Active Mathematics In Classrooms: Finding Out Why

Children Make Mistakes – And Then Doing Something

To Help Them.
Allan L. White
University of Western Sydney

Introduction necessary theoretical background. The titles of the nine

workshops and booklets are:
This is the third in our series that presents a summary
of a workshop from the primary teacher professional 1. Word problems need not be so difficult
learning program that was very successfully conducted 2. Finding out why children make mistakes – And then
in Brunei Darussalam under the supervision of Professor doing something to help them
Ken Clements. While it is a mistake to think that you can 3. Developing number sense throughout the primary
take a successful program from one culture and expect mathematics curriculum
it to be an instant success in another. Nevertheless, an 4. Fractions: Going beyond area models to assist
examination can provide a vehicle for reflection of current student learning
classroom practices. For some readers, an examination 5. Decimals: Going beyond skill and drill into reality
may provide a welcome revision of some earlier learning 6. Percentages, linked to decimals and fractions
or remind them of practices long since rejected and 7. Language issues in upper primary mathematics
replaced by their current approaches. classrooms
8. Expanding the modes of communication (including
The program was called Active Mathematics in technology) in primary mathematics classrooms
Classrooms (AMIC) and had three main groups of 9. Teaching mathematics through a problem-solving
teacher participants: workshop writers; workshop and problem-posing approach.
leaders; and classroom teachers. A summary of
the process is shown diagrammatically below. A full The first workshop was presented in our last edition
description of the program can be found in Square One (White, 2005).It is the intention of the writer to now
(Vol 15, No. 2, pp. 8-15). present material from the second workshop (adapted
from Salim, 2003). As you can see from the teacher cycle
The program writers produced nine workshops and the in the figure above, the teachers began by participating in
Ministry of Education published a series of booklets to a workshop that covered the material to be used in their
support the workshops and provide the teachers with the classroom.

The Workshop
9 units prepared by writers
and presented to future The workshop has a theoretical foundation based
workshop leaders on Newman (1977), who maintained that when a
person attempted to answer a standard, written,
mathematics question then that person had to
Saturday meeting at MOE as preparation
for school workshops - writers and
be able to pass over a number of successive
Leadership Cycle
leaders hurdles: Reading (or Decoding), Comprehension,
Transformation, Process Skills, and Encoding.
Along the way, it is always possible to make
a Careless error. While there are many other
School workshops in 5 Teachers prepare a report for theoretical approaches available to teachers,
schools - Wednesdays next workshop
Newman’s offers one of the easiest to use and
adapt and has proven popular among teachers.
Teacher Cycle
In the past, teaching in the period BC (before
constructivism) teachers often had negative
Teachers and leaders
implement lessons and conduct
Teachers and leaders complete feelings about the mistakes that children would
pre- and post-tests
between unit tasks and readings
make, regarding them as “unfortunate” events that
need to be eliminated and possibly avoided at
all times. However, the strategy of more drill and
Figure 1. AMIC cycle for writers, leaders and classroom
practice has been replaced by regarding errors as
teachers (White & Clements, p. 152 ,2005a).
valuable sources of student thinking. As teachers
it is difficult to escape from children’s mistakes so
it is worthwhile finding out why children make the

Square One, Vol 15, No 4, December 2005 Page 15

mistakes in the first place (and often continue to repeat Because teachers are not fortune tellers who can predict
the mistake). Mistakes can become entrenched, so how children think all the time, they cannot read what
error analysis is the first step towards doing something is going on in children’s heads. What can they do, then,
relevant which will remove the cause of the mistake. to find out why their pupils make mistakes on written
mathematical tasks? The special interview technique
Although there are many ways to find out why children called the “Newman error analysis procedure” provides
make mistakes on mathematics tasks, the Newman error one useful method for solving the error identification and
analysis/interview procedure is highly regarded. analysis dilemma.

Workshop Objectives
15 Mathematics Questions: Write the answer to each
• To learn the Newman “analysis of errors” strategy for
question in the space provided. The teachers were asked
identifying why children make mistakes on written
to give the following questions to their sample of year
mathematical tasks.
5 students. How would your study perform on this list?
• To use the Newman error analysis interview to find
Why not use one of your own recently completed tests
out why real Primary 5 pupils made mistakes on a
and conduct the Newman analysis?
mathematics test.
• To study the types of errors made by different pupils
1. Balkis has $4.80 and Said has $6.34. How much
in a Primary 5 class on some written mathematics
more money does Said have than Balkis? Answer:
tasks, and develop error profiles for the pupils.
2. Here are three fractions: 1/3, 1/4, 1/10. Write these
fractions in order of size, from smallest to largest.
Workshop Activities
Smallest Fraction:_____ Middle Fraction: _____
The teachers were expected to:
Largest Fraction: _____
• Discuss mistakes which children would be likely 3. If Raman gives the shopkeeper $5 for a box of flour
to make on some simple paper-and-pencil that costs $3.65, how much change should he get?
mathematics questions. Answer:
• Administer a pencil-and-paper mathematics test 4. Ice cream cost 85 cents each, and apples cost
to a Primary 5 class, marks the pupils’ scripts, and 45 cents each. How much altogether would 7 ice-
prepare Newman “interview data” sheets. creams and 5 apples cost? Answer:
• Conduct Newman Interviews with Primary 5 children 5. I had a piece of rope, but someone cut 15 metres
who attended the workshop for this purpose. off it. When I measured the rope that remained, I
found it was only 57 metres long. How long was my
Activity 1: Examination of some common mathematical original piece of rope? Answer:
errors. The teachers discussed the following examples in 6. It takes Kim 15 minutes to walk to school. If school
small groups before joining a large group discussion. starts at 8:05 a.m., at what time must she leave
home so that she can get to school just as school
3+ =7 A common answer error is 10 starts? Answer:
“A plus sign means to add” 7. Radiah is 12 days older than Mary. If Radiah’s
. 35 A common answer error is 912 birthday is January 29, on what day is Mary’ s
+67 “Add the numbers in each column and birthday? Answer:
write the sums under the line 8. A shop is open from 1 o’clock in the afternoon to 5
42 a common answer error is 35. o’clock in the evening. For how many hours is the
-17 “When you subtract, you take the shop open? Answer:
smaller number from the larger”. 9. If the time is now 9 o’clock, what was the time 4
300 A common answer error is 163 or 174 hours ago? Answer:
-136 “You can’t subtract a number from zero, 10. Find the value of 940 - 586. Answer:
so you change zero into nines and the 3 11. What time will it be 35 minutes after the time shown
becomes two”
on the clock? (10 minutes past 6). Answer:
“You can’t subtract from zero, so
you borrow from the 3 and the zeros 12. Two exercise books and a pencil cost $2 altogether.
become tens”. If pencil costs 30 cents, how much does an exercise
1 + 2 A common answer error is 3/5 book cost (each book costs the same amount of
2 3 “When you add fractions, you add money as the other)? Answer:
across the top and across the bottom”. 13. There are 12 apples on the table. If you picked up
2.06 + 1.3 + 0.38 A common answer error is 2.57. one-third of the apples, how many apples would
“Line up the numbers and add”. there still be left on the table? Answer:
This is a form of decimal blindness - 14. Pensri works in a shop each day from March 22
they don’t see the point. to March 26. If she earns $10 a day, how much
$5.40 x 0.15 A common answer error is $81.00. altogether did she earn? Answer:
“After you do the calculation, bring 15. Bella takes exactly 3 hours to walk 15 km. How long
down the decimal point”. would it take her (if she walks at the same speed) to
walk 20 km? Answer:

Page 16 Square One, Vol 15, No 4, December 2005

Conducting The Newman Interview READING ERRORS (coded as R). An error would be
• Talk to the child in a friendly way, briefly, in an classified as READING if the child could not read a key
attempt to assist him/her to feel relaxed. Tell him/her word or symbol in the written problem to the extent that
your reason for talking to him/her is to help him/her this prevented him/her from proceeding further along an
with his/her Mathematics. appropriate problem- solving path.
• Tell the child you want him/her to do some of the COMPREHENSION ERRORS (coded as C). The child
problems that he/she did before, once again. had been able to read all the words in the question,
• Give the child a new test paper and answer sheet, but had not grasped the overall meaning of the words
and ask him/her to answer a question for which he and, therefore, was unable to proceed further along an
originally gave an incorrect answer. Encourage the appropriate problem-solving path.
child to show any working. Say nothing more until TRANSFORMATION ERRORS (coded as T). The child
he/she is finished. had understood what the questions wanted him/her
• Ask the child some or all of the five key Newman to find out but was unable to identify the operation, or
questions/requests (see below) for the particular sequence of operations, needed to solve the problem.
question being considered. Do not help the child
at any stage, but note (briefly) any of the child’s PROCESS SKILLS ERRORS (coded as P). The child
answers which are especially revealing. identified an appropriate operation, or sequence of
• Decide on the Newman error classification, that is to operations, but did not know the procedures necessary
say the classification that you believed corresponds to carry out these operations accurately.
to where the pupil got off track on the original pencil- ENCODING ERRORS (coded as E). The child correctly
and- paper administration of the question (ie., before worked out the solution to a problem, but could not
the interview). express this solution in an acceptable written form.
• Repeat 3, 4 and 5 for the next error.
Between-Unit Activities
The Five Newman Questions/Requests
After completing a trial run during the workshop, the
1. Please read the question to me. teachers were expected to complete the following before
2. Tell me, what is the question asking you to do? the next workshop and perform each of the following
3. Which method do you use to get your answer? activities with their class:
4. Show me how you get your answer, and “talk aloud”
• Administer the pencil-and-paper test to the class;
as you do it, so that I can understand how you are
• Analyse errors made by some pupils using the
thinking.
Newman interview technique;
5. Now, write down your actual answer.
• Fill in data sheets indicating pupils’ Newman error
categories.
During Step 4, listen carefully to what the child says
and decide where you think, when the child originally 1. Teacher’s Evaluation and Mark Sheets: Workshop
attempted the question in the whole-class context, the participants were expected to fill in the mark sheets
first breakdown point is. Hence classify the mistake/error. provided as well as an evaluation form. These
(Go at least one step beyond what you think is the first were submitted at the next workshop, after each
breakdown point) participant had conducted the Newman interviews
with at least two pupils.
The Newman Error/Mistake Analysis Procedure
2. Prepare a Report for the Next Workshop: Each
1. To identify reading errors (R): “Read the question to workshop participant was expected to give a
me. If you don’t know a word tells me.” 5-minutes report on his/her findings at the next
2. To identify comprehension errors (C): “Tell me, what workshop. They reported to workshop participants
the question asked you to do.” what they observed when performing the activities
3. To identify transformation errors (T): “Now tell me with their class and were asked to comment on the
what method you used to find the answer.” advantages and disadvantages of the activity.
4. To identify process skills error (P): “Now go over
each step of your working, and tell me what you 3. Between-Unit Reading: They were also expected to
were thinking.” read the following papers
5. To identify encoding errors (E) - an inability to a. Anthony, G. (1998). It’s all right to be wrong.
express an answer in an acceptable form: “Tell me, The Australian Mathematics Teacher, 54(4), 34-
what is the answer to the question? Point to your 37
answer.” b. Clements, M. A. (2004). Analysing Errors Made
by Pupils on Written Mathematics Tasks.
If, when the child attempts the question for the second Sultan Hassanal Bolkiah Institute of Education,
time, he/she gets the correct answer and, after the Universiti Brunei Darussalam
teacher has listened to the answers to the Newman c. Marinas, B., & Clements, M. A. (1990).
requests, the teacher is convinced that the child originally Understanding the problem: A prerequisite to
made a careless slip, then, the error would be classified problem solving in mathematics. Journal of
as CARELESS (coded as X). Science and Mathematics Education in South

Square One, Vol 15, No 4, December 2005 Page 17

 East Asia, 13(1), 14-20 gets a correct answer (as a result of the help given by
d. Smith, R. (1989). What’s going on in their the interviewer) the error is (often incorrectly) classified as
Heads? Mathematics in School, 18(5),33. a “Careless error.” In Newman interviews it is important
that the interviewer does not help students over a hurdle.
The only reading that would be difficult to obtain is the Usually, in Newman interviews, the interviewer is wise
second one and so what follows is a very brief summary to go one step beyond the first hurdle the interviewee
of the key points of this reading. In this paper different misses for any question. That is because sometimes
types of interviews that have been used successfully in a student knows how to do something but is unable
education research were considered. The main example to express that knowledge verbally in an interview.”
of a structured interview approach discussed was the (Clements, 2003, pp 4-5).
Newman (1977) diagnostic technique, a variant of which
is the technique developed by Casey (1978). The paper And further he states “In Newman research it is normal
also looked at data arising from teaching experience for students to have answered the written question twice
interviews (as used in the Erlwanger (1975) study, and before being asked the sequence of Newman questions
used by interviewers in constructivist-oriented research- for that test question. The first time usually occurs in a
(see, for example, Wright, 1990). This summary will whole-class test situation. The Newman interview should
concentrate on the Newman diagnostic interviews. take place as soon as possible after that first test, and
steps should be taken to try to ensure that students
I will not represent the Newman material already listed who took the test do not talk to each other about the
above: the process and questions. From a research solutions they gave (or their teacher does not talk to the
methodology perspective, it is important to realise that class about the test, or any question on the test). The
these key questions were carefully linked to M. Anne second time the student attempts the written question
Newman’s theory on why students make mistakes on is usually in the interview situation, before the Newman
written mathematical tasks (see Newman (1977) and questions are asked.
Newman (1983) for discussion of her theory, and of her
research based on that theory). When the interviewer begins to pose the Newman
questions to a student, the interviewer should already
“Newman also allowed for “careless errors” and errors have two pieces of data concerning the student’s
due to a lack of motivation on the part of students (see interaction with the question. The first datum is the
Newman, 1977, 1983). Newman error analysis research answer given in the whole-class test situation; the
by Marinas and Clements (1990), Singhatat (1991), second is the answer given when the child attempted
and Clements and Ellerton (1992), in Southeast Asia the question for a second time. If the test is such that
reported that about 70% of errors made by students on “student working” is shown, then this working provides
standard word problems could be attributed to a lack of extra data. Responses given to the questions asked in
comprehension or to an inability to select an appropriate the Newman interview provide important additional data.
sequence of operations (that is, in Newman’s terms, to Note that the aim of the whole exercise is to decide, by
carry out the required “Transformation”). The strength considering all the data available, why the student made
of such data suggests that one of the urgent agendas the mistakes he/she made on the original test questions
of mathematics education research is to establish a (which were questions on the test administered in the
way of addressing this state of affairs. Above all else, whole-class context).
Newman research points to the students’ lack of a deep
understanding of mathematical vocabulary, semantic The decision on why a student made a mistake on a
structure, and the absence of links between the students’ particular test question is usually made by the interviewer
formal language and mathematical skills, and their immediately after the interviewer has asked the student
personal worlds. In other words, the Newman research the Newman questions in relation to that question. The
procedure provides a framework for investigating each of interviewer should have a coding sheet on which the
the major themes discussed above (Clements & Ellerton, student’s original answer was shown. The interviewer
1996). Clements (1999) argued that mathematical should enter on the coding sheet the second answer
modelling should be present whenever a learner attempts given by the student (before the Newman questions
to solve a real-life problem or a mathematics word were asked). The coding sheet should also provide
problem. In that sense, Newman’s transformation often space for the interviewer to write a code (R, C, T, P,
involves mathematical modelling. Often, this process of E, or X) indicating the decision made on the Newman
transformation or mathematical modelling is not evident error category (on the original whole-class test) that the
in mathematics classroom discourses.” (Clements, 2004, interviewer deemed to be appropriate given all the data
pp. 2-3). now available. The coding sheet should also have some
space available if the interviewer wants to make a brief
Professor Clements also provided some advice for comment on any matter arising in the Newman interview.
inexperienced users. “ From my long experience of
observing interviewers attempting to follow Newman After the interview is complete, the following data will be
procedures, I can say that inexperienced interviewers available so far as an error made by the student when
often find it difficult to stop themselves from helping attempting the original (whole-class) test: the original
students to answer the questions. Then, when a student (whole-class test) answer given to the question;’ the

Page 18 Square One, Vol 15, No 4, December 2005

second answer given to the question; the Newman Clements, M. A. (1982). Careless errors made by sixth-grade
classification made by the interviewer as a result children on written mathematical tasks. Journal for Research in
Mathematics Education, 13(2), 136-144.
of student responses to the Newman questions.” Clements, M. A. (2004). Analysing Errors Made by Pupils on Written
(Clements, 2003, p.5). Mathematics Tasks. Sultan Hassanal Bolkiah Institute of
Education, Universiti Brunei Darussalam
Regarding the issue of careless errors, Professor Clements, M. A. (1999). Language aspects of mathematical modelling
in primary schools. In M. A. Clements & Leong Yong Pak (Eds.),
Clements states (2003, pp. 5-6). “ In Newman research Cultural and language aspects of science, mathematics, and
a careless error has been defined as one which occurred technical education (pp. 363-372). Gadong: Universiti Brunei
even though the student knew (from a cognitive Darussalam.
perspective) exactly how to gain a correct answer to Clements, M. A., & Ellerton, N. F. (1992). Overemphasising process
skills in school mathematics: Newman analysis data from five
the question at the time the incorrect answer was given countries. In W. Geeslin & K. Graham (Eds.), Proceedings of
and would be expected to give the correct answer when the Sixteenth International Conference on the Psychology
responding to the same question at some later time. of Mathematics Education (Vol. 1, pp. 145-152). Durham,
New Hampshire: International Group for the Psychology of
Thus, if a student gave an incorrect response in the Mathematics Education.
Clements, M. A., & Ellerton, N. F. (1996). Mathematics education
original whole-class test situation but then gave a correct research: Past, present andfuture. Bangkok: UNESCO.
answer immediately before the Newman interviews, Ellerton, N. F:, & Clarkson, P. C. (1992). Language factors in
then the interviewer would suspect that an X (Careless mathematics education. In B.Atweh & J. Watson (Eds.),
error) classification of the error might be appropriate. Research in mathematics education in Australasia 1988-1991
(pp. 153-178). Brisbane: Mathematics Education Research
Data from the Newman interviews should then enable Group of Australasia.
the interviewer to decide whether the X-classification is Ellerton, N. F., & Clements, M. A. (1996). Newman error analysis: A
really appropriate. If, during the Newman interview for comparative study involving Year 7 students in Malaysia and
that question, it becomes clear that the student was not Australia. In P. Clarkson (Ed.), Technology in Mathematics
Education (pp. 186-193). Melbourne: Mathematics Education
sure which of the two answers that he had given - the Research Group of Australasia.
incorrect one, given when the test was administered to Ellerton, N. F., & Clements, M. A. (1996). Newman error analysis. A
the whole class, or the correct one, given just before the comparative study involving Year 7 students in Malaysia and
Newman interview took place - was correct, then the Australia. In P. C. Clarkson (Ed.), Technology and mathematics
education (pp. 186-193). Melbourne: Mathematics education
original error should not be classified as X (Careless)”. Research Group of Australasia.
Ellerton, N. F., & Clements, M. A. (1997). Pencil and paper tests
The conclusion of the article discusses the importance under the microscope. In F. Biddulph & K. Carr (Eds.), People
of producing summary sheets and how to best extract in mathematics education (pp. 155-162). Waikato, NZ:
Mathematics Education Research Group of Australasia.
patterns of errors that can provide direction for future Erlwanger, S. H. (1975). Case studies of children’s conceptions
classroom planning. The two modifications are also of mathematics - Part 1. Journal of Children’s Mathematical
discussed. Behavior, 1(3), 157-283.
Marinas, B., & Clements, M. A. (1990). Understanding the problem:
While this is a far from satisfactory summary, if the A prerequisite to problem solving in mathematics. Journal of
Science and Mathematics Education in South East Asia, 13(1),
reader is keen then the main ideas can be found in the 14-20.
list of papers by Clements in the reference section. The Newman, M. A. (1977). An analysis of sixth-grade pupils’ errors on
reference list provides a good foundation for further written mathematical tasks. Victorian Institute for Educational
reading. Alternatively, if you contact me at al.white@uws. Research Bulletin, 39, 31-43.
Newman, M. A. (1983). Strategies for diagnosis and remediation.
edu.au, I will send you a copy of Clements’ 2003 paper. Sydney: Harcourt, Brace Jovanovich..
Salim, R. (Rozina bte Awg Hj Salim) (2003). Finding out why children
make mistakes – And then doing something to help them.
Conclusion Bandar Seri Begawan, Brunei Darussalam: Brunei Darussalam
In this attempt to condense another of the workshops, it Ministry of Education.
Singhatat, N. (1991). Analysis of mathematics errors of lower
is hoped that enough material was provided to do justice secondary pupils in solving word problems. Penang: SEAMEO-
to the original. While there are many ways for finding out RECSAM.
why children make mistakes, the approach presented Smith, R. (1989). What’s going on in their Heads? Mathematics in
in this unit has been popular with teachers. It offers an School, 18(5),33.
White, A. L. (2005). Active Mathematics in Classrooms: Word
easily understood framework, backed by research, for problems need not be so difficult. Square One, 15 (3), 15-19.
working with student errors. So how did your students White, A. L., & M. A. Clements (2005a). Energising upper-primary
do? mathematics classrooms in Brunei Darussalam: The Active
Mathematics In Classrooms (AMIC) Project. In H. S. Dhindsa,
I. J. Kyeleve, O. Chukwu, & J.S.H.Q. Perera (Eds.). Future
References directions in science, mathematics and technical education (pp.
Anthony, G. (1998). It’s all right to be wrong. The Australian 151-160 ) (Proceedings of the Tenth International Conference).
Mathematics Teacher, 54(4), 34-37. Brunei: University Brunei Darussalam.
Casey, D. P. (1978). Failing students: A strategy of error analysis. In P. White, A. L., & M. A. Clements (2005b). Energising upper-primary
Costello (Ed.), Aspects of motivation (pp. 295-306). Melbourne: mathematics classrooms in Brunei Darussalam: The Active
Mathematical Association of Victoria. Mathematics In Classrooms (AMIC) Project (Reprinted) Square
Clements, M. A. (1980). Analysing children’s errors on written One, 15 (2), 8-15.
mathematical tasks. Educational Studies in Mathematics, 11(1), Wright, B. (1990). Interactive communication: Constraints and
1-21. possibilities. In L. P. Steffe & T. Wood (Eds.), Transforming
children’s mathematics education: International perspectives (pp.
235-243). Hillsdale, NJ: Lawrence Erlbaum.

Square One, Vol 15, No 4, December 2005 Page 19