Ethnicity, Teachers' Expectations, and the Academic Achievement Process in Canada

Author(s): Rodney A. Clifton

Source: Sociology of Education, Vol. 54, No. 4 (Oct., 1981), pp. 291-301
Published by: American Sociological Association
Stable URL: https://www.jstor.org/stable/2112570
Accessed: 26-02-2019 09:32 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

American Sociological Association is collaborating with JSTOR to digitize, preserve and

extend access to Sociology of Education

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY, TEACHERS' EXPECTATIONS, AND THE ACADEMIC
ACHIEVEMENT PROCESS IN CANADA*

RODNEY A. CLIFTON
University of Manitoba

Sociology of Education 1981, Vol. 54 (October):291-301

This research examines the extent to which differences in achievement between

German-speaking and French-speaking students are accountedfor by the expectations of their
teachers. A causal model incorporating both normative and cognitive expectations of teachers
is used. The results suggest that teachers base their expectations on the intellectual ability and
previous performance of students rather than upon their ethnicity or other ascribed
characteristics. Moreover, the data indicate that expectations affect the teachers' evaluation of
learning to a greater degree than they affect actual learning.

There has been considerable research and faster than children with other as-
on the effects of teachers' expectations cribed characteristics.
upon the performances of their students The present study is an empirical in-
(Dusek, 1975; Braun, 1976; Cooper, 1979), vestigation of the extent to which the
much of it stimulated by the publication of ethnicity of students shapes expectations
Pygmalion in the Classroom (Rosenthal of their teachers and how both these vari-
and Jacobson, 1968). Researchers and ables affect students' achievement
teachers have been concerned that the (Dusek, 1975:680; Braun, 1976:191).
performances of students may be lowered,
rather than raised, because expectations
BACKGROUND
for performances are low (Finn,
1972:387), especially if students are disad- Within Canada, the examination of
vantaged or come from minority groups. stratification and mobility has
Expectations of teachers may be influ- traditionally been related to class and
enced by both ascribed characteristics of ethnicity. John Porter (1965) argues that
students that are essentially irrelevant to ethnicity has been a principle cause of
performances (ethnicity, physical attrac- class structures, and there are a number of
tiveness, sex, etc.) and achieved charac- ways that ethnicity could either hinder or
teristics that are relevant (intellectual facilitate upward mobility. The most
ability, previous performance, academic common theme has been that ethnic sub-
aspirations, etc.). Teachers' expectations, cultures may shape the development of
if based upon irrelevant ascribed charac- abilities, aspirations, and motivation in
teristics, may shape teacher-student in- order to preserve culture-specific values
teractions and performances in such a way and attitudes. Children socialized within
that initial false definitions of the situation ethnic subcultures have developed
may, over time, become true (Merton, abilities, aspirations, and motivation
1957:423). A number of scholars (Rist, which match certain occupational status
1970; Carnoy, 1974; Bowles and Gintis, positions within society. In turn, the
1976) have argued that schools play a shaping of abilities, aspirations, and moti-
major role in perpetuating the existing so- vation within ethnic subcultures could set
cial stratification system, in part because limits on educational achievement which
teachers assume that children with certain could further limit occupational achieve-
ascribed characteristics can learn more ment (Breton and Rosenborough, 1968:4;
Blishen, 1970:120).
* I would like to thank two anonymous reviewers Empirical evidence (Porter 1965:90) in-
for helpful comments and suggestions. Address cor-
dicates overrepresentation of British,
respondence to Dr. Rodney A. Clifton, Depart-
ment of Educational Administration and Foun-
Jewish, German, and other northern
dations, University of Manitoba, Winnipeg, European groups in the upper levels of the
Manitoba, R3T 2N2 Canada. occupational structure in Canadian soci-

291

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 292 CLIFTON

ety and underrepresentation of the main language at home and, as a result,

French, Italian, native Indian and Inuit were classified as English-speaking.
groups. More recent research on ethnic Third, the French-speaking and German-
immigrants in Toronto (Li, 1978) has il- speaking students were the two largest
lustrated that German and British adults ethnic groups represented in the data set.
have very similar educational and occu- Thus, there was an adequate number of
pational achievement. Additionally, it was students from both groups for analyses.
shown that occupational achievement was Even though ethnicity has been an im-
largely the result of educational achieve- portant factor in the analyses of the occu-
ment. Previous research on the data used pational stratification system in Canada, it
in the present study (King, 1968) showed has not generally been included in the
that German-speaking students main- analyses of inequalities in education nor in
tained very similar performances and re- the way in which education fits into the
tention rates to Dutch-, English-, occupational stratification system (Breton
Hungarian-, and Polish-speaking students. et al., 1975:83). From previous research it
The performances and retention rate for is expected that ethnicity would have im-
Yiddish-speaking students were substan- portant direct and indirecf effects upon
tially higher than all other groups while educational performances. Ethnicity is
the performances and retention rate for probably related to ability, aspirations,
French-speaking students were substan- and motivation and the educational sys-
tially lower than all other groups. The tem may simply sort students on these
differences between the ethnic groups criteria. Moreover, teachers may hold
were maintained even when size of com- differential expectations based upon
munity, father's and mother's education, ethnicity and these expectations may af-
father's occupation, and number of chil- fect the performances of students. The
dren in the family were controlled. main concern of this research is with the
For the present research, data on the relationships between ethnicity, teachers'
total population of students enrolled in expectations, and educational perfor-
Grade 9 (over 90,000 students) in the mances.
Province of Ontario were available. From
this data set, only French-speaking and
THE MODEL
German-speaking students were selected
for analysis. There are three reasons for In an extensive review of 345 studies,
selecting the students from these two Rosenthal and Rubin (1978) suggested that
ethnic groups. First, past research showed the expectations that experimenters have
that the German-speaking students were for the behavior of their subjects and the
very similar in performances and retention expectations that teachers have for the
to students who spoke other northern performance of their students are com-
European languages while the French- municated through both verbal and non-
speaking students had much lower per- verbal means. In most cases, when re-
formances and a much lower retention searchers or teachers expect certain levels
rate than students from other ethnic of performances from their subjects or
groups. This fits the notion that, in On- students they communicate these expec-
tario, the French-speaking citizens repre- tations in such a way that neither the re-
sent a "white siege culture" (Breton, Bur- searchers, teachers, subjects, nor stu-
net, Hartmann, Isajiw, and Lennards, dents know exactly how the information is
1975:88). Second, even though it may transferred (Rosenthal, 1976). Moreover,
seem logical to compare minority students the subjects and students often do not
(French, for example) with majority (En- know how or why they internalize the ex-
glish) students, the English-speaking stu- pectations in such a way that their behav-
dents were not selected because Cana- ior is affected.
dian Census data revealed that this group Brophy and Good (1970) and Rosenthal
represented a variety of ethnic/linguistic (1976) have postulated theoretical linkages
origins. That is, a number of students from which may help explain how expectations
other ethnic groups spoke English as their operate within classrooms. First, it is

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY AND CANADIAN ACADEMIC ACHIEVEMENT 293

postulated that a climate is established mances whereas cognitive expectations

within a classroom so that differential ex- relate to academic achievement. Norma-
pectations for performances are estab- tive expectations concern the manner in
lished and accepted by both the students which students are seen, by their
and the teachers. Second, certain students teachers, as meeting the behavioral stan-
begin to be treated differentially in ac- dards of the classroom whereas cognitive
cordance with the differential expecta- expectations concern the manner in which
tions held by the teachers. Third, in re- students are seen as meeting the academic
sponse to differential treatment, each stu- standards of the classroom. These two
dent tends to exhibit behavior which com- different types of expectations have been
plements and reinforces the teachers' ex- identified in the literature on academic
pectations. Finally, over a period of time, and behavioral difficulties pupils en-
the achievement levels of the students and counter in school (Rubin and Balow,
the social organization of the classroom 1979). Nevertheless, only Williams (1976)
reflects, to some degree, differences in the has integrated the two types of expecta-
expectations of the teachers and the dif- tions into research on the manner in which
ferential expectations are exhibited in the teachers mediate the relationships be-
results of achievement tests. tween the socioeconomic status and in-
Within classrooms it is possible to dis- tellectual abilities of students, on the one
tinguish between two interrelated dimen- hand, and their academic performances,
sions of the educational enterprise. There on the other.
is an instructional dimension that pertains These linkages, in terms of a causal
to knowledge and skills and there is a so- model, are illustrated in Figure 1. It is
cial control dimension that pertains to the noted that both normative and cognitive
management of social interaction (see expectations are assumed to cause three
Parsons, 1959; Bidwell, 1972). Report different types of achievement: assigned
cards, which are issued to students and grades, standardized achievement in En-
their parents, and cumulative records, glish, and standardized achievement in
which are consulted by teachers and other mathematics.
professionals in the schools, convey Previous research has shown that
teachers' assessments of students on both teachers' expectations have effects upon
instruction and social control. Since both standardized achievement tests (Brophy
dimensions are seen as being important, it and Good, 1974; Dusek, 1975; Braun,
is logical to assume that teachers take 1976; Cooper, 1979). However, even
both of them into consideration in devel- though previous research has focused on
oping expectations for their students. standardized achievement tests, it is hy-
From this perspective, two types of ex- pothesized that teachers' expectations,
pectations may be identified: normative particularly cognitive expectations, will
expectations and cognitive expectations have a strong impact on teacher assigned
(Williams, 1976). Normative expectations grades because teachers may either con-
relate to normative or behavioral perfor- sciously or unconsciously bias their grad-

6. Normative Expectations-

4. Assigned Grades _

1. Socioeconomic Status. - - . 8. Assigned Grades

2. Intellectual Ability _ _ 9. Standardized Achievement

(English)
3. Ethnicity - _ 10. Standardized Achievement
(Mathematics)

5. Academic Aspirations _ --

7. Cognitive Expectations _

Autumn 1959 Spring 1960 Autumn 1960 Spring 1961

Figure 1. The Theoretical Model

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 294 CLIFTON

ing of tests which they have constructed. effects are via normative and cognitive
Both teacher assigned grades and stan- expectations.
dardized achievement tests are included There is also substantial evidence that
as dependent variables. the sex of the pupils, affects various as-
West and Anderson (1976) have argued pects of the academic achievement pro-
that the literature supports the notion that cess (Richer, 1979). Moreover, within the
student achievement causes expectations teacher expectations literature there is
and expectations, in turn, cause achieve- specific evidence of differential influence
ment. In the present study, teacher as- attributed to the sex of pupils (Kehle,
signed grades are included at two points in 1974). For this reason, analyses for each
time: spring 1960 and spring 1961. As a sex are conducted.
result, it is hypothesized that grades as-
signed in the spring of 1960 affect both METHODOLOGY
normative and cognitive expectations in
Data
the autumn of 1960 and that all of these, in
turn, affect both assigned grades and Data were obtained from the Carnegie
standardized achievement in the spring of Human Resources Data Bank, a five-year
1961'.1 panel study of practically the total popu-
It is also probable that the academic lation of students (90,719) who were en-
aspirations of students play a major role rolled
in in the first year of high school,
mediating any relationships between Grade 9, in the Province of Ontario at the
ethnicity, socioeconomic status, and both beginning of the 1959-1960 academic
teachers' expectations and students' year. Students were classified on their re-
achievement (Williams, 1972). As a result, sponses to the question, "What is the main
this variable is incorporated into the language spoken in your home?" On this
model. basis, 79 percent of the students indicated
There is substantial evidence that both that they were from English-speaking
ethnicity and socioeconomic status affect homes, 5 percent indicated that they were
the expectations of teachers (Marwit, from French-speaking homes, 1 percent
Marwit, and Walker, 1978; Feldman and indicated that they were from German-
Orchowsky, 1979). The model presented speaking homes, and 15 percent indicated
here hypothesizes that these two vari- that they were from homes in which other
ables, as well as intellectual ability, not languages were spoken. The French-
only affect normative and cognitive ex- speaking and German-speaking students
pectations, but they also affect all of the were selected for this research because
other variables. That is, it is hypothesized both represented large homogeneous
that socioeconomic status, intellectual ethnic groups with the German-speaking
ability, and ethnicity have both direct and students generally outperforming the
indirect effects upon the achievement of French-speaking students. In total, 4,733
students. Part of the indirect effects are students spoke French and 1,045 students
via assigned grades in Grade 9 and aca- spoke German as their main language. Of
demic aspirations and part of the indirect these, 2,440 French-speaking students (52
percent) and 707 German-speaking stu-
dents
1 When the students entered Grade 10, in the (68 percent) completed Grade 10 in
autumn of 1960, the majority of their teachers al- the spring of 1961.2
ready knew them from the previous year. In Ontario,
high school teachers specialize in subject areas and 2 The other students dropped out of school, or
may teach students from Grade 9 to Grade 13. In thismoved to another province or country, with the
respect, over 85 percent of the students who did notmajority dropping out or moving during July and
drop out of school or move to another province or August. As noted previously, the retention rate for
country remained in the same school for both the the French-speaking students was the lowest of all
1959-1960 and 1960-1961 academic years. Also, the ethnic groups while the retention rate for the
teachers generally review the cumulative records of German-speaking students was about average (King,
their students who have recently transferred to the 1968). There were also slight differences between the
school. Thus, it is usual for teachers to have consid-
sexes: for the French-speaking students the retention
erable information about their students' previous rate for the females was slightly higher than the rate
performances.
for the males whereas for the German-speaking stu-

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY AND CANADIAN ACADEMIC ACHIEVEMENT 295

Variables point scale which ranged from " much

below average" to "much above average"
Complete descriptions of the question-
was used. The three items were summed
naires and tests that were administered to
to form the normative expectations scale.
the students can be found in Brehaut
Cognitive Expectations (X7). The
(1964) and D'Oyley (1964). The following
teachers were requested to provide a
ten variables are used in the present
similar rating on each student's chance of
study.
completing Grade 13 successfully, the
Socioeconomic Status (x1). To create
final year of high school. It must be recog-
this variable three student-reported mea-
nized that for both normative and cogni-
sures are summed: father's occupation, an
tive expectations considerable between-
eight-category scale based on Blishen and
teacher differences may be obliterated by
McRoberts (1976); father's and mother's
averaging the responses of different
education, five-point scales ranging from
groups of teachers.
"no secondary schoorl to "university de-
gree" .
Standardized Achievement Tests (xg,
x10). Two tests are used: the CATE, an
Intellectual Ability (X2). The students'
English language achievement test and the
scores on the Canadian Academic Apti-
CATM, a standardized achievement test
tude Test III (non-verbal reasoning) were
in mathematics.
taken as the indicator of intellectual
ability.
Advantages of the Data
Ethnicity (X3). This variable is created
from the respondents' answers to a ques- This data set has four advantages for the
tion about the major language that is spo- analysis of ethnicity, teachers' expecta-
ken in their homes. Students who spoke tions, and academic achievement. First,
French were coded as 1 and students who the data included a variable which can be
spoke German were coded as 2. used to measure ethnicity rather accu-
Assigned Grades (X4, X8). In both Grade rately.3 Second, the data were obtained
9 and Grade 10 average grades were cal- from the total population of students in
culated from the marks, for all relevant Grade 9 in the Province of Ontario during
subjects, which were assigned by the the 1959-1960 academic year. Thus, gen-
teachers. Assigned grades were obtained eralizations may at least apply to this
from the school records. province.4 Third, data were collected over
Academic Aspirations (X5). Subjects a number of years so that the variables
were asked to respond to a question about were measured at different periods during
their own educational and work plans. the students' academic careers (see Figure
Seven responses were provided, ranging 1). Fourth, in many studies only the link-
from "leave school as soon as possible to ages between the characteristics of pupils
obtain a job" to "complete secondary and the expectations of teachers have
school and enter university". been examined, and often only in con-
Normative Expectations (X6). Teachers trived or experimental situations (Dusek,
in each school were asked to rate each 1975; Braun, 1976; Cooper, 1979). In the
student's reliability as indicated by per- present study these relationships are ex-
formance on curricular and extra- amined as they existed within actual
curricular duties, cooperation with school settings rather than laboratory set-
teachers and students, and industry in tings (see Rosenthal and Rubin, 1978:
school work. The teachers were requested 383).
to meet as a group and give ratings which
would represent their combined judgment I This was not true for more recent large-scale

for each student. For each item, a five- research on related aspects of the education of
Canadian students (see Breton, 1972; Porter, Porter,
and Blishen, 1973).
dents the retention rate for the males was slightly 4 Since the data were collected 20 years ago, it is
higher than the rate for the females. In view of the not known if the results are applicable at the present
differential retention rates, generalizations based time. Thus, until more recent longitudinal data are
upon this analysis must be considered with great collected on sufficiently large ethnic subsamples
caution. these results must be regarded as tentative.

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 296 CLIFTON

RESULTS cluded in the analysis. However, before

The relevant correlation coefficients, assigned grades and academic aspirations
means, and standard deviations are pre- in Grade 9 were incorporated, the effects
sented in Table 1. Ethnicity has positive of ethnicity were substantially higher.
correlations with most other variables. Specifically, the effects of ethnicity de-
Analyses of variance examining the re- creased by 48 percent for normative ex-
lationships between ethnicity and each of pectations and by 62 percent for cognitive
the other variables in the model revealed expectations.
that, except for the academic aspirations Quite different results are observed for
of females, the German students had sub- the female students (Panel B). When as-
stantially higher mean scores (p.S.0001) signed grades in Grade 9 and academic
than the French students. aspirations are included, the effects of
Both normative and cognitive expecta- ethnicity are not decreased as much as for
tions are very highly related to the three the males. That is, for normative expecta-
dependent variables, assigned grades and tions the effect of ethnicity decreased by 7
standardized achievement in both English percent and for cognitive expectations the
and mathematics. As expected, cognitive effect of ethnicity decreased by 9 percent.
expectations have higher correlations with Nevertheless, for both males and fe-
the three performance scores than with males past performances have the most
normative expectations, and both norma- powerful effects upon both normative and
tive and cognitive expectations have cognitive expectations of teachers while
higher correlations with teacher assigned the second most powerful effect on cogni-
grades than with the standardized tive expectations is from academic aspi-
achievement tests. rations. This means that what originally
To identify the direct and indirect ef- may have been interpreted as moderately
fects of the independent and intervening strong ethnic bias on the part of teachers,
variables upon the dependent variables a is, to a large degree, the effect of previous
path analysis was undertaken (see Dun- achievement and aspirations. If students
can, 1975; Heise, 1975). The path coeffi- have done well and have high aspirations
cients, path regressions, and R2 are re- then, other things being equal, their
ported in Table 2. Only coefficients re- teachers expect them to be responsible in
flecting the fully recursive model are pre-following the generally accepted norms
sented. and to complete high school. These find-
Beginning with normative and cognitive ings are congruent with Dreeben's (1968:
expectations it is observed that, for males 44) claim that, in schools "pupils learn to
(Panel A), the effects of ethnicity are small accept principles of conduct, or social
when all five exogenous variables are in- norms, and to act according to them."

Table 1. Correlations, Means and Standard Deviations for All Variables by Sexa

Standard
Means Deviations
XI X2 X3 X4 X5 X6 X7 X8 X9 XIO Males Females Males Females
XI .155 .133 .151 .290 .054 .177 .129 .194 .147 6.48 6.32 3.03 3.02
X2 .143 .235 .438 .189 .189 .298 .341 .392 .466 27.71 27.35 8.82 8.25
X3 .125 .192 .253 .115 .193 .206 .219 .333 .330 1.19 1.16 0.39 0.37
X4 .205 .395 .115 .294 .496 .570 .714 .561 .541 57.76 52.46 14.16 13.76
X5 .210 .128-.021 .274 .160 .343 .274 .344 .177 4.91 4.70 1.92 1.69
X6 .027 .179 .182 .408 .083 .642 .620 .340 .338 9.54 10.04 2.46 2.38
X7 .168 .309 .156 .611 .247 .564 .665 .550 .432 2.49 2.54 1.16 1.09
X8 .133 .268 .103 .661 .150 .502 .651 .523 .542 61.73 64.75 12.21 11.41
Xg .148 .410 .259 .515 .206 .301 .515 .499 .536 55.15 59.05 13.07 12.38
X1o .115 .410 .211 .474 .121 .305 .442 .459 .534 16.91 16.28 5.17 5.11
a Correlations for the males are above the diagonal.
1 = Socioeconomic Status, 2 = Intellectual Ability, 3 = Ethnicity, 4 = Assigned Grades (Grade 9),
5 = Academic Aspirations, 6 = Normative Expectations, 7 = Cognitive Expectations, 8 = Assigned Grades
(Grade 10), 9 = Standardized Achievement (English), 10 = Standardized Achievement (Mathematics).

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY AND CANADIAN ACADEMIC ACHIEVEMENT 297
Table 2. Path Coefficient, Path Regre$sions, and R2 by Sexa

Dependent Predetermined Variables

Variables XI x2 X3 X4 X5 X6 X7 R2
A. Males
X4 .070 .391 .152 .220
(.327) (.628) (5.50)
X5 .262 .137 .048 .107
(.166) (.030) (.234)
X6 --.031 -.045 .081 .493 .023 .253
(-.025) (-.013) (.505) (.086) (.030)
X7 .041 .035 .050 .484 .176 .365
(.016) (.005) (.148) (.040) (.107)
X8 - .003 .027 .003 .438 .017 .234 .251 .639
(-.013) (.037) (.095) (.377) (.109) (1.16) (2.63)
Xg .024 .122 .165 .284 .112 -.087 .331 .458
(.102) (.181) (5.52) (.262) (.763) (-.464) (3.72)
YIlO .024 .251 .160 .304 -.039 .018 .149 .401
(.042) (.147) (2.12) (.111) (-.104) (.037) (.659)
B. Females
X4 ..149 .369 .026 .179
(.677) (.614) (.960)
X5 .202 .112 -.067 .058
(.113) (.023) (-.309)
X6 --.073 .002 .143 .409 -.011 .191
(-.057) (.006) (.921) (.071) (-.016)
X7 .018 .063 .081 .550 .086 .392
(.007) (.008) (.238) (.044) (.055)
X8 .005 -.016 -.025 .420 -.059 .158 .328 .556
(.021) (-.022) (-.764) (.348) (-.396) (.756) (3.43)
Xg -.007 .196 .155 .244 .050 -.026 .285 .394
(-.029) (.293) (5.19) (.219) (.369) (-.136) (3.24)
XIO --.010 .232 .102 .248 -.022 .040 .188 .325
(-.017) (.144) (1.41) (.092) (-.068) (.085) (.880)
a Path regressions are in parentheses.
I = Socioeconomic Status, 2 = Intellectual Ability, 3 = Ethnicity, 4 = Assigned Grades (Grade 9),
5 = Academic Aspirations, 6 = Normative Expectations, 7 = Cognitive Expectations, 8 = Assigned Grades
(Grade 10), 9 = Standardized Achievement (English), 10 = Standardized Achievement (Mathematics).

It must be noted, however, that ethnic the effects of ethnicity on assigned grades
bias has not been completely eliminated are extremely small while ethnicity has
by including assigned grades and aca- moderate effects upon achievement in
demic aspirations in the analysis of both English and mathematics.
teachers' expectations. There are still sig- Also, except for cognitive expectations
nificant effects of ethnicity upon both for males, teachers' expectations have
cognitive and normative expectations. stronger effects upon assigned grades than
Moreover, it is not known to what extent upon standardized achievement in English
academic achievement and aspirations in and mathematics. This is understandable
Grade 9 were, in themselves, biased as a because the grades have been assigned by
result of ethnicity and the indirect effects the teachers while the achievement tests
this may have had upon the teachers' ex- have been evaluated independently of the
pectations in Grade 10. In any event, the teachers (see Cooper, 1979). Moreover,
results presented here indicate that when for both sexes and both standardized
socioeconomic status, intellectual ability, achievement measures, cognitive expec-
assigned grades, and academic aspirations tations have more powerful effects than
are equal, there are differences between normative expectations. Additionally,
the German and French students. For cognitive expectations have more sub-
both sexes, teachers have higher expecta- stantial effects upon achievement in En-
tions for the German students than for theglish than achievement in mathematics. At
French students. the same time, for both males and fe-
For both males and females in Grade 10, males, teachers' expectations have posi-

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 298 CLIFTON

tive and relatively strong effects upon as- portance of considering sex in under-
signed grades. For males, both normative standing the academic achievement pro-
and cognitive expectations have almost cess (see Maccoby and Jacklin, 1974;
identical effects. For females, on the other Richer, 1979). In order to provide a closer
hand, cognitive expectations are more examination of these conditional relation-
than twice as powerful as normative ex- ships, as well as to provide estimates of
pectations. This suggests that quite dif- the strength of the indirect effects of
ferent processes occur when teachers as- ethnicity on performances, the direct and
sign grades to males and females. For indirect effects (Alwin and Hauser, 1975)
males, grades are assigned equally on the of the predetermined variables are pre-
basis of normative and cognitive expecta- sented in Table 3.
tions while for females, grades are as- In this table, for both males and fe-
signed to a greater degree on the basis of males, very little of the ethnic effects are
cognitive expectations than on normative mediated through teachers' expectations.
expectations (cf. Maccoby and Jacklin, Nevertheless, the indirect effects of
1974:247-248). ethnicity are consistently stronger for fe-
The results, to this point, indicate some males than for males, illustrating the con-
important ethnic-by-sex interaction effects ditional relationship between ethnicity,
upon both expectations and achievement. teachers' expectations, and the achieve-
Therefore, the effects of sex across the ment variables. It is also noted that the
ethnic groups are examined in more detail. indirect effects are lower for both stan-
These comparisons are based upon metric dardized achievement measures than for
coefficients rather than standardized assigned grades.
coefficients (Kim and Mueller, 1976). Also, this table illustrates that past
It is observed that for both normative performances have, by far, the strongest
and cognitive expectations the effects of mediated effects through teachers' ex-
ethnicity are stronger for females than forpectations upon all three achievement
males. Also, for assigned grades in Grade measures. Thus, the data offer little sup-
10, the effects of ethnicity are stronger forport to common arguments suggesting
females than for males but the difference teachers' expectations have strong
is not as important as those noted above mediating effects between the ascribed
because the coefficients are extremely characteristics of students and their
small. For standardized achievement in achievement. And, the evidence suggests
English the effects of ethnicity are similar that the indirect effects of ethnicity are
for both sexes, while for standardized slightly stronger upon grades which have
achievement in mathematics the effects of been assigned by the teachers than upon
ethnicity are stronger for males than for standardized achievement tests.
females. For assigned grades in Grade 10, Nevertheless, the teachers' expecta-
the effects of normative expectations are tions, by themselves, do affect both as-
stronger for males than for females while signed grades and standardized achieve-
the effects of cognitive expectations are ment measures. This finding has been
stronger for females than for males. How- previously reported (Brophy and Good,
ever, for standardized achievement in En- 1974; Dusek, 1975; Cooper, 1979). But,
glish, the effects of both normative and the present study extends past research by
cognitive expectations are stronger for showing that the mediated effects are from
males than for females, while for stan- past performance and intellectual ability
dardized achievement in mathematics the rather than from ethnicity or, for that
effects of both normative and cognitive matter, from socioeconomic status. That
expectations are stronger for females than is, the data support the argument that
for males. teachers base their expectations upon
These results indicate that the relation- meritorious factors rather than ascribed
ships between ethnicity, teachers' expec- factors. And, furthermore, it is the
tations, and each of the three performance meritorious factors of past performance
variables are conditional upon the sex of and intellectual ability, as mediated
the students. This illustrates the im- through teachers' expectations, which af-

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY AND CANADIAN ACADEMIC ACHIEVEMENT 299

Table 3. Total, Indirect, and Direct Effects of the Predetermined Variables upon Assigned Grades and
Standardized Achievement

Dependent Predetermined Total Indirect Effects via Direct

Variables Variables Effects X4X5 X6X7 Effects
A. Males
X8 Xl .065 .065 .003 -.003
X2 .298 .273 - .002 .027
X3 ..140 .106 .031 .003
X4 .675 .237 .438
Xs .067 .050 .017
Xg XI .112 .072 .016 .024
X2 .318 .180 .016 .122
X3 .244 .069 .010 .165
X4 .401 .117 .284
X5 .168 .056 .112
XIO XI .054 .024 .006 .024
X2 .404 .149 .004 .251
X3 .227 .058 .009 .160
X4 .385 .081 .304
X5 - .012 .027 - .039

B. Females
X8 XI .092 .092 - .005 .005
X2 .247 .242 .021 -.016
X3 .044 .020 .049 - .025
X4 .665 .245 .420
X5 - .032 .027 - .059
Xg XI .073 .073 .007 -.007
X2 .365 .152 .017 .196
X3 .179 .005 .019 .155
X4 .390 .146 .244
X5 .075 .025 .050
X,0 X, .044 .054 .000 -.010
X2 .379 .135 .012 .232
X3 ..133 .010 .021 .102
X4 .367 .119 .248
X5 - .007 .015 - .022

1 = Socioeconomic Status, 2 = Intellectual Ability, 3 = Ethnicity, 4= Assigned Grades (Grade 9),

5 = Academic Aspirations, 6 = Normative Expectations, 7 = Cognitive Expectations, 8 = Assigned Grades
(Grade 10), 9 = Standardized Achievement (English), 10 = Standardized Achievement (Mathematics).

fect student achievement to the greatest in actual achievement. At a later period cf

degree. time this may lead to greater variation in
achievement in both the educational in-
stitution and the occupational world.
CONCLUSION
The evidence presented here suggests
The literature suggests that the expec- that teachers appear to engage in dis-
tations of teachers serve to sustain pre- crimination based upon past performance
existing achievement variations among and intellectual ability to a much greater
students (Cooper, 1979). Furthermore, it degree than upon ethnicity or socioeco-
has been suggested that, in Canada, edu- nomic status. The data also suggest that
cational institutions may be perpetuating, teachers evaluate student achievement on
in part, the vertical mosaic (Porter, 1965) the basis of both normative and cognitive
because teachers assume that children expectations, whereas performance on
from certain ethnic groups can learn more standardized achievement tests is influ-
and faster than children from other ethnic enced mainly by cognitive expectations.
groups. Moreover, it is argued that The fact that the evaluation of student
teachers' expectations play a role in achievement is based upon both norma-
translating the differences among students tive and cognitive criteria suggests that, to
in potential achievement into differences some degree. ideas about meritocratic

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 300 CLIFTON

processes are not completely consistent pational status positions within society
with what happens in schools (Dreeban, (Breton and Rosenborough, 1968:4;
1968). In a pure meritocratic system, Blishen, 1970:120). The differences in
teachers would evaluate the academic achievement between the French- and
performance of students entirely upon German-speaking students may reflect, to
their cognitive expectations which, in some degree, differences in the subcul-
turn, would be based upon their past per- tural backgrounds which include abilities,
formance and their intellectual ability. aspirations, and motivations.
Nevertheless, the effects of ethnicity, as
mediated through teachers' expectations,
REFERENCES
are relatively small, and it is within this
context that meritocratic factors may not Alwin, Duane F. and Robert M. Hauser
be operating in schools. 1975 "The decomposition of effects in path
analysis." American Sociological Review
The present data are limited to high
40:37-47.
school students in Grades 9 and 10, and it Bidwell, C. E.
is possible that teachers' expectations had 1972 "Schooling and socialization for moral com-
much greater effects before the students mitment." Interchange 3:1-27.
reached high school. That is, the linkages Blishen, Bernard R.
1970 "Social class and opportunity in Canada."
between ethnicity, intellectual ability,
Canadian Review of Sociology and An-
past performances, cognitive expecta- thropology 7:110-27.
tions, and standardized achievement ob- Blishen, Bernard R. and Hugh A. McRoberts
served here may have been the result of 1976 "A revised socioeconomic index for occu-
pations in Canada." Canadian Review of
the expectations of teachers in previous
Sociology and Anthropology 13:71-79.
years. Perhaps ethnicity, via teachers' ex- Bowles, S. and H. Gintis
pectations, has cumulative effects upon 1976 Schooling in Capitalist America. New
abilities, performances, and aspirations York: Basic.

throughout the students' academic careers Braun, Carl

1976 "Teacher expectation: Sociopsychological
(see West and Anderson, 1976), and some
dynamics." Review of Educational Research
significant portion of the effects from 46:185-213.
abilities, performances, and aspirations, Brehaut, Willard
as measured in this research, is an indirect 1964 A Preliminary Look at the Carnegie Students
in Grade 12 in Ontario Schools. Carnegie
effect that has previously resulted from
Study Bulletin No. 8, Department of Edu-
ethnicity. In fact, Rist (1970) suggests that cational Research, University of Toronto.
the indirect effects of ethnicity, through Breton, Raymond
teachers' expectations, are much greater 1972 Social and Academic Factors in the Career
in the first few years of school. Decisions of Canadian Youth. Ottawa: De-
partment of Manpower and Immigration.
From another perspective, the analyses Breton, Raymond and Howard E. Rosenborough
illustrate that ethnicity is quite strongly 1968 "Ethnic differences in status." In Bernard R.
related to intellectual ability, past perfor- Blishen, F. E. Jones, K. D. Naegele and J.
mances, academic aspirations, and stan- Porter (eds.), Canadian Society. Toronto:
Macmillan.
dardized achievement in both English and
Breton, Raymond, Jean Burnet, Norbert Hartmann,
mathematics. This suggests that ethnicity Wsevolod Isajiw, and Jos Lennards
has important direct and indirect effects 1975 "Research issues on Canadian cultures and
upon educational performances as mea- ethnic groups: An analysis of a conference."
sured by standardized achievement tests, Canadian Review of Sociology and An-
thropology 12:81-94.
hut the indirect effects are not, to any
Brophy, J. E. and T. L. Good
great extent, the result of teachers' ex- 1970 "Teacher communication of differential ex-
pectations. Rather, some of the effects are pectations for children's classroom perfor-
directed through other factors which may mance: Some behavioral data." Journal of
be to some degree endemic within the Educational Psychology 61:365-74.
1974 Teacher-Student Relationships: Causes and
subcultural backgrounds of the children.
Consequences. New York: Holt, Rinehart,
That is, children may be socialized within and Winston.
ethnic subcultures so that their abilities, Carnoy, M.
aspirations, and motivation are shaped to 1974 Education as Cultural Imperialism. New
fit into certain educational and occu- York: David McKay.

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms
 ETHNICITY AND CANADIAN ACADEMIC ACHIEVEMENT 301

Cooper, Harris M. transgressions." Journal of Educational

1979 "Pygmalion grows up: A model for teacher Psychology 70:911-15.
expectation communication and perfor- Merton, Robert
mance influence." Review of Educational 1957 Social Theory and Social Structure. Glen-
Research 49:389-410. coe: The Free Press.
D'Oyley, V. R. Parsons, Talcott
1964 The Canadian Tests, Grades 8, 9, and 10. 1959 "The school class as a social system: Some
Carnegie Study Bulletin No. 5, Department of its functions in American society." Har-
of Educational Research, University of To- vard Educational Review 29:297-318.
ronto. Porter, John
Dreeben, Robert 1965 The Vertical Mosaic. Toronto: University
1968 On What is Learned in School. Don Mills, of Toronto Press.
Ontario: Addison-Wesley. Porter, Marion R., John Porter, and Bernard R.
Duncan, Otis D. Blishen
1975 Introduction to Structural Equation Mod- 1973 Does Money Matter? Prospects for Higher
els. New York: Academic Press. Education. Toronto: Institute for Be-
Dusek, Jerome B. havioural Research, York University.
1975 "Do teachers bias children's learning?" Re- Richer, Steven
view of Educational Research 45:661-84. 1979 "Sex-role socialization and early school-
Feldman, R. S. and S. Orchowsky ing." Canadian Review of Sociology and
1979 "Race and performance of student as deter- Anthropology 16 (May):195-205.
minants of teacher nonverbal behavior." Rist, R. C.
Contemporary Educational Psychology 1970 "Student social class and teacher expecta-
4:324-33. tions: The self-fulfilling prophecy in ghetto
Finn, Jeremy D. education." Harvard Education Review
1972 "Expectations and the educational environ- 40:411-51.
ment." Review of Educational Research Rosenthal, Robert
42:387-410. 1976 Experimenter Effects in Behavioral Re-
Heise, David R. search. Enlarged Edition. New York: Irv-
1975 Causal Analysis. New York: Wiley. ington.
Kehle, T. J. Rosenthal, Robert and L. Jacobson
1974 "Teachers' expectations: Ratings of student 1968 Pygmalion in the Classroom. New York:
performance as biased by student charac- Holt, Rinehart, and Winston.
teristics." The Journal of Experimental Edu- Rosenthal, Robert and Donald B. Rubin
cation 43:54-60. 1978 "Interpersonal expectancy effects: The first
Kim, J. 0. and C. W. Mueller 345 studies." The Behavioral and Brain Sci-
1976 "Standardized and unstandardized coeffi- ences 3:377-86.
cients in causal analysis." Sociological Rubin, R. A. and B. Balow
Methods & Research 4:423-38. 1979 "Prevalence of teacher identified behavior
King, A. J. C. problems: A longitudinal study." Ex-
1968 "Ethnicity and school adjustment." Cana- ceptional Children 45 (October):102-11.
dian Review of Sociology and Anthropol- West, C. K. and T. H. Anderson
ogy 5:84-91. 1976 "The question of preponderance causation
Li, Peter S. in teacher expectancy research." Review of
1978 "The stratification of ethnic immigrants: Educational Research 46:613-30.
The case of Toronto." Canadian Review of Williams, Trevor
Sociology and Anthropology 15:31-40. 1972 "Educational aspirations: Longitudinal evi-
Maccoby, E. E. and C. N. Jacklin dence on their development in Canadian
1974 The Psychology of Sex Differences. Stan- youth." Sociology of Education 45
ford: Stanford University Press. (Spring): 107-33.
Marwit, K. L., S. J. Marwit, and E. Walker 1976 "Teacher prophesies and the inheritance of
1978 "Effects of student race and physical at- inequality." Sociology of Education 49
tractiveness on teachers' judgments of (July):223-36.

This content downloaded from 111.68.97.183 on Tue, 26 Feb 2019 09:32:38 UTC
All use subject to https://about.jstor.org/terms