for time use in classrooms
that were used to some extent as reference points in the study.
Snapshot data are summarized as amount of class time spent on three composite variables
(interactive instruction, passive instruction, and organization-management) and student
off-task rates. Based on empirical research, the test developer recommends that 50% or
more of class time be spent in interactive instruction, because in US studies interactive
time was found to increase test scores. 141 Passive instruction includes seatwork and
copying should take up no more than 35% of class time. Passive instruction can be used
judiciously to have students review what they have learned, but repetitive work may
result in inattentiveness and limited new learning. Organizing and management include
giving directions, dealing with intrusions and visitors, and such activities as taking roll
and lining up for a fire drill. These elements of the class should take up no more than
15% of class time. It is normal for students to stop paying attention at some point in
time, but observed student off-task behaviors should not exceed 6% of the time.

The study adapted the Snapshot for use in low-income classes outside the US; in
particular, a category of copying was added to activities and codes related to discipline
were de-emphasized. The developer did not did not foresee the possibility of teachers
leaving the classroom unattended (which in the US is illegal) but this was observed
frequently in other countries, and a category was added to describe a teacher out of the

137. Kurz 2003; not used in the current study

138. Stallings 1975, 1980
139. Stallings, Needles, and Stayrook, 1979; Stallings 1985, 1986
140. Stallings 1986
141. Schaffer et al., 1992; Stallings, 1980; Teddlie et al., 1989.
 52

room or disengaged from students. 142 The instrument and manual were translated into
French and Portuguese prior to training and reviewed by language experts and educators
for accuracy. The five-day training included practice in coding vignettes of potential
classroom activities and two days of classroom observations and coding practice. About
25 people were trained in each country. In Morocco, Tunisia and Ghana, trainees were
inspectors or pedagogical advisors, and in Pernambuco they were university students;
participants had to demonstrate competency in scoring and comprehension of the
instrument’s principles, and were certified.

Data from the snapshot were initially entered into a program using Microsoft Access.
However, the software was found to be inadequate for large-scale use, and new software
was developed with grant funds and piloted in India on an activity that was not financed
by the grant. Further modifications are being made as the needs are better understood.

The training provided on the classroom Snapshot proved very popular, and many
supervisors realized that they could use the same parameters in their everyday work. On
the downside, one-hour observations were time-consuming. Also, repeated training was
necessary to clarify the doubts that some participants had regarding the coding use.
(Efforts have been made to simplify the instrument.)

A principal components factor analysis with varimax rotation on a dataset with all
countries merged showed 19 factors with eigenvalues above 1.00. This large number was
to be expected, since that this instrument measures many variable that have relatively low
intercorrelations. (Though some were as high as 0.85, most ranged from 0 to 0.5).
Several of the Snapshot variables had significant loadings (> 0.30) in more than one
factors. The largest factors were related to (a) off-task behavior and management, (b)
interactive instruction and not seatwork, (c) reading, books, and instruction, (d) teaching
materials other than books, (e) copying, assignments, and seatwork, (f) group size and
cooperative learning, and (g) teacher management activities. Factors 1 and 7 seem to
measure the obverse, with many factor loadings negative on one and positive on the
other. Three variables of low incidence did not have significant factor loadings, and two
of these that pertain to other adults in the class have been removed from future
administrations.

Data Analyses, Challenges, and Solutions

Data were analyzed in the field as well as at the World Bank. The Tunisian CNIPRE and
the Secretariat of Education in Pernambuco submitted reports with analyses of their
respective datasets that have been included in the analyses presented in this report. The
World Bank team carried out additional analyses on the data of all four countries. To
make them compatible, certain procedures and assumptions were used. Although they
magnified measurement error, they were the only means to compare time loss across
countries. Parametric and multivariate statistics were calculated with some variables
after ascertaining that normality and homoskedasticity assumptions had not been unduly
violated.

142. “Teacher out of the classroom” was added after this problem was observed in Brazil, but later it was
modified as “teacher alone” to describe a teacher who was not interacting with children: sitting and grading
tests or something else, writing lessons on the blackboard.
 53

Recodes to create interval-scale data. In some countries, questions asking for specific
numbers had been converted to ranges with open ends. (For example, the number of days
schools were closed was often coded as 1 for 1-3 days, 2 for 4-7, days, and 3 for 7 or
more). In those cases, the items were recoded; the midpoint and frequency of the ranges
were used to calculate weighted averages and use them in further analyses. On some
occasions it proved impossible; the school codes and some data on school closures of
Tunisia had been entered incorrectly and could not be analyzed properly.

Corrections and assumptions for survey administration dates were as follows:

Governments administered surveys after deliberations lasting several months and at a

time convenient to them. The team was apprised of the dates, but it was not possible to
control in advance which day and time of the year the surveys were administered. To
arrive at a specific number of days lost based on the responses, the team calculated the
median date of the periods that surveys were administered in each country and projected
respondents’ answers to the end of the school year based on the fraction of school days
that had already passed. 143 (See more details in Annex B-4). An assumption was made
that the percentage of wasted time would be the same for the remaining school year, that
variations would be would be randomly distributed, and that surveys were equally likely
to be filled at any day of the week. This was important because absenteeism rates may
vary during the week; in Pernambuco absenteeism could be 55% higher, depending on
the day (Figure A-4). However, on many surveys the exact date was not noted, and the
potential bias could not be estimated.

• Different countries have different blocks of vacations interspersed, but it proved

difficult in some of them to get a clear number of school days past as a function of
the school calendar at the time the surveys were given. For that reason, the 180-
200 school days were assumed to be interspersed evenly in the course of nine
months. Corrections were then made to the data. For example, Tunisian
principals stated that schools had closed for about 4.50 days, but this happened
near the end of the school year, while countries reporting fewer days did so closer
to the beginning of the school year.

• Early departure was meant to measure instances of teachers dismissing their last
class early and leaving the school, but due to a lack of clarity in the relevant
survey questions, early dismissal was calculated for every class; where actual
early departure times were unavailable, a 5-minute period was assumed. (Future
surveys will provide clearer definition of what is meant by teacher early
departure.)

Triangulation of responses and computation of time loss involved the following

processes:

143. Brazil expects 800 teaching hours (20 hours per week) in grades 1-4, and 1000 grades 5-8. Tunisia
expects 160 days for grades 1-2 (704 hours in 32 weeks of 22 hours each) and 190 days for grades 3-6.
Grades 3-4 have 800 hours in 32 weeks of 22 hours each, while grades 5-6 have 960 hours. Aside from
there days the system has about 48 days of holidays. Schools in the Tunisian Priority Education Program
have an additional two hours of work a day plus remediation for all students. In all cases, the teachers must
be there for 190 days.
 54

Efforts were made to capture and calculate various segments of time lost. Spreadsheets
were developed to compare the frequencies of similar survey answers among the
principals, teachers, students, and parents of each country (and ultimately across
countries). Variation among responses was studied on the issues of school closures,
attendance, tardiness, homework, and reasons for absenteeism. Reports about time loss in
the previous year were also compared with those of the current year. Then algorithms
were developed to estimate the various sources of time loss reported by the various
stakeholders, and variability of comparable measures was studied. Averaged across
schools within a country, it was found that various stakeholders differed relatively little in
their estimations of time loss from a particular source. The time loss averages are shown
in Table 1 of this report.

In a country or district some schools or just some classes may fail to start on the
appointed day. Some sections may start and stop for a period due to teacher
reassignments or other issues. To calculate the percentage of time lost through small
segments, the concepts of school-days and class-days were used. Based on the number of
official school days, the numbers of existing classes were multiplied by the number of
days they should be functioning, and the amount of time lost by specific grades in
specific schools was deducted. (If only schools were involved in a question, then school-
days were used.) This method was used when the questionnaires obtained more specific
information, as in Pernambuco. It was also used to calculate the amount of time teachers
were delayed or dismissed classes early. For example, 17.5% of Moroccan teachers were
reported to be late for 9.8 minutes coming to school. This amounted to a 3.4% loss of a
50-minute hour if spread among all teachers. To calculate the number of days lost, the
percentage was deducted from the number of days schools were open and teachers
present.

Time losses were added up. The number of days lost and reported teacher absenteeism
were projected as a percentage of time given the number of days in the school year
already past at the time that surveys were conducted (see assumptions section above.)
Losses were computed in each case by deducting various numbers from net number of
days available. When schools are closed, fewer days are left. So, teacher absences were
calculated as a function of the days schools were open. Also, teachers cannot be late
when schools are closed or when they are absent, so delays were calculated on the
number of days schools were open and teachers present. (For subsequent surveys a
computer program may be needed to make such calculations.)

The Classroom Snapshot data were in the form of percentage of scans that showed a
certain characteristic (scans were usually 10 unless a class started late or was dismissed
early). Thus, they had a minimum value of 0 and a maximum of 100. (Teacher activity
data summed to 100% of time calculations, whereas student activity data were dependent
on the size of the group a teacher was addressing.) Some non-parametric statistics were
carried out, but most of the time the data were treated as parametric. The Snapshot data
were at the teacher level and included a few classroom-related variables. To these,
school-level questionnaire replies were linked and were used for various analyses. The
Snapshot data had the same format across countries, so they were analyzed comparatively
to compare performance among countries and among selected grades.
 55

This section discusses the output produced by the classroom Snapshot, with a focus on
the major activities taking place in class, how the teacher is addressing the students
(whole group, small group, one-on-one instruction), and what materials, if any, are being
used. Emphasis was on recording teacher activities. (Student activities may be to some
extent different, particularly if they work in groups or carry out independent work, but
this does not happen frequently in the countries where the research took place.) Table A-
6 shows the average overall performance of the observed classrooms in the participating
schools on a country-by-country basis. Some classes have been shown to carry out a
certain activity 100% of the time, though the maximum number is usually lower. Because
activities change often, standard deviations are large, sometimes larger than the means.

Table A 6: Classroom Snapshot – Country-level Performance and Criteria

Composite Variable U.S Classroom Brazil Ghana Morocco Tunisia
Criteria % % % %
(Number of teachers) 539 410 197 150
Interactive Instruction 50% or more 52.4 59.9 62.8 61.2
Passive Instruction 35% or less 19.6 10.3 19.9 25.6
Total Instructional Time 72.1 70.2 82.6 86.7
Organizing/Management 15% or less 27.9 28.0 17.8 13.3
Student Off Task Rate 6% or less 19.3 21.1 9.2 9.9

Key
 10% above recommended minimum for interactive instruction
 Above recommended minimum for interactive instruction
 Above recommended maximum for organizing, management and Student Off Task Rate
 10% above recommended maximum for organizing, management and Student Off Task Rate

Interactive (active) instruction. This includes reading, explanation, discussion, practice

drills, and kinesthetic exercises or projects. Brazil and Ghana reported over 50%
interactive instruction while Morocco and Tunisia reported over 60%. Although this
seems encouraging, this category includes lecturing, which if it goes on past a few
minutes turns into “passive” instruction. The instrument does not distinguish well among
shorter and longer periods of lecturing. 144

Passive instruction refers to students doing individual work seated at their desks. It
includes teacher’s monitoring of seatwork and copying, and is often related to lower
achievement. This is why it is recommended that no more than 35% of the time be spent in
it. However, some practice is necessary particularly in math, and the time may be well-
used if insight rather than repetition is promoted in solitary study. Students spent a low
10.3% of their time in such activities in Ghana and a high 25.6% of the time in Tunisia.

Students’ engagement rates. The figures of interactive and passive instruction combined
give the total percentage of time that students are engaged in learning activities. In
Morocco and Tunisia figures are in excess of 80% while in Brazil and Ghana they are just
a bit above 70%. Thus, Ghanaian classes give students about 16% less instructional time
than Tunisian classes.

144. Student activities in Tunisia, where teachers control classes, closely followed teachers, but they may
not be elsewhere. However, this aspect of the Snapshot was not sufficiently discussed in the research.
 56

Organization-management. This remaining time is strictly speaking not instructional,

though it may involve necessary tasks such as instructions for the next day or notebook
delivery. Activities may also involve social interactions among teachers and students 145
or describe a teacher who works alone, does not interact with students, or may have left
the room. Such activities should take up no more than 15% of the time. Of the four
countries, only Tunisia is below 15%, while students in Ghana and Brazil spend about
28% of their class time in such ‘management’ activities. Teacher alone rates were almost
zero in Tunisia and 4% in Morocco, but they rose to 12% in Pernambuco and 15% in
Ghana. This implies time spent in activities that involve no teacher feedback and
monitoring.

The relationship between discipline, management activities and instruction is apparent

when the Snapshot data of all countries are analyzed together. Discipline and off task
behavior were correlated 0.30; management was correlated 0.43 with off task behavior
and -0.59 with interactive learning (Table A-21). Also averaging activities across
snapshots and producing correlations makes it possible for large patterns to emerge.
Management correlates highly with students being disciplined (r=0.73) and practice drill
(0.86), but highly negatively with assignments (r= -0.93) because the teacher is busy.
Copying is an opportunity for classroom management (r= -0.55) and social interaction by
students but also for reading (r=0.64).

Despite the variability and particular circumstances of different countries, the concepts of
interactive and passive use of time, management, and student off-task rates have
consistent relationships among them (Table A-7).

Table A 7: Relationships between categories of time use in the Classroom Snapshot

Cross-Country Teacher Student off- Interactive Passive

Correlation Reading Instruction Task Learning Learning Management
Teacher reading 1 -0.28 -0.14 0.31 -0.12 -0.23

Instruction -0.28 1.00 -0.05 0.38 -0.18 -0.25

Student Off Task -0.14 -0.06 1.00 -0.24 -0.21 0.43

Interactive learning 0.31 0.38 -0.24 1.00 -0.56 -0.59

Passive learning -0.12 -0.18 -0.21 -0.56 1.00 -0.30

Management -0.23 -0.25 0.43 -0.59 -0.30 1.00

**significant at the 0.01 level (2-tailed).

*significant at the 0.05 level (2-tailed).
Note: N=1651
Students off-task. This very important variable of the Snapshot is computed separately
and on the basis of visual observations of students who seem uninvolved with classroom
activities. (More psychophysiologically direct means of assessing involvement are
beyond reasonable cost or ease of use.) It estimates the number of person-minutes

145. In the US children have discipline problems, so social interaction variables are important.
 57

wasted in a class (e.g. 50 students unengaged for 20 minutes). In all countries of the
study, the off-task behavior among students is above the 6% recommended maximum.
However, Morocco and Tunisia have rates around 9%, (with socializing a low 3% and
being disciplined 1% in Tunisia) whereas in Ghana about 21% of the students appear not
to pay attention to the classroom proceedings.

An absent-minded student is probably not learning, so off-task behavior reflects

involvement in learning and has been used as a proxy for learning in some studies. High
levels of off-task behaviors are found in classrooms with low performance on tests.
Further more, there is a relationship between student off-task behavior and teacher
management activities. Classrooms with lower rates of instruction tend to have higher
rates of off-task behaviors. This may happen in part because often several minutes pass
in each class before instruction begins or tasks change, and during that time students high
rates of off-task behavior. In Ghana, for example the teacher was disengaged from
students 15% of the time and organized teaching for another 10% of the time, while
students were off task 21% of the time. In Brazil teacher off task correlates 0.47 with
student off task rate.

Off-task rates are certainly related to classroom activities (Table A-8). In Pernambuco the
three classes with the highest percentage of reading time (67%) had off task rates of
about 10%; by comparison, the five classes that spent no time in reading had off task
rates of 83%. In Tunisia where off-task rates are lower and time used efficiently,
correlations with activities tend to be lower due to the restriction of range.

Table A 8: Correlations between Student Off-Task and Engagement in Instruction

Average
Teacher Off- No. students
Correlation with Interactive Passive Management class size
task in class
(range
Pernambuco -0.18 0.23 0.49 0.46 -0.25 25
(3-65)
Ghana -0.41 -0.15 0.30 0.33 -0.01 26
(2-82)
Morocco 0.08 -0.06 0.05 0.03 0.06 31
(18-55)
Tunisia -0.17 0.0 0.11 0.22 -0.14 26
(11-40)
Source: Stallings 2006

In Ghana there are few books, or paper and pencils necessary for seatwork, so the
teachers may rely more upon class discussions. According to the Ghana USAID study,
only 63.7% of the teachers were following the approved timetable – a proxy for
curricularly relevant time. Twenty-three percent of the teachers do not have lesson plans,
and of the 77% who do, 35% do not use them. They teach “off their heads”, risking
disorganization. The teacher disciplining variable was an average of 1.3% indicating that
the misbehavior was going unchecked. Similarly, in Pernambuco, the highest off task
variable in all grades was socializing among students, 12% of the time. However, the
“student disciplined” variable is a low 2%, indicating that teachers rarely stop the off-task
behavior (Stallings 2006).
 58

Stability of Snapshot measurements. It is reasonable to expect that the amounts of time

involved in various class activities would change from hour to hour (one reason why
variance of time parameters within schools is greater than variance between schools.)
However, initial studies with this instrument showed considerable stability (Stallings and
Kaskowitz 1974). Similarly, in Morocco and Tunisia, correlations between the first and
second administration to the same teachers showed moderate but significant correlations
(Tables A-9, A-10). The highest correlations involved management and student off task
behaviors. This suggests that some teachers systematically use time more or less
efficiently and commensurately their students pay attention less or more. With the
Snapshot data at hand, teachers can receive feedback and training.

Table A 9: Morocco - Correlations between Repeated Observations of Same Teachers

1st Observation 2nd Observation

Mean Mean

Interactive Instruction 61.35 63.76

Reading 13.15 18.0*
Instruct Explain 26.89 26.21
Discussion 6.19 6.94
Practice Drill 13.59 11.04*
Kinesthetic 0.41 0.20
Projects 0.30 0.71
Passive Instruction 20.59 19.44
Monitoring Seatwork 16.13 13.92
Monitoring Copying 4.46 5.52
Organizing Management 18.42 17.31
Giving Assignments 6.58 6.08
Managing with Students 9.17 8.91
Disciplining Students 1.41 1.07
Managing Alone 0.10 0.24
Student Off Task Rate 9.22 9.34
Being Disciplined 4.17 4.54
Socializing 4.08 3.87
Uninvolved 0.97 0.97

Some studies ask teachers to self-report on how they spend classroom time. Tunisian
teachers were asked during the survey. As might be expected, interactive time was
overestimated at 69.4% compared to the real average of 61%.

Classroom observers coded the Snapshot with reading activities mainly when the teacher
used books, (r=0.39), but did not usually code activities as reading when the teacher used
chalkboard or students used notebooks. Also, independent student activities involving
reading were not analyzed in sufficient detail. Since some students must have been reading
 59

the written contents, the possibility arises that reading activities were underestimated
through the Snapshot.

Table A 10: Correlations between repeated observation sessions

Student
Interactive Passive Organizing/
off Reading Seatwork Copying Instruction Discussion
instruction instruction Management
Task
Morocco 0.21 0.19 0.31 0.04 -0.12 0.18 0.21 0.39 0.41
Tunisia 0.32 0.16 0.46 0.26 0.26 0.19 0.37 0.44 0.38
 60

Table A 11: Average percentages of Time Use in 4th Grade

Student off interactive Organization-
Discipline Task learning Passive learning management
Country % % % % %
Pernambuco 1.75 18.21 52.89 18.61 28.50
Ghana 1.44 25.92 52.50 12.50 35.00
Morocco 4.20 9.37 62.85 20.03 17.78
Tunisia 0.88 9.64 61.70 26.32 11.98

Table A 12: Ghana average percentages of time use by location

Location Rural Peri-urban Urban
Interactive 59.2 59.5 63.6
Reading 9.8 6.03 6.3
Instruction 19.2 20.2 22.9
Discussion 22.9 26.6 26.7
Practice Drill 6.3 6.0 7.7
Passive 9.7 12.1 10.5
Seatwork 6.8 9.5 8.8
Copying 2.9 2.6 1.7
Management 27.7 28.1 263
Total engaged time 68.9 71.6 74.1
Student Off Task 21.1 19.9 22.0

Table A 13: Morocco average percentages of time use by location

Location Rural Urban
Interactive 62.6 63.4
Reading 15.8 16.1
Instruction 28.7 24.4
Discussion 6.8 7.4
Practice Drill 9.9 14.1
Passive 21.3 18.3
Seatwork 15.4 14.2
Copying 5.8 4.1
Total Engaged Time 83.9 81.7
Management 16.8 18.6
Student Off Task 9.6 8.7
 61

Table A 14: Pernambuco average percentages of time use by location

Location interior peri-urban urban
Interactive 57.5 48.5 49.0
Reading 8.3 5.1 5.9
Instruction 37.5 29.1 28.7
Discussion 6.9 6.0 5.5
Practice Drill 1.8 1.3 0.8
Passive 19.3 20.8 17.1
Seatwork 16.7 17.8 13.2
Copying 2.6 3.0 3.9
Total Engaged Time 76.8 69.3 66.1
Management 23.2 30.7 34.5
Student Off Task 16.2 21.1 24.3

Table A 15: Pernambuco Average Percentages of Time Use by Grade

Grade 2 4 8
Interactive 52.9 52.8 51.8
Reading 7.8 7.9 4.3
Instruction 30.2 31.0 37.1
Discussion 5.9 7.1 5.8
Practice Drill 2.2 1.4 0.7
Passive 16.0 18.8 23.2
Seatwork 13.5 16.6 18.9
Copying 2.6 2.2 4.3
Total Engaged Time 68.9 71.6 74.0
Management 31.1 28.4 25.1
Student Off Task 22.6 18.5 17.8

Table A 16: Ghana Average Percentages Of Time Use By Grade

Grade 1 2 3 4 5 6
Interactive 71.1 58.6 50.0 57.5 50.2 62.4
Reading 14.4 10.4 8.8 16.3 4.1 6.7
Instruction 16.7 18.8 16.3 11.3 17.0 22.4
Discussion 20.6 21.4 22.5 26.3 24.2 27.6
Practice Drill 16.7 7.3 2.5 3.8 4.1 5.2
Passive 4.4 10.5 8.8 17.5 14.3 9.6
Seatwork 3.9 7.4 2.5 11.3 11.6 7.1
Copying 0.6 3.1 6.3 6.3 2.8 2.6
Total Engaged Time 75.5 71.1 58.8 75.0 64.5 72.0
Management 23.3 29.0 38.8 25.0 33.1 26.1
Student Off Task 19.7 24.7 26.5 21.3 23.8 16.6
 62

Activity changes in an instructional hour

Activities in class do not take place at random; there are sequences that sometimes differ
from one country to another. Figures 8-10 show the flow of various activities in Tunisia,
Ghana, and Pernambuco. (It was not possible to conduct this analysis for Morocco.) Students
tend to be off task more frequently in the last parts of the hour. There does appear to be
engaged in seatwork more towards the end of the class period while instruction is more
frequently observed in the beginning of the class (see Figure A-10 for Tunisia). This is one
reason why it was important to observe an entire hour and take snapshots rather than visit
classes at some point during the period.

Correlations between the serial number of snapshots and various activities confirm the
tendency for drill, adult management and seatwork to be done towards the end of the period.
Classes in Ghana lasted only 30 minutes, and this may be one reason why management takes
up so much time compared to other activities. Also reading along with instruction takes
place earlier in Ghana.

Figure A-1. Activity Changes in Brazil

100%

90%

80%

70%

60%

50%

40%

30%

20%

10%

0%
5 10 15 20 25 30 35 40

Adult Management Adult Social Interaction Assignments

Being Disciplined Classroom Management Copying
Discussion Instruction/Demonstration Kinesthetic
Practice Drill Projects Reading
Social Interaction Student Uninvolved Written Assignments/Se

Note: Each row in the table is a snapshot, and each column is an activity. The
intersection of the row and column is a count of that activity during that snapshot.
This table shows the trend of activities over time.
 63

Figure A-2. Activity Changes in Ghana

100%

80%

60%

40%

20%

0%
5 10 15 20 25 30 35 40

Adult Management Adult Social Interaction Assignments

Being Disciplined Classroom Management Copying
Discussion Instruction/Demonstration Kinesthetic
Practice Drill Projects Reading
Social Interaction Student Uninvolved Written Assignments

Note: class time in Ghana lasts 30 minutes

Figure A-3. Activity Changes in Tunisia

100%

80%

60%

40%

20%

0%
5 10 15 20 25 30 35 40 45 50

Adult Management Adult Social Interaction Assignments

Being Disciplined Classroom Management Copying
Discussion Instruction/Demonstration Kinesthetic
Practice Drill Projects Reading
Social Interaction Student Uninvolved Written Assignments/Seatwork

Note: Each row in the table is a snapshot, and each column is an activity. The intersection of the
row and column is a count of that activity during that snapshot. This table shows us the trend of
activities over time
 64 Annex B

Annex B. Survey Forms and Instructions

B-1. Classroom Snapshot

This form is one snapshot. Ten snapshots are produced for each class hour.

Snapshot 1 T: Teacher A: Other Adult I: Individual

Number of Adults Present Nunber of Students Present

MATERIALS
Chalk Comput. / Manip- Visual Co-
Books Notebk None
board Calc. ulative Aids operative
ACTIVITY T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Reading A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Copying A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Assignments A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Instruction/ T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Demonstration A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Discussion A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Practice/Drill A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Written T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Assignments/ A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Seatwork I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Kinesthetic A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
Projects A 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
I 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE 1SLE
T 1SLE
Social Interaction A 1SLE
I 1SLE
Student Uninvolved I 1SLE
Being Disciplined T 1SLE
A 1SLE
Classroom T 1SLE
Management A 1SLE
I 1SLE
Adult Social Interaction TAA Adult Management T AA
 65

B-2. Basic Skills Teacher Profile – General

Values are calculated as follows:
Teacher US Criteria
Active Instruction 50% or more
Reading
Instruction/Explanation
Discussion
Drill and Practice
Kinesthetic
Projects

Passive Instruction 35% or less

Monitoring Copying
Monitoring Seatwork
Classroom Management 15% or less

Giving Assignments
Managing with Students
Disciplining Students
Managing Alone
Off-Task Socializing with Students 0%
Socializing with Others
Students
Off-Task Being Disciplined 6% or less
Socializing
Uninvolved
 66

B-3. Classroom Snapshot Activities as used in 2004-2005

Code Item Description
Academic Activities
Reading Aloud/Oral reading Students are reading aloud
One or more students are reading connected text from a text, trade book,
periodical, their own writing, or reproduced material. When reading aloud,
generally students take turns reading sections from the material. The teacher
or student may also read aloud while the rest of the class follows along in
their own texts. Although one person reads at a time, all students are coded
as engaged in the oral reading activity with the teacher. Code this over the T
using the 1, S, L, or E code. Use Books in conjunction with Reading Aloud.
If students are reading a play aloud and are working together in order to
make a presentation, code Reading Aloud in conjunction with Cooperative
Groups (i.e.: students rehearsing for a Readers' Theatre presentation).
Reading in unison is also coded here.
Copying Students are copying work or exercises from the chalkboard. The primary
purpose of the activity is to transfer the text on the board verbatim to the
students’ paper or copybooks. If students are interacting with the material to
learn or practice concepts or procedures or to apply knowledge (work math
problems, conjugate verbs), then code under Written Assignments/Seatwork.
When the teacher is copying on the blackboard and the students are copying
at their seats, code the teacher in the Classroom Management Alone. When
the teacher is monitoring while the students are copying, code the teacher
with students under Copying.
Receiving Assignments The teacher is explaining an activity, the procedures to be followed, the
amount of work to be finished, or rewards for completing the assignment.
The discussion is not focusing on the academic content, but on the
information that students need to carry out the assignment. Discussion of
grades and clarification of behavior expectations is coded here.
Instruction/Demonstration An adult or some form of media is informing some grouping of students
about a subject. Code this category also if a teacher models a procedure or
shows students how to do something (e.g., science experiment, math
problem, use of materials).
Discussion Academic discussion, verbal exchange, or slow-paced question/answer
session takes place regarding the lecture material, assignments, or problems.
This code may be used in conjunction with Cooperative Groups, to show
that a cooperative group is discussing an assignment.
Rote learning/Practice/Drill One or more students are verbally involved in reinforcing, repetitive, or rote
work. For example, learning definitions or answers to given questions, or
any text, Verbal and manipulative games which give further practice in using
a learned concept are coded here.
Written Assignments/ Seatwork One or more students are writing papers, doing computation, or are involved
in any other silent written work. If a student is writing and reading, use the
Written Assignment activity. If the teacher is monitoring, code this activity
in the T line using 1, S, L, or E. Written test taking is coded here. Please
note in the Log when test-taking occurs.
Note: Whenever the teacher is actively monitoring or listening, the activity should be coded on the T line.
Projects Projects are hands-on activities that result in a product and can extend over
one or more class sessions. Examples include constructing a bird house,
painting a mural, making puppets, making a pottery bowl, or creating a
book. Doing science experiments, agriculture, or shop also would be coded
as a type of project. The Projects category may often be coded with
Cooperative Groups when two or more people are involved in a joint
project. Please describe any project in your Log.
Note: Whenever the teacher is actively monitoring or listening, the activity should be coded on the T line.
 67

Code Item Description

Academic Activities
Off-Task Two or more students are interacting about work or subjects other than
Social Interaction class-related activities. This would include physical or negative interaction
between or among students that disrupts the class.
Uninvolved/No Activity This category is recorded when one or more students are not involved in
instructional activities. If the teacher has not specified an instructional
activity then this will be recorded by placing an E (Everyone) in the T
column under “None” in the materials column. Code students who are
arriving or departing as uninvolved. Code students who are waiting for the
teacher in order to begin an assignment or take an exam as uninvolved.
Being Disciplined One or more students are being reprimanded for their behavior or are being
sent out of the room for disciplinary reasons. This may include corporal
punishment. When corporal punishment occurs please record the incident
in the snapshot log.
Classroom Management with Adults and students are involved in classroom management: passing out
Students papers, changing activities, putting away materials, preparing to leave.
Mark No Materials. Even though materials are being handled, they are not
being used. Coded on right hand side of form.
Classroom Management Without (coded at the bottom of the form)
Students
Adult Social Interaction Two or more adults are interacting about subjects other than class-related
activities. Circle the T for the teacher or A for other adult on the left-hand
side of the form to indicate they are not working with the students.
Adult Social Interaction Two or more adults are interacting about subjects other than class-related
activities. Circle the T for the teacher or A for other adult on the left-hand
side of the form to indicate they are not working with the students.
Adult Management One or more adults are alone (without students) performing duties related
to the classroom but not directly related to any activity which is occurring
at the time of the Snapshot. Examples are a teacher correcting papers at
his/her desk, copying material onto the board, setting up materials, or
arranging books on a shelf. Circle the T for the teacher or the A for other
adult on the left hand side of the form to indicate they are not working with
students.
Note: The first ten activities reflect academic activities. The last four activities indicate students or adults who are not
involved in an academic activity. When completing the Snapshot, first record adults and students involved in academic
activities. Then record the students who are off-task or who are involved in management.
Books This category refers to printed materials that students do not write in
directly. The Book category includes trade books, textbooks, anthologies,
periodicals, students' own produced writing, and material copied from
books or periodicals.
Paper/Pencil (notebooks) This designation refers to consumable materials that students work with to
develop concepts or skills and include notebooks, workbooks, worksheets,
journals, sets or files of teacher or commercially prepared work, or blank
sheets of paper on which students work problems, write answers, or write
essays and stories. If students are copying work from the chalkboard onto
their papers, use this category.
Chalkboard Code this category when teachers or students are working at the
chalkboard located in the classroom. Also code this category when
students are using individual chalkboards at their seats.
Computer/Calculator This category includes machines that are used for instruction and require
students to employ more than one modality or computers or calculators
that provide practice, drill, or instruction. Multimedia, such as film
projectors, VCRs with monitors, CD ROM technology, and laser discs are
included. Code audio-cassette tape recorders in this category. Please
describe the machine(s) in your log.
Manipulables Code this category for hands-on materials that provide instruction or
practice. Any physical object that the student manipulates for learning is
included. Rulers, compasses, and learning materials that students use to
 68

Code Item Description

Academic Activities
construct or create something are coded as manipulables. If students are
using an abacus, code it here. Standard classroom materials such as
pencils, crayons, or erasers are not coded here. They are coded under the
Paper/Pencil category.
Visual Aids Use this category to describe visuals that teachers use to accompany
instruction and enhance student understanding. Visual aids include
information written on the overhead projector, maps, charts, photos,
posters, flipcharts, pocket charts, slides, or felt boards and other teaching
learning materials. This category does not include the chalkboard –
activities using the chalkboard should be coded under Chalkboard.
Cooperative Learning Use this category when students work on an assignment in small groups
toward a specific outcome. The assignment should have a final intended
product, such as a mural, a revised essay, a group report, a map, a science
experiment done as a group, something done in shop, etc.
No Materials/None This is coded when materials are not being used. No Materials is always
coded with the last four activities.
Coding Rules
1. Teacher can be placed only once for each snapshot.
2. Do not code the students who are working with the teacher on a separate line – they are part of the teacher
code.
3. One student or small groups of students or large groups of students can be marked more than once on the
line.
Note: subsequent modifications of this instrument have resulted in a slightly different variable list and
definitions.
 69

B-4. Calculations of Time Loss From Survey Data

Surveys asked the following comparable questions from principals, teachers, students, and in
Morocco enumerators. The averages are listed below.

School Closure

 Brazil (Pernambuco): Excluding the days that were holidays in the school calendar, how many
other days was the school closed (2004)? Principals = 2.19, Teachers = 2.70
 Ghana: How many days was the entire school closed last year (2003)? Principals = 3.21,
Teachers = 3.11
 Morocco: How many days does the school have no classes aside from the official vacation
during the current school year? During the previous school year? Principal (2004) = 0.59,
Principal (2003) = 1.13
 Tunisia: How many days wasn’t there class this year (2004)? Principal = 4.50

Teacher Absence

 Brazil (Pernambuco): How many teachers were missing today? Principals = 1.76
 Ghana: How many are absent today? Principals = 0.90 teachers
 Morocco: How many teachers are absent today? Principals = 0.60, Enumerator = 1.97 teachers
 Tunisia: How many teachers are absent today? Principals = 0.97 teachers

Teacher Delay

 Brazil (Pernambuco): How many teachers arrived late today? Principal = 1.19
 Ghana: How many teachers were late today? Principal = 0.83 teachers
 Morocco: Did you notice any teachers delay arriving to school? If yes, how many?
Enumerator = 3.43 teachers
 Tunisia: How many teachers arrive late today? Principal = 0.02 teachers

Teacher Early Dismissal

 Brazil (Pernambuco): How many teachers left early yesterday? Principal = 0.64
 Ghana: How many teachers left before normal time of closing? Number extracted from
USAid Study = 0.06 teachers
 Morocco: Did you notice any teacher’s early departure from the classroom? If yes, how
many? Enumerator = 0.68 teachers
 Tunisia: No data. Imputed from Morocco

Student Absence

 Brazil (Pernambuco): Since you began classes, how many days have you missed?
Students = 3.99 days
 Ghana: Since the beginning of the school year, how many days were you absent from
school? Students = 1.96 days
 Morocco: During the school year, how many lessons have you missed? Students = 1.59
 70

 Tunisia: Since the beginning of the year, how many times were you absent? Students =
2.93 times

Student Delay

 Brazil (Pernambuco): Since you began classes, how many days have you arrived late?
Students = 2.86 days
 Ghana: Since the beginning of the school year, how many days were you late in arriving in
school? Students = 2.23 days
 Morocco: During the current school year, how many times did you arrive late to school?
Students = 1.92 times
 Tunisia: Since the beginning of the year, how many times were you late? Students = 2.29

How Time Loss was calculated at Various Stages

The mid-point was calculated of the dates surveys were administered in each country and
respondents’ answers were projected to the end of the school year. The length of school day was
included in calculations where it was absolutely necessary to do so, since there is considerable
variation in school day lengths across schools. Finally, questions from two or more survey types were
averaged when they provided the same information. The numbers of days teachers were absent or
late were deducted from the number of days left after school closures.

Brazil (Pernambuco)

 School Closure: Average from principal and teacher (Principal average = 4.29, teacher average =
5.29 days)
 Teacher Absence: From principal (Principal = 12.76 days)
 Teacher Delays: From principal (Principal = 5.50). Teacher self reports on delays last month =
6.06 days. Assumptions for both: a 6.4 minute delay, 50-minute class period, and 5 hour school
day
 Teacher Early Dismissal: From principal (Principal = 2.30 teachers). Teacher self-reports on early
departures last month = 2.93 teachers. Assumption for both: a 5 minute delay, 50-minute class
period and 5 hour school day
 Student Absence: From student (Student = 7.82 days)
 Student Delay: From student (Student = 5.61 days)
 Researcher found by calling schools that 26 of 180 schools started 15 days late. Average time lost
was 2.175 days in the sample of 180 schools

Ghana

 School Closure: Average principal and teacher reports on closure last year (2003) (Principal =
3.22, teacher = 3.12 days)
 Teacher Absence: From principal (Principal = 43.01 teachers). Teacher self-reports on
absence last year = 13.90 and student reports on teacher absence = 10.82 days
 Teacher Delays: From principal = 39.75 days (assuming 30 minute delay and 5.5 hour school
day; both cited in Ghana’s Time on Task report)
 Teacher Early Dismissal: USAid study cites that 1.5% of teachers depart early. Using those
figures, Days lost = 2.42
 71

 Student Absence: From student (9.04 days)

 Student Delay: From student (10.610 times)

Morocco

 School Closure: Average of two principal questions on school closure (in 2004 = 1.62 and 2003 =
1.13 days)
 Teacher Absence: Average principal and enumerator question on absence in school on survey
date (Principal = 6.19, enumerator = 20.53). Note that teacher self report on absence = 11.67 and
student reports on teacher absence = 9.79 days
 Teacher Delays: From enumerator = 6.94 days.
 Teacher Early Dismissal: From enumerator There was a very low response rate to this question.
Average early departure time = 9.58 minutes
 Student Absence: From student (4.30). Parents report = 26.235 days from one-week averages that
was not used.
 Student Delays: From student (5.19 times)
 Data are available from the enumerator survey on teacher delay entering classroom, with an
average delay of 7.83 minutes on average, resulting in losses of 4.88 days. Teacher delay entering
classroom after break time = 3.184 days with average delay of 6.07 minutes

Tunisia

 School Closure: from principal (Principal = 5.15 days). Data file had errors. Average
estimated from reasons Tunisian schools were closed.
 Teacher Absence: from Principal (Principal = 11.55). Teacher self-reports on absence last
year = 11.28 and student reports = 11.28 days
 Teacher Delays: from principal (Principal = 1.27 days). Assumed 15 minute delay, 50-minute
class periods, and a 5 hour school day
 Teacher Early Dismissal: No data. Imputed by projecting ratio of delay to departure in
Morocco
 Student Absence: From students (3.35 days). Parents reported 7.65 days
 Student Delay: From students (2.63 times since the beginning of school year)
 72 Annex C

Annex C. Policy Discussion Exercise

Improving service delivery

Helping governments get what they pay for
Consider the country you are working on, look at the schematic and answer the questions below.
- How important are these factors in reducing schooling time as far as you know?
- What policy dialogue or public debate has taken place already?
- What aspects of policy dialogue were more and less successful?
- What decisions and activities are needed to deal with the factors reducing schooling time?
- Who is responsible?
- What expenditures will be needed to increase instructional time?
- What incentives might work at each level of time loss?
- Who might act as a champion?
Class time as allotted by a government: _________ hours or days
Remaining after school closures
Strikes
Weather
Building problems
Teacher training, study leaves
Extra holidays
Civil unrest-security
Other
Remaining after teacher absenteeism
Tardiness
Early departures
Longer breaks
Important illness issues? (HIV)
Other
Student absenteeism
Widespread illnesses
Need for work
Limited instruction-little need to attend
Other
Class time devoted to any learning task
Do supervisors/inspectors function?
Do principals supervise learning?
Are there textbooks for all?
Other issue?
Learning time relevant to
curriculum
Class up to date with
curriculum?
Students know prerequisites?
Other issue?
 73 References

REFERENCES
Abadzi, H. 2006. Efficient Learning for the Poor: Insights from the Frontier of Cognitive Neuroscience.
Washington, DC: World Bank.
Álvarez, Jesús, Vicente García Moreno, and Harry Anthony Patrinos. 2006. Institutional effects as determinants
of Learning outcomes: Exploring state. Variations in Mexico. Washington DC: World Bank.
Amadio, M. 1996. “Primary School Repetition: A Global Perspective.” Geneva: IBE-UNESCO, International
Bureau of Education-UNESCO.
Amadio, M. 1997. “Primary Education: Length of Studies and Instructional Time.” Educational Innovation and
Information, 92, Pp.2-7.
Amadio, M. 1998. Instructional time and teaching subjects during the first four years of primary education.
Educational Innovation and Information. No. 96. Geneva: IBE.
Ainsworth, M. et al. 2002. The Impact of Adult Mortality on Primary School Enrollment in Northwestern
Tanzania. Africa Region Human Development Working Papers Series. Washington, DC: World Bank.
Anderson, L.W. 1976. “An Empirical Investigation of Individual Differences in Time to Learn,” Journal of
Educational Psychology, 68, 2, p. 226-233.
Anderson, L.W and R.B Burns. 1989a. Research in Classrooms: The Study of Teachers, Teaching, and
Instruction. Oxford, England: Pergamon Press.
Anderson, L.W., D.W Ryan, and B.J. Shapiro. (eds) 1989b. The IEA Classroom Environment Study. Oxford:
Pergamon Press.
Ali, M. and T. Reed. 1994. A School and Parental Survey of Book Provision Issues in NWFP, International
Book Development, Ltd.
Arancibia, Violeta. 1987 Manejo instruccional del profesor en la sala de clase. Harvard Institute for
International Development, BRIDGES causual paper.
Atkinson, R. C.1972. Ingredients for a theory of instruction. American Psychologist, 27, 921-931.
Attar, M. 2001. “Ghana Country Report,” In: Pillai, S. (Ed.) Strategies for Introducing New Curricula in West
Africa, (final report of the seminar held in Lagos, Nigeria, 12-16 November 2001) Geneva: IBE-
UNESCO.
Baker, V. 1988. “Schooling and disadvantage in Sri Lankan and other rural situations.” Comparative Education,
24, 3, Pp. 377-388.
Baker, D., B. Goesling and G. LeTendre. 2002. “Socioeconomic Status, School Quality and National Economic
Development.” Comparative Education Review 46, 3, Pp. 291-312.
Barrier Émilie, Sekou Fernández, Jeannot Saa Tinguiano, Gononan Traore. 1998. Évaluation du système
éducatif guinéen. Centre international d’études pédagogiques – Sèvre et Cellule nationale de
coordination de l’évaluation du système éducatif – Conakry.
Benavot, A. 2002. "Educational Globalisation and the Allocation of Instructional Time in National Education
Systems." Report submitted to the International Bureau of Education (Geneva) as a contribution to
UNESCO's Education for All Global Monitoring Report 2002.
Benavot, Aaron and Massimo Amadio. 2004. A Global Study of Intended Instructional Time and Official
School Curricula, 1980-2000. Background paper commissioned by the International Bureau of
Education for the UNESCO- EFA Global Monitoring Report (2005): The Quality Imperative. Geneva:
IBE.
 74

Benavot, Aaron. 2006. The Diversification of Secondary Education: School Curricula in Comparative
Perspective. UNESCO: IBE Working Papers on Curricular Issues No. 6.
http://www.ibe.unesco.org/organization/index_archive/index_archive_eng.htm
Bennell, P. Hyde, K. and N. Swainson. 2002. The Impact of HIV/AIDS Epidemic on the Education Sector in
Sub-Saharan Africa: A Synthesis of the Findings and Recommendations of Three Country Studies.
Center for International Education: University of Sussex: Institute of Education.
Ben Jaafar, S. 2006. An Alternative Approach to Measuring Opportunity-to-Learn in High School Classes.
Alberta Journal of Educational Research 52(1).
Ben-Peretz, M. and R, Bromme. 1990. The Nature of Time in Schools: Theoretical Concepts, Practitioner
Perceptions. New York and London: Teachers College Press; Teachers College, Columbia University.
Berliner, D.C., and C. W. Fisher. 1985 “One More Time.” In: Fisher, C. W. and D.C. Berliner, (eds.)
Perspectives on Instructional time, New York: Longman, Pp. 333-347.
Berliner, D.C. 1988. Simple views of effective teaching and a simple theory of classroom instruction. In D.
Berliner and B. Rosenshine (eds.) Talks to the teachers, New York: Random House, Pp. 93-100.
Bloom, B. 1971. “Handbook on Formative and Summative Evaluation of Student Learning,” New York:
McGraw Hill.
Bloom, B. 1974. “Time and Learning,” American Psychologist, 29, Pp. 682-688.
Bobonis, G., Miguel, E. and C. Sharma 2004. “Iron Deficiency Anemia and School Participation,” Mimeo,
University of California, Berkely.
Bjork, R A., and E.L. Bjork. 1992. “A New Theory of Disuse and an Old Theory of Stimulus Fluctuation.” In
A. Healey, S. Kosslyn, and R. Shiffrin (eds.), From Learning Processes to Cognitive Processes:
Essays in Honor of William K. Estes. Vol. 2, pp. 35–67. Hillsdale, NJ: Erlbaum
Bray, M. 2000. Double – Shift Schooling: Design and Operation for Cost-Effectiveness. 2nd Edition, London:
The Commonwealth Secretariat and IIEP-UNESCO.
Bray, M. 2006. Private supplementary tutoring: comparative perspectives on patterns and implications.
Compare, 36(4): 515–530.
Bridges E. and M. T. Hallinan.1978. Subunit Size, Work System Interdependence, and Employee Absenteeism,
Educational Administration Quarterly, 14: 2, pp 24-42.
Bright, G.W. 1988. “Time-on-Task in Computer and Non-Computer Estimation Games,” Journal of Computers
in Mathematics and Science Teaching, 7, 4, pp. 41-46.
Brophy J. & T. Good. 1986. Teacher behavior and student achievement. In M.C. Witrock (ed.) Handbook of
research on teaching (Third ed.) New York: Macmillan.
Buchmann C. and E. Hannum. 2001. “Education and stratification in developing countries: A review of theories
and research,” Annual Review of Sociology, 27, pp 77-103.
Carnoy, M., A. Gove, and J. Marshall. 2004. “Why Do Students Achieve More in Some Countries than in
Others? A Comparative Study of Brazil, Chile, and Cuba.” Paper presented in October 2004 at the
World Congress of Comparative Education in Havana, Cuba.
Carroll, J. 1963. “A model of school learning,” Teachers College Record, 64, Pp. 723-733.
Cerdan-Infantes, Pedro and Chrsteel Vermeersch. 2007. More Time is Better: An evaluation of the full-time
school program in Uruguay. Washington DC: World Bank, Policy Research Working Paper No. 4167.
Chaudhury, N. and J. Hammer 2003. Ghost doctors: Absenteeism in Bangladeshi health facilities. Policy
Research Working Paper 3065. Development Research Group, Public Services. Washington, DC: The
World Bank.
 75

Chaudhury, N., J. Hammer, M. Kremer, K. Mularidharan, and H. Rogers. 2004a. Roll Call: Teacher Absence in
Bangladesh. World Bank: DEC.
Chaudhury, N., J. Hammer, M. Kremer, K. Mularidharan, and H. Rogers. 2004b. Teacher Absence in India.
(draft) World Bank: DEC.
Chaudhury, N., J. Hammer, M. Kremer, K. Mularidharan, and H. Rogers. 2004c. Teacher and Health Care
Provider Absence: A Multi-Country Study. World Bank: DEC.
Coleman, J. S., E. Q. Campbell, C. J. Hobson, J. McPartland, A. Mood, F. Weinfeld, and R. L. York. 1966.
Equality of Educational Opportunity. Washington, DC: Office of Education, National Center for
Education Statistics, GPO.
Connelly, F.M. and D.J. Cladinin. 1993. “Cycles, Rhythms and the Meaning of School Time,” In: Anderson,
L.W. and H.J. Walberg, (eds.) Time Piece: Extending and Enhancing Learning Time, Virginia:
National Association of Secondary School Principals.
Crouch L, H Abadzi, M Echegaray, C Pasco, J Sampe 2005. “Monitoring basic skills acquisition through rapid
learning assessments: A case study from Perú.” Prospects, 35, 2, Pp. 137 - 156
Das, J., S. Dercon, J. Habyarimana, and P. Krishnan 2005. Teacher shocks and student learning: Evidence from
Zambia. World Bank Policy Research Working Paper
Das, Jishnu and Jeffrey Hammer. 2006. Money for Nothing: The Dire Straits of Medical Practice in Delhi,
India. World Bank Policy Research Working Paper 3669.
Das, Jishnu and Jeffrey Hammer. 2004. Strained Mercy: The Quality of Medical Care in Delhi. World Bank
Policy Research Working Paper 3228.
Daun, H. 2000. “Primary Education in Sub-Saharan Africa – A moral Issue, an Economic Matter or Both?”
Comparative Education, 36, 1, Pp. 37-53.
De, Anuradha and Jean Dreze. 1999. Public Report on Basic Education in India (PROBE). New Delhi: Oxford
University Press.
Dia, E.C. 2003. “Instructional Time in Primary School: The Cases of Burkina Faso and the Gambia,” Paper
prepared for HDNED. Washington DC: World Bank.
Diouf, Adama. 2005. Temps scolaire et qualité des apprentissages dans les collèges et lycées. Diourbel,
Sénégal (unpublished study).
Downey, C.J., B.E. Steffy, F.W. English, L.E. Frase, and W.K. Poston. 2004. The Three-Minute Classroom
Walk-Through. Thousand Oaks, CA: Corwin Press.
Doyle, W. 1983. “Academic work.” Review of Educational Research, 53, Pp. 159-200.
Duflo, E. and R. Hanna 2005. “Monitoring Works: Getting Teachers to Come to School.” MIMEO, MIT.
EARC. 2003. Teacher Time on Task. United States Agency for International Development, Grant No. 641-G-
00-03-0055 unpublished report. Educational Assessment and Research Center, 35 Dabebu Rd. Osu,
Ghana.
EDUCA. 2005. “Uso del Tiempo en la Escuela Dominicana.” Encuesta EDUCA-GALLUP. (financed by
USAID). Santo Domingo, República Dominicana.
Ehrenberg, R. G., D. I. Rees, and E. L. Ehrenberg. 1991. "School District Leave Policies, Teacher Absenteeism,
and Student Achievement." Journal of Human Resources, 26, 1, Pp. 72-105.
El-Sanabiy, N. 1989. Determinants of Women's Education in the Middle East and North Africa: Illustrations
from Seven Countries. Washington, DC: World Bank.
Fairhurst, G, W. Gibbs, P. Jain, D. Khatete, G. Knamiller, G. Welford, and P. Wiegand 1999. The Effectiveness
of Teacher Resource Centre Strategy. United Kingdom: Department for International Development.
 76

Filmer D. and L. Pritchett. 1999. “The Effect of Household Wealth on Educational Attainment: Evidence from
35 countries.” Population and Development Review 25, Pp. 85-120.
Fisher, C.W., Filby, N.N., Marliave, R.S., Cahen, L.W., Dishaw, M.M., Moore, J.E. and Berliner, D.C. 1978.
Teaching Behaviors. Academic Learning Time and Student Achievement. Technical Report V-1, Final
Report of Phase III-B, Beginning Teacher Evaluation Study, San Francisco: Far West Laboratory for
Educational Research and Development.
Fisher, C. W. and Berliner, D.C. (eds.) 1985. Perspectives on Instructional Time. New York: Longman.
Fredrick, W. and Walberg, H. 1980. “Learning as a Function of Time,” Journal of Educational Research. 73, 4,
Pp. 183-194.
Fuller, B. 1987. “What School Factors Raise Achievement in the Third World?” Review of Educational
Research. 57, Pp. 255-292.
Fuller, B. and Clarke, P. 1994. “Raising School Effects While Ignoring Culture? Local Conditions and the
Influence of Classroom Tools, Rules and Pedagogy.” Review of Educational Research. 64, Pp. 119-
157.
Fuller, B. L. Dellagnelo, A. Strath, et al. 1999. “How to Raise Children’s Early Literacy? The Influence of
Family, Teacher, and Classroom in Northeast Brazil,” Comparative Education Review. 43, Pp. 1-35.
Fusaro, J.A. 1997. “The Effect of Full Day Kindergarten on Student Achievement: A Meta-analysis.” Child
Study Journal, 27, 4, Pp. 269-277.
Gemechu, D. T. 2001. Ethiopia Country Report (In: Final report of the seminar held in Nairobi, Kenya), 25-29
June 2001, Geneva: IBE-UNESCO.
Gettinger, M. 1984. “Individual differences in time needed for learning: A Review of the Literature.”
Educational Psychologist. 19, Pp. 15-29.
Glewwe, P., Jacoby, H.G., E.M. King 2001. “Early Childhood Nutrition and Academic Achievement: A
Longitudinal Analysis.” Journal of Public Economics, 81 pp. 345-368.
Glewwe, P. and M. Kremer. 2005. “Schools, Teachers, and Education Outcomes in Developing Countries”
Second draft of chapter for Handbook on the Economics of Education.
Gomes-Neto, J.B., E. Hanushek, R.H Leite, and R.C. Frota-Bezzera. 1997. “Health and Schooling: Evidence
and Policy Implications for Developing Countries.” Economics of Education Review, 16, 3, pp. 271-
282.
Goodlad, J.L. 1980. A place called school. New York, NY: McGraw Hill
Gore, M.C and J. F. Dowd. 1999. Taming the Time Stealers: Tricks of the Trade from Organized Teachers.
Thousand Oaks, CA: Corwin Press.
Gould, B. and U. Huber. 2003. The social demand for schooling in HIV/AIDS affected populations in Tanzania:
Summary results from a field survey. Draft paper for presentation at the Project Results Dissemination
Workshops, University of Dar es Salaam.
Greenwood, C. 1991. “Longitudinal Analysis of Time, Engagement and Achievement in At-Risk Versus Non-
Risk Students.” Exceptional Children. 57, 6, Pp. 521-535.
Grossman, M. 1973. The Correlation between Health and Schooling. NBER Working Paper No. 22. New York,
NY: National Bureau of Economic Research.
Hanushek, E.A. and V. Lavy. 1994. School Quality, Achievement Bias and Dropout Behavior in Egypt. LSMS
Working Paper Number 107. Washington D.C.: World Bank.
Haut Commissariat au Plan. 2004. Carte de La pauvreté communale. Royaume du Maroc
Heyneman, S. and W. Loxley. 1983. “The Effect of Primary School Quality on Academic Achievement across
Twenty-Nine High and Low Income Countries.” American Journal of Sociology, 88, 6, Pp. 1162-1194.
 77

Hunter, S. and J. Williamson. 2000. Children on the Brink: Executive Summary, USAID, Washington, D.C.
Huyvaert, S. H. 1998. Time is of the Essence: Learning in Schools, Needham Heights, MA: Allyn & Bacon.
Independent Evaluation Group (IEG). 2006a. Argentina: First Secondary Education Project (Loan 3794);
Decentralization and Improvement of Secondary Education and Polimodal Education Development
Project (Second Secondary Education Project; Loan 3971); Third Secondary Education Project (Loan
4313). Project Performance Assessment Report. (Draft) Washington DC: World Bank.
Independent Evaluation Group (IEG). 2006b. From Schooling Access to Learning Outcomes: An Unfinished
Agenda. Independent Evaluation Group, World Bank. An Evaluation of World Bank Support to
Primary Education. Washington, DC: World Bank.
Independent Evaluation Group (IEG). 2006c. Indonesia: East Java and East Nusa Tenggara Junior Secondary
Education Project (Loan 4042); Central Indonesia Junior Secondary Education Project (Loan 4062);
Sumatra Junior Secondary Education Project (Loan 4095). Project Performance Assessment Report.
Washington DC: World Bank, Independent Evaluation Group Report 36427-ID. Washington, DC:
World Bank.
International Bureau of Education. 2002. “A New Challenge: Scaling Up the Educational Response to
HIV/AIDS.” Educational Innovation and Information, 110, Geneva.
Jacobson, S. L. 1991. “Attendance Incentives and Teacher Absenteeism.” Planning and Changing. 21, 2, Pp. 78-
93.
Jamison, D. and M. Lockheed. 1987. “Participation in Schooling: Determinants and Learning Outcomes.”
Economic Development and Cultural Change. 35 2, Pp. 279-306.
Jencks, C. 1972. Inequality: A Reassessment of the Effects of Family and Schooling in America. New York:
Basic Books.
Karweit, N. 1976. “A Reanalysis of the Effect of Quantity of Schooling on Achievement.” Sociology of
Education. 49, Pp. 236-246.
Karweit, N. 1983. Time on Task: A Research Review. Baltimore, MD: The John Hopkins University, Center for
Social Organization of Schools.
Karweit, N. 1985. “Should We Lengthen the School Term?” Educational Researcher. 14, Pp. 9-15.
Karweit, N. 1988. “Time on Task: The Second Time Around” NASSP Bulletin, 72, 505, Pp. 31-39.
Kelly, A.E. 2003. “A Report on the Literacy Network and Numeracy Network Deliberations.” OECD Seminar,
Brockton, MA, 29–31. January.
Kelly, M.J. 2000. “The Impact of HIV/AIDS on the Rights of the Child to Education.” Paper Presented at
SADC-EU Seminar on The Rights of the Child in a World with HIV and AIDS, Harare, October 23.
Kelly, M.J. and T. Comb 2002. Perspectives in Education. Pretoria: South Africa.
Kim, Kyeh. 1999. Comparative Study of Instructional Hours in West Africa. Informal report, AFTHR.
Washington, DC: World Bank:
Kim, K. 1999. “Status Update on Comparative Study of International Hours in West Africa.” Washington, DC:
World Bank.
King, E. M., P. F. Orazem, and E. M. Paterno. 1999. Promotion With and Without Learning: Effects on Student
Dropout. Washington, DC: World Bank.
King, E. M. and B. Ozler. 2001. What's Decentralization Got To Do With Learning? Endogenous School
Quality and Student Performance in Nicaragua. Washington DC: World Bank.
Knight, Stephanie L. 2006. “History and Use of the Stallings Observation System in the World Bank
International Time on Task Project.” Paper presented at the American Educational Research
Association Conference, San Francisco. April.
 78

Knight, Stephanie. 2001. Using Technology to Update Traditional Classroom Observation Instruments. Paper
presented at the American Educational Research Association Conference.
Kremer, M., S. Moulin, D. Myatt, and R. Namunyu. 1997. The Quality-Quantity Tradeoff in Education:
Evidence from a Prospective Evaluation in Kenya. Washington DC: World Bank.
Kremer, M., K. Muralidharan, N. Chaudhury, J. Hammer, and H. Rogers. 2005. “Teacher absence in India.”
Journal of the European Economic Association, 3, 2-3, Pp. 658-667.
Kurz, T.B. 2003. “Using Systematic Classroom Observation in the Current Era of Accountability: Lessons from
the Past.” Paper Presented to American Educational Research Association. Chicago, IL April 22.
Lavy, Victor. 2001. Estimating the Effect of School Time of Instruction on Student Achievements. Hebrew
University Economics Working Paper No. 01-4 Social Science Research Network Electronic Library.
Leonard, Lawrence J. 1999. “Towards Maximizing Instructional Time: The Nature and Extent of Externally-
Imposed Classroom Interruptions." Journal of School Leadership 9, 5, September. Pp 454-474.
Leonard, Kenneth L., Melkiory C. Masatu , and Alex Vialou. 2005. Getting doctors to do their best: Ability,
Altruism and Incentives. (NSF Grant 00-95235). Washington DC: World Bank.
Leslie, J. and Jamison, D.T. 1990. Health and Nutrition Considerations in Education Planning. Food and
Nutrition Bulletin, 12, 191-204.
Lewis, M. and Lockheed, M. 2006. Inexcusable Absence: Why 60 Million Girls Still Aren’t in School and What
to Do About it. Washington DC: Center for Global Development.
Lezotte, L. W., & Bancroft, B. A. 1985. School improvement based on Effective Schools research: A promising
approach for economically disadvantaged and minority students. The Journal of Negro Education,
54(3): 301-312.
Linden, T. 2001. Double-Shifts Secondary School: Possibilities and Issues. Secondary Education Series,
Washington DC: World Bank.
Lockheed, M., B. Fuller, and R. Nyirongo. 1989. “Family Effects on Students’ Achievement in Thailand and
Malawi.” Sociology of Education, 62, Pp. 239-256.
Lockheed, M and A.M. Verspoor. 1992. Improving Primary Education in Developing Countries. New York:
Oxford University Press.
Lloyd, C. B., S. El Tawila, W. Clark, and B. Mensch. 2003. The impact of educational quality on school exit
in Egypt. Comparative Education Review, 47 (4):444.67.
Marshall, J.H. 2003. “School Attendance and Academic Achievement: A Simultaneous Equations Approach.”
Stanford University Working Paper.
Mason, A.D. and S. Khandker. 1997. Household Schooling Decisions in Tanzania, Poverty and Social Policy
Department, The World Bank, Washington D.C.
Matsagouras, Elias. 1987. Time is also learning (Ο χρόνος είνaι κaι ... μάθηση). Contemporary Education
(Σύγχρονη Εκπaίδευση), 37: 38-44.
Millot, B. 1994. Macroeconomics of Educational Time and Learning, International Encyclopedia of Education,
6(2), 3545-3550.
Millot, B. and J. Lane 2002. “The Efficient Use of Time in Education.” Education Economics, 10, Pp. 209-228.
Muñoz Izquierdo, C. 1979. El síndrome del atraso escolar y el abandono del sistema educativo. Revista
Latinoamericana de Estudios Educativos, 9.
Murillo, M.V., M. Tommasi, L. Ronconi and J. Sanguinetti. 2002. The Economic Effects of Unions in Latin
America: Teachers’ Unions and Education in Argentina. Inter-American Development Bank, Latin
American Research Network Working Paper R-463.
 79

Njie, F. 2001. “Gambia Country Report.” In: Pillai, S. (ed.) Strategies for Introducing New Curricula in West
Africa. (Final report of the seminar held in Lagos, Nigeria) Geneva: IBE-UNESCO.
Onwu, G. 1999. “Inquiring into the Concept of Large Classes: Emerging Topologies in an African Context”, In:
Naidoo, P. and Savage, M. (eds.) Using the Local Resources Base to Teach Science and Technology.
Westville, South Africa, University of Durban, African Forum for Children's Literacy in Science and
Technology (AFCLIST) publication.
Operations Evaluation Department (OED). 2001. Morocco: Country Assistance Evaluation. Washington DC:
The World Bank. Report no. 22212. Washington, DC: World Bank.
Operations Evaluation Department (OED). 2004. Books, Buildings, and Learning Outcomes: An Impact
Evaluation of World Bank Support to Basic Education in Ghana. Washington, DC: World Bank.
Operations Evaluation Department (OED). 2005. Niger: First Education Project (Credit 1151-NIR), Primary
Education Development Project (Credit 1740-NIR), Basic Education Sector Project (hybrid) – (credit
2618-NIR). Project Performance Assessment Report. Washington, DC: World Bank: Independent
Evaluation Group.
Operations Evaluation Department (OED). 2005a. “Uruguay. Vocational Training and Technological
Development Project (L1594-UR), Basic Education Quality Improvement Project (Loan 3729-UR),
Second Basic Education Quality Improvement Project (Loan 4381-UR).” Project Performance
Assessment Report. Washington, DC: World Bank: Independent Evaluation Group.
Osborne, Jason W. (2000). Advantages of hierarchical linear modeling. Practical Assessment, Research &
Evaluation, 7(1). http://PAREonline.net/getvn.asp?v=7&n=1.
Perie, M., D. Baker and S. Bobbitt. 1997. Time Spent Teaching Core Academic Subjects in Elementary
Schools. (NCES 97-293). Washington DC: US Department of Education, National Center for
Educational Statistics.
Pianta, Robert C., Jay Belsky, Renate Houts, and Fred Morrison. 2007. Opportunities to Learn in America’s
Elementary Classrooms. Science, 315, 1795-1796.
Pitkoff, E. 1993. "Teacher Absenteeism: What Administrators Can Do." NASSP Bulletin, 77, 551, Pp. 39-45.
Quartarola, B. 1984. A Research paper on Time on Task and the Extended School Day/Year and their
Relationship to Improving Student Achievement, Sacramento, CA: Research, Evaluation, and
Accreditation Committee, Association of California Administrators. (ERIC Document Reproduction
Service No. ED 016 890).
Rabenstein, R.L., N.A. Addy, B.J. Caldarone, Y. Asaka, L.M. Gruenbaum, L.L. Peters, D.M. Gilligan, R.M.
Fitzsimonds, and M.R. Picciotto. 2005. “Impaired Synaptic Plasticity and Learning in Mice Lacking
Adducing, an Acting-Regulating Protein.” The Journal of Neuroscience 25: 2138–2145
Rao, G. V. L. N. 1999. Teachers Absenteeism in Primary School: A Field Study in Select Districts of Madhya
Pradesh and Uttar Pradesh, New Delhi: District Primary Education Programme.
Reimers, F. 1993. “Time and Opportunity to Learn in Pakistan's Schools: Some Lessons on the Links between
Research and Policy.” Comparative Education, 29, 2, Pp. 201-12.
Reynolds, A. 1992. “What is competent beginning teaching? A review of the literature.”Review of Educational
Research, 62, 1-35.
Rogers, H. J. Lopez-Calix, N. Chaudhury, J. Hammer, N. Córdoba, M. Kremer, and K. Muralidharan. 2004.
“Teacher Absence and Incentives in Primary Education: Results from a National Teacher Tracking
Survey in Ecuador.” In: Ecuador: Creating Fiscal Space for Poverty Reduction: A Fiscal Management
and Public Expenditure Review. 2, Pp. 136-162.
OED 2001. Morocco: Country Assistance Evaluation. Report no. 22212. Washington DC: The World Bank.
 80

OED. 2002a. “Brazil. Innovations in Basic Education Project (Loan 3375–BR); Second Northeast Basic
Education Project (Loan 3604–BR); Third Northeast Basic Education Project (Loan 3663–BR); School
Improvement Project—FUNDESCOLA I (Loan 4311–BR).” Project Performance Assessment Report.
Washington, DC: World Bank, Independent Evaluation Group.
OED. 2004. Books, Buildings, and Learning Outcomes: An Impact Evaluation of World Bank Support to Basic
Education in Ghana.
Romberg, T.A. 1980. “Salient Features of the BTES Framework of Teacher Behavior.” In: C. C. Denham and
A. Liebermen (eds.) Time to Learn, Washington DC: U.S. Department of Education, Pp. 73-93.
Rosenshine, B. V. 1979. “Content, time and direct instruction.” In Peterson and H. Walberg (eds.) Research on
Teaching: Concepts, findings and Implications. Berkeley, CA: McCutchean, Pp. 28-56.
Roth, J., J. Brooks-Gunn, M. Linver and S. Hofferth. 2003. “What happens during the school day?: Time diaries
from a national sample of elementary school teachers.” TCRecord.org. on-line article #11018.
(http://www.tcrecord.org/PrintContent.asp?ContentID=11018).
Sathar, Z. A., C. B. Lloyd, M. ul Haque, M. J. Grant, and M. Khan. 2005. Fewer and better children :
Expanding choices and school fertility in rural Pakistan. Population Council, Policy Research
Division, New York.
Schaffer, E., Stringfield, S., and Wolfe, D. 1992. An Innovative Beginning Teacher Induction Program: A Two-
Year Analysis of Classroom Interactions. Journal of Teacher Education, 43, 3, pp.181-192.
Schaffer, Eugene C., Kenny Khoo, Danielle Switzer and Kathy Neumann. 2006. “Results of the World Bank
Surveys Investigating the Quality of Education in Tunisia, Morocco, Ghana, and Brazil.” Paper
presented at the Annual Meeting of the American Educational Research Association, San Francisco,
California, 2006. (session 71.05)
Schenker, I. 2001. “New Challenges for school AIDS education within an evolving HIV pandemic.”
Prospects.,Vol. XXX: 3, pp 415-434
Schenker I., and J. Nyrenda. 2002. “Preventing HIV/AIDS in Schools.” Educational Practice Series-9.
International Academy of Education and the International Bureau of Education. Geneva.
Schenker, I. 2003. AIDS. History of Childhood Encyclopedia, NY: McMillan References.
Sen, A. 2002. Pratichi Education Report: An Introduction. (www.amartyasen.net/pratichi.htm)
Smith, B. 2000. “Quantity matters: Annual instructional time in an urban school system.” Educational
Administration Quarterly, 36, 5, Pp. 652-82.
Smyth, J. 1985. “A context for the study of time and Instruction.” In: Fisher, C. W. and Berliner, D.C. (eds.)
1985. Perspectives on Instructional time. New York: Longman, Pp. 3-27.
Stallings, J. 1975. “Implementation and child effects of teaching practices in follow through classrooms.”
Monograph of the Society for Research in Child Development. 40, Serial # 163.
Stallings, J. 1980. “Allocated Academic Learning Time Revisited, or Beyond Time on Task.” Educational
Researcher, Vol. 9, No. 11, December, pp. 11-16. (Summarized in Effective Schools Research
Abstracts Volume 1, Issue 6)
Stallings, J. 1985a. Instructional time and staff development. In C. W. Fisher and D. C. Berliner (Eds.),
Perspectives on instructional time (pp. 283-98). New York and London: Longman.
Stallings, J. 1985b. “Implications from the research on teaching for teacher preparation.” In I. E. Housego & P.
P. Grinunett (eds.), Teaching and teacher education: Generating and utilizing valid knowledge for
professional socialization (pp 161-179). Vancouver: The University of British Columbia, Centre for
the Study of Teacher Education, Faculty of Education.
Stallings, J. 1986. Under what conditions do children thrive in the Madelaine Hunter Model? A report of Project
Follow Through, Napa, California. ERIC AN ED265141.
 81

Stallings, J., & D. Kaskowitz. 1974. Follow through classroom observation evaluation, 1972-1973 (SRI Project
URU- 7370). Menlo Park, CA: Stanford Research Institute.
Stallings, J., M. Needles, & Stayrook N. 1979. How to change the process of teaching basic reading skills, in
secondary schools. Final report to the National Institute of Education. Menlo Park, CA: SRI
International.
Stallings, Jane A. 2006. “Results of Classroom Observations in Tunisia, Morocco, Ghana, and Brazil.” Paper
presented at the Annual Meeting of the American Educational Research Association, San Francisco,
California, 2006. (session 71.05)
Stevens, F. E. 1993. “Applying an Opportunity-to-Learn Conceptual Framework to the Investigation of the
Effects of Teaching Practices via Secondary Analyses of Multiple-Case-Study Summary Data.” The
Journal of Negro Education, 62, 3, Pp. 232–248.
Stallings, J., and I. Freiberg. 1991. “Observation for the improvement of teaching.” In H. Waxman & H.
Walberg (eds.), Effective teaching: Current research (pp. 107-133). Berkeley: McCutchan.
Stringfield, S., Schaffer, E., & Devlin-Scherer, R. 1986. A study of the generalizability of teacher change quasi-
experiments. Journal of Classroom Interaction, 21,9-15.
Stringfield, S., C. Teddlie, Wimpelberg, R., & P. Kirby. 1990, January. “Design and first analyses from a 5-year
follow-up of more and less effective schools in the Louisiana School Effectiveness Study. Paper
presented at the International Congress for School Effectiveness, Jerusalem.
Subbaro, K., A. Mattimore, and K. Plangemann. 2001. Social Protection of Africa’s Orphans and Other
Vulnerable Children, Africa Region Human Development Working Papers Series. August.
Washington, DC: World Bank.
Suryadarma, D., A. Suryahadi., S. Sumarto, and H. Rogers. 2004. The Determinants of Student Performance in
Indonesian Public Primary Schools: The Role of Teachers and Schools. Washington, DC and Jakarta,
Indonesia: World Bank and SMERU.
Tan, J.P., J. Lane, and P. Coustère. 1997. “Putting Inputs to Work in Elementary Schools: What Can Be Done in
the Philippines.” Economic Development and Cultural Change. 45, 4, Pp. 857-879.
Tanzania Development Research Group (TADREG). 1993. Parents Attitudes and Strategies Towards Education
in Rural Tanzania. Tanzania.
Taylor, D. L., C. Teddlie, J. Freeman, and M. Pounders. “A Child's Day at School: Variations in More and Less
Effective Schools in Low- And Middle-Socioeconomic Status Contexts.” Journal of Education for
Students Placed At Risk, 3, 2 1998): 115-132.
Teddlie, C., Kirby, P., & Stringfield, S. 1989. Effective versus ineffective schools: Observable differences in the
classroom. American Journal of Education, 97, 3, pp. 221-236
Teddlie, C & Reynolds, D, 2000. The International Handbook of School Effectiveness Research, London and
New York, Falmer Press.
Tietjen, K., A. Rahman, and S. Spaulding. 2004. “Time to Learn: Teachers’ and Students’ Use of Time in
Government Primary Schools in Bangladesh.” Basic Education and Policy Support (BEPS) Activity
Creative Associates International, Inc. (USAID).
UNAIDS website: http://www.unaids.org
UNESCO Institute of Statistics website: http://www.uis.unesco.org/.
UNESCO. 1994. World Education Report, Paris: UNESCO.
UNESCO. 1998. World Education Report, Paris: UNESCO.
UNESCO. 2000. World Education Report, Paris: UNESCO.
 82

UNESCO-IBE. 2000. World Data on Education: A Guide to the Structure of National Education Systems. Paris:
UNESCO.
UNESCO. 2001. UNESCO’s Strategy for HIV/AIDS Preventative Education, Paris: IIEP.
UNESCO. 2002. Education for All: Is the World On Track? Paris: UNESCO.
UNESCO. 2003. EFA Global Monitoring Report, 2004: Gender and Education for All, The Leap to Equality,
Paris: UNESCO.
UNESCO. EFA Global Monitoring Report 2005: Education For All: The Quality Imperative, Paris: UNESCO
UNESCO-IIEP website: http://www.unesco.org/iiep/
UNICEF. 1996. The State of the World’s Children, Oxford, UK: Oxford University Press.
Usman, S., Ahmadi, and Daniel Suryadarma. 2004. When Teachers are Absent: Where do They Go and What is
the Impact on Students? Jakarta: SMERU Research Institute.
Vermeersch, C. and M. Kremer 2004. School Meals, Educational Achievement and School Competition:
Evidence from a Randomized Evaluation. Policy Research Working Paper no. 3523, Washington DC:
World Bank.
Verwimp, P. 1999. “Measuring the Quality of Education at Two Levels: A Case Study of Primary Schools in
Rural Ethiopia.” International Review of Education, 45, 2, Pp. 167-196.
Wachter, J. C. and Carhart, C. 2003. Time-Saving Tips for Teachers. Thousand Oaks, CA: Corwin Press.
Walberg, H. J. 1984. “Improving the Productivity of America’s School.” Educational Leadership, 8, Pp. 19-27.
Walberg, H. J. 1988. “Synthesis of Research on Time and Learning.” Educational Leadership, 45 (6), Pp. 76-85.
Walberg, H. J. and Fredrick, W.C. 1991. Extending Learning Time, Washington, DC: U.S. Department of
Education.
Wang, J. 1998. “Opportunity to Learn: The impacts and policy implications.” Educational Evaluation and
Policy Analysis, 20, 3, Pp. 137-156.
Watkins, Kevin. 2000. The Oxfam Education Report. Bath, UK: Redwood Books.
Waxman, H. C. & H. J. Walberg (eds.). 1994. Effective teaching: Current research. (pp. 277-93) Berkeley, CA:
McCutchan.
Wiley, D.E. and A. Harnischfeger. 1974. “Explosion of a Myth: Quantity of Schooling and Exposure to
Instruction: Major Education Vehicles.” Educational Researcher, 3, Pp. 7-11.
Wolfe, B. and J. Behrman. 1984. “Who is Schooled in Developing Countries? The Roles of Income, Parental
Schooling, Sex, Residence and Family Size.” Economics of Education Review, 3(3), Pp. 231-245.
Wood, C. 1999. Time to Teach, Time to Learn: Changing the Pace of School. Greenfield, MA: Northeast
Foundation for Children.
World Bank. 1993. Investing in Health. World Development Report 1993: New York: Oxford University Press.
World Bank. 1996. India: Primary Education Achievement and Challenges, Washington, DC: World Bank.
World Bank. 1997. Primary Education in India. Washington, DC: World Bank World Bank. 1999. Tanzania
Social Sector Review, A World Bank Country Study, Washington DC: World Bank.
World Bank. 1999/2000. “World Development Indicators.” Washington, D.C.: World Bank.
World Bank. 2001. Expanding and Improving Upper Primary Education in India. Education Sector Unit. South
Asia Region. Washington, DC: World Bank.
World Bank Human Development Network. 1999. Education Sector Strategy. Washington D.C: World Bank.
 83

World Bank Human Development Network. 2002. Education and HIV/AIDS: A Window of Hope. Washington,
DC: World Bank.
World Development Report. 2004. Making Services Work For Poor People. Washington, DC: World Bank.
World Bank and Inter-American Development Bank (IADB). 2000. Brazil Secondary Education in Brazil: Time
to Move Forward. Washington DC: World Bank.
Worthen, B.R., L.M. Van Dusen and P. J. Sailor. 1994. “A Comparative Study of the Impact of Integrated
Learning Systems on Students’ Time-on-Task”, International Journal of Educational Research, 21, 1,
Pp. 25-37.
Yair, G. 2000. “Not Just about Time: Instructional Practices and Productive Time in School.” Educational
Administration Quarterly, 36 (4), 485-512.