You are on page 1of 10

Psychology in the Schools, Vol.

35(4), 1998
© 1998 John Wiley & Sons, Inc. CCC 0033-3085/98/040391-10

EDUCATIONAL CHARACTERISTICS OF STUDENTS


WITH HIGH OR LOW SELF-CONCEPT
ian hay
Griffith University, Brisbane

adrian f. ashman and christina e. van kraayenoord


The University of Queensland

In response to methodological concerns associated with previous research into the educational char-
acteristics of students with high or low self-concept, the topic was re-examined using a significantly
more representative sample and a contemporary self-concept measure. From an initial screening of
515 preadolescent, coeducational students in 18 schools, students significantly high or low in self-
concept were compared using standardized tests in reading, spelling, and mathematics, and teacher
interviews to determine students’ academic and nonacademic characteristics. The teachers were not
informed of the self-concept status of the students. Compared to students with low self-concept,
students with high self-concept were rated by teachers as being more popular, cooperative, and per-
sistent in class, showed greater leadership, were lower in anxiety, had more supportive families,
and had higher teacher expectations for their future success. Teachers observed that students with
low self-concept were quiet and withdrawn, while peers with high self-concept were talkative and
more dominating with peers. Students with lower self-concepts were also lower than their peers in
reading, spelling, and mathematical abilities. The findings support the notion that there is an inter-
active relationship between self-concept and achievement. © 1998 John Wiley & Sons, Inc.

Although it is a popularly held belief that low self-concept explains why students have diffi-
culty in school or vice versa, the research supporting this belief is equivocal (Black, 1991; Carlton,
1981; Frisby & Tucker, 1983; Hattie & Marsh, 1996; Swann, 1996). The significance of self-concept
in education has been particularly influenced by Coopersmith’s (1967) research on the educational
characteristics of students with high or low self-concept. He concluded that when students with high
self-concept were compared to their low self-concept peers they were more involved in classroom
discussions, more confident, more popular with peers, more persistent in performing tasks, more con-
cerned about public affairs rather than personal problems, less afraid to disagree, less sensitive to
criticism, less passive, and less anxious, and that they set themselves higher long-term goals, relat-
ed better to their parents, expressed their own opinions more.
Coopersmith’s research, although systematic and based on detailed interview and questionnaire
procedures has methodological weaknesses. He focused on “normal” boys from White middle-class
families (Robinson, 1990). Hence, his findings are unlikely to generalize across more heterogeneous
populations. Robinson (1990) claimed that Coopersmith’s self-concept research was a “portrait . . .
stereotype of the all-American high school boy (Caucasian) . . . [and] . . . high scores may indicate
no more than conformity to a predefined ideal” (p. 6). Furthermore, the reliability of Coopersmith’s
Self-Esteem Inventory has been criticized. Wylie (1974) stated that the “development of this inven-
tory and the amount of available information about it do not make it an instrument of choice for self-
concept research on children” (p. 174).
It has been argued, however, that the influence of high or low self-concept on education will be
further clarified when researchers use measures of self-concept that are multidimensional and psy-
chometrically stronger than Coopersmith’s Self-Esteem Inventory (Hattie & Marsh, 1996). Hattie
and Marsh maintained that researchers need to explore the dynamics of self-concept and a range of
education variables, particularly students’ learning, behavior, and self-concept. Osborne (1996) also
urged researchers to revisit the educational characteristics of students with high or low self-concept
to identify the significance of self-concept on students’ development.

Requests for reprints should be sent to Ian Hay, Faculty of Education, School of Cognition, Learning, and Special Edu-
cation, Griffith University, Nathan, Brisbane, Queensland, Australia, 4111. E-mail: I.Hay@edn.gu.edu.au

391
392 Hay, Ashman, and van Kraayenoord

Coopersmith’s (1967) assertion that students with high self-concept scores were associated with
positive educational and psychological characteristics has also been re-examined by researchers in-
vestigating students’ response style on self-concept questionnaires (Bachman & O’Malley, 1984;
Baumeister, Tice, & Hutton, 1989; Campbell & Fehr, 1990). These investigators have claimed that
individuals with low self-concept scores simply reflect a more cautious conservative self-presenta-
tional style than any emotional problem. In addition, high self-concept scores are associated with in-
dividuals who have a more assertive, outspoken, ambitious, and rigid personality. The desire of in-
dividuals to report extremely high self-concept scores has been called a form of narcissism (Gecas,
1982; Paulhus, 1984; Raskin & Terry, 1988) and described as “a pattern of grandiosity used to bol-
ster and enhance a fragile sense of self-esteem” (Raskin & Novacek, 1989, p. 66). More recently,
Gibbons and McCoy (1991) found that persons with high self-concept were more likely to denigrate
others than their low self-concept counterparts. Osborne (1996) has suggested that the difference be-
tween individuals with high or low self-concept may reflect a difference in their ability to self-
appraise, producing differences in their ability to self-assess performance accurately, self-define and
self-describe, maintain positive self-evaluations, and protect their ego. Osborne’s notions about high
and low self-concept, in part, challenge Coopersmith’s (1967) almost simplistic notion that high
self-concept is “good” and low self-concept is “bad.”
The aim of this study was to reexamine Coopersmith’s assertion that students with high self-
concept have more positive characteristics in the domains of cooperation, persistence, leadership,
anxiety, expectations for future schooling, family support, behavior in class, and peer interactions
than those students with low self-concept. This re-examination was based on teacher interview data
and the use of a contemporary measure of self-concept, the Self-Description Questionnaire-1 (SDQ-
1; Marsh, 1988). The SDQ-1 assesses self-concept in four nonacademic areas (Physical Abilities,
Physical Appearance, Peer Relations, and Parent Relations), three academic areas (Reading, Math-
ematics, and General School) as well as General Self, and derived from these subscales is an over-
all Total-Self score. A second aim of this study was to compare the educational characteristics of stu-
dents with high or low self-concept on standardized tests of reading, spelling, and mathematics.
Method
Participants
A sample of 515 Year (Grade) 5 coeducational students was drawn from 18 metropolitan and
rural schools located in the greater Brisbane (Australia) area. The mean age of the screening sample
was 120.6 months (SD ⫽ 4.32 months; range 113 months–135 months). Schools were randomly se-
lected from all government and nongovernment schools in the area to reflect a range of socioeco-
nomic and size of school settings. The socioeconomic status (SES) of the 18 schools was estimated
using the Australian Occupational Prestige Scale (Daniel, 1983). This ranks employment status us-
ing a one to seven scale (the lower the number the higher the prestige rating). The schools’ SES rank-
ings were estimated by averaging individual students’ parent occupational ratings. The SES mean
for all schools was 4.9 (SD ⫽ .76, range: 3.4–6.1). Based on the obtained data and the suggested
Daniel (1983) cut off points of less than 4.2 as high SES, 4.2 to 5.5 as middle SES, and above 5.6 as
low SES, it was estimated that four schools were high SES, eleven middle SES, and three low SES.
Participants were identified as either high or low in self-concept on the basis of their Total-Self
score out of 40, on the Self-Description Questionnaire-1. Marsh (1988) stated that because of the
significant skewness in the distribution pattern towards high self-concept scores only SDQ-1 scores
above the 75th percentile should be considered high and only scores less than the 25th percentile
should be considered low. To ensure in this study that only students with high or low self-concepts
were compared the more stringent cut off points were used, that is, less than the 15th percentile for
the low self-concept group and above the 85th percentile for the high self-concept group.
High or Low Self-Concept 393

Students with high or low self-concept were compared for differences in academic abilities in
reading, mathematics, and spelling. From this group with high or low self-concept a further sub-
population of students were selected for teacher interviews. Across the 18 schools (24 classes in all)
the highest and lowest scoring boy and girl from each Year 5 class were selected for teacher inter-
views.

Instruments
Self-Description Questionnaire-1 (SDQ-1). This 76-item test was developed by Marsh (1988)
to measure self-concept in four nonacademic areas (Physical Abilities, Physical Appearance, Peer
Relations, and Parent Relations), three academic areas (Reading, Mathematics, and General School)
and General Self. Children are asked to read declarative sentences (e.g., “I’m good at mathematics.”
“I make friends easily.”) and select one of five alternative responses: False; Mostly False; Sometimes
False/Sometimes True; Mostly True; or True. The norms are based on a sample of 3,562 New South
Wales (Australia) students. The raw scores are converted to stepped percentile and t scores, with the
SDQ-1 reporting combined and separate male and female norms. On the separate gender norms,
compared to boys, for the same raw score, girls’ SDQ-1 Total-Self percentiles were often 1 to 3 points
higher. This indicates that the boys in the norming sample tended to record slightly higher Total-Self
scores than the girls. Marsh reported an internal reliability coefficient of .92 for the full scale score
and a test–retest reliability coefficient of .87 over a 6 month period. The SDQ-1 was administered in
a class group situation, with the presentation taking from 10 to 15 minutes. The students’ raw scores
and separate gender percentile levels were used in this study.
Tests of Reading Comprehension (TORCH). This test was developed by Mossenson, Hill, and
Masters (1987) as a series of untimed reading passages for students in Years 3 to 10 and contains 14
passages (either a short story or a descriptive article) of about one page in length. Students are pre-
sented with a single passage of text together with a retelling of the passage in different words in the
form of a cloze activity that contains gaps corresponding to details in the original text. In this study,
one passage was used which was age-appropriate for Year 5 students. The authors reported a
test–retest reliability coefficient of .91 to .93 for the different TORCH passages. From students’ raw
scores, comparative TORCH scaled scores can be obtained. Students’ TORCH scaled scores were
recorded for this study.
The South Australian Spelling Test. This test was developed by Westwood (1979) to identify
students’ spelling achievement and progress on a representative sample of primary (elementary)
school words. The list contains 70 words of increasing difficulty. The administration involved the
tester saying the word, placing the word in a sentence, saying the word again and then having the
students write the word. Westwood reported a test–retest reliability coefficient of .96 after 1 week.
Raw scores were recorded for this research.
Test of Whole Number Computations (TWNC). This is a sequenced test of four basic mathe-
matics operations; addition, subtraction, multiplication, and division (ACER, 1979). The test con-
sists of a booklet containing 39 items of increasing difficulty. Students record their workings in the
booklet as they progress through the test and, after 30 minutes, they are asked to stop, even if they
have not finished all items. The TWNC has a test–retest reliability coefficient of .86 over a 6 to 8
week period. The test was administrated in class groups and raw scores were recorded.

Teacher Questionnaire
A 9-item teacher questionnaire was developed to collect educational characteristics data on stu-
dents’ with very high and very low self-concept. The four open-ended questions were: How would
you describe the student’s behavior in class? How would you describe the student’s interactions with
394 Hay, Ashman, and van Kraayenoord

peers? How would you describe the student’s home life? How do you think the student will achieve
in high school? The five rating questions were on a 1 to 10 scale, where 10 is the highest, how would
you rate the student for: (a) popularity with peers (b) cooperation within the classroom (c) persis-
tence with school work (d) leadership within the class and (e) anxiety? To estimate the stability and
reliability of the teachers’ responses, three weeks after the initial interviews nine of the same teach-
ers were asked to re-rate the last five questions. Twenty of the same students were re-rated involv-
ing equal numbers (five each) of boys and girls with the highest and lowest self-concept. The
test–retest correlation coefficients were peer popularity (r ⫽ .92), cooperation (r ⫽ .86), persistence
(r ⫽ .82), leadership (r ⫽ .92), and anxiety (r ⫽ .90). This supports the notion that the teachers main-
tained a consistent perspective for the same students with very high or very low self-concept.

Procedures
The four tests were administered according to procedures outlined in the manuals, in class
groups, involving two administration sessions of about 30 minutes. In session one, the TORCH and
Westwood spelling tests were administered and a week later the SDQ-1 and the TWNC tests were
administered. A very small number of students identified as less able readers by their teachers had
the self-concept test read to them. One month after the SDQ-1 data were collected in the 18 schools,
students from participating schools were identified as high or low in self-concept on the basis of their
SDQ-1 results. Teacher rating data were collected using the teacher questionnaire (four open-ended
questions; five rating questions). This instrument was administered to 7 male and 14 female Year 5
teachers. The teachers were not informed of the self-concept status of the selected students. The ques-
tions were read to the teachers and their responses recorded.

Results
For the low self-concept group the Total-Self raw score cut off was set at 25 or less (11th per-
centile for girls and 12th percentile for boys), and for the high self-concept group the Total-Self raw
score cut off was set at 36 and above (86th percentile for girls and 92th percentile for boys). This re-
sulted in a low self-concept group of 82 (girls n ⫽ 44, Total-Self M ⫽ 22.3, SD ⫽ 3.06; boys
n ⫽ 38, Total-Self M ⫽ 21.8, SD ⫽ 3.05) and a high self-concept group of 98 (girls n ⫽ 43, Total-
Self M ⫽ 37.4, SD ⫽ 1.01; boys n ⫽ 55, Total-Self M ⫽ 37.06, SD ⫽ 1.15). A multivariate F test
demonstrated a significant overall difference for students with high or low self-concept in reading,
spelling, and mathematics test results, Hotelling’s F (3, 176) ⫽ 16.06, p ⬍ .001. Univariate F tests
identified that students with high self-concept outperformed their peers with low self-concept in read-
ing, spelling, and mathematics (Table 1).
For the class teacher interviews, of the 98 students with high self-concept across the 18 schools
(24 classes in all), 44 students were identified as the highest in self-concept in their class (22 girls,
22 boys; SDQ-1 Total-Self-Concept Percentile M ⫽ 94.64, SD ⫽ 4.25) and of the 82 students with
low self-concept, 47 students were identified as the lowest in self-concept in their class (24 girls, 23
boys; SDQ-1 Total-Self-Concept Percentile M ⫽ 5.87, SD ⫽ 4.17). The small variation in group size
was due to the inclusion in the sampling of three composite classes with mixed year (grade) levels.
The fewer Year 5 students in these classes meant that there were not always a boy or girl who met
the 15 percentile high or low criteria. The first four teacher interview questions were analyzed for
frequency of theme (idea) statements (see Weber, 1985). If teachers gave more than one statement
about a student only the first statement was selected for analysis to avoid “double counting” of
that student (see Hays, 1994). Statements were categorized as either positive, negative, or neutral,
and the frequency data tabulated by group and the ␹2s calculated for each question. The text analy-
ses (Table 2) illustrate that teachers identified students with high self-concept to be better behaved
High or Low Self-Concept 395

Table 1
Comparison Between Students With High or Low Self-Concept for Standardized Tests
of Reading, Spelling, and Mathematics
High Self-Concept (n ⫽ 98) Low Self-Concept (n ⫽ 82)
M SD M SD Univariate F df: 1, 178

Reading 44.76 12.13 31.97 13.57 44.52***


Spelling 38.47 4.89 33.11 7.89 31.07***
Mathematics 24.99 5.56 21.56 5.55 17.01***
***p ⬍ .001.

in class, more popular with peers, have more supportive families, and have higher teacher expecta-
tions for future schooling, than their peers with low self-concept.
Teachers’ responses for the five rating questions (popularity, cooperation, persistence, leader-
ship, and anxiety) for students with the highest or lowest self-concept produced a significant multi-
variate result, Hotelling’s F ⫽ 30.93 (5, 85), p ⬍ .001. As shown in Table 3, the very high self-
concept scorers were identified by their teachers as being more popular in school, cooperative in
class, persistent with school work, having higher levels of leadership, but lower levels of anxiety
than the very low self-concept scorers.

Discussion
Based on the comparison of students’ level of self-concept and their standardized test results a
relationship between academic achievement and self-concept is indicated. Although not demon-
strated with this study, it is speculated that this is a cyclical relationship with academic achievement
influencing academic self-concept, which influences students’ level of academic effort, which in turn
influences achievement (Borkowski, 1996; Chapman, Lambourne, & Silva, 1990; Hay, 1995; Hay,
Ashman, & van Kraayenoord, 1994, 1997; Short & Weissberg–Benchell, 1989). The notion of a
cyclical relationship suggests that for students who achieve positive academic feedback the cycle is
more positive and self-enhancing; however, for students receiving more negative academic feedback,
the cycle would be more of a downward spiral of low levels of self-concept leading to lower levels
of effort, and lower levels of academic performance.
Teachers reported that students with high self-concept had more positive characteristics in the
domains of cooperation, persistence, leadership, anxiety, expectations for future schooling, family
support, behavior in class, and peer interactions than students with low self-concept. The findings
are consistent across quantitative and more qualitative research techniques. The present study goes
beyond merely replicating Coopersmith’s (1967) influential research but updates it by using an im-
proved methodology, a significantly more representative sample of students, and a self-concept
screening instrument that has received praise for its strong psychometric and theoretical construct
characteristics that includes academic and nonacademic dimensions (Boyle, 1994; Hattie, 1992). The
suggestion of Robinson (1990) that a high self-concept score may indicate no more than conformi-
ty to a predefined teacher ideal is less supported by this investigation’s findings, because the teachers
did not know the self-concept status of the students and research by Marsh, Parker, and Smith (1983)
and Morvitz and Motta (1992) proposed that teachers are not very accurate in identifying the self-
concept status of their students.
396 Hay, Ashman, and van Kraayenoord

Table 2
Teacher Statements of Students With Highest or Lowest Self-Concepts for Behavior in Class,
Interactions With Peers, Home Life, and Expectations for High School
Behavior in Class
Frequency
Teachers’ Theme Responses Classification Highest Self-Concept Lowest Self-Concept

1. Good ⫹ 16 3
2. Academically oriented ⫹ 12 5
3. Cooperative ⫹ 2 1
4. Changes sometimes good/bad N 4 3
5. Quiet, withdrawn N 1 9
6. Talkative, attention seeking — 4 1
7. Shows little attention to work — 1 8
8. Not academically oriented — 1 4
9. Lazy — 1 2
10. Aggressive, bad temper — 1 2
11. Needs assistance — 1 4
12. Bad — 0 5

Frequency of Teachers’ Statements


Self-Concept Positive Neutral Negative Total N ␹2 (df ⫽ 2)

High 30 5 9 44 24.59***
Low 9 12 26 47 10.51**
␹2 (df ⫽ 1) 18.88*** 7.21** 9.01** 91 Total ␹2 35.51***

Peer Interactions
Frequency
Teachers’ Theme Responses Classification Highest Self-Concept Lowest Self-Concept

1. Good, popular ⫹ 16 1
2. Liked ⫹ 6 8
3. Wide (large) circle of friends ⫹ 5 0
4. Good at Sports + 4 2
5. Few friends N 4 13
6. Bossy, dominating — 7 1
7. Does not get on with others — 1 9
8. Teases, fights — 1 1
9. Loner — 0 2
10. Not a leader — 0 10

Frequency of Teachers’ Statements


Self-Concept Positive Neutral Negative Total N ␹2 (df ⫽ 2)

High 31 4 9 44 28.14***
Low 11 13 23 47 5.28
␹2 (df ⫽ 1) 19.6*** 8.2** 5.62* 91 Total ␹2 33.42***

continued
High or Low Self-Concept 397

Table 2
Continued

Home Life
Frequency
Teachers’ Theme Responses Classification Highest Self-Concept Lowest Self-Concept

1. Parents supportive, cooperative ⫹ 14 1


2. Parents are caring ⫹ 9 3
3. Good ⫹ 6 1
4. Parents concerned, interested ⫹ 4 4
5. Divorced, separated N 5 5
6. English a second language in home N 1 2
7. Aboriginal family N 1 1
8. Seems OK N 0 3
9. Indulgent — 2 4
10. Unable to help the student academically — 1 1
11. Parents are dominating — 1 3
12. Little support — 0 11
13. Finds child difficult to manage at home — 0 4
14. Suspected sexual or physical abuse — 0 4
in the home

Frequency of Teachers’ Statements


Self-Concept Positive Neutral Negative Total N ␹2 (df ⫽ 2)

High 33 7 4 44 34.69***
Low 9 11 27 47 12.43**
␹2 (df ⫽ 1) 25.77*** 5.39** 15.96*** 91 Total ␹2 47.12***

Expectations for High School


Frequency
Teachers’ Theme Responses Classification Highest Self-Concept Lowest Self-Concept

1. Good ⫹ 26 2
2. OK, all right ⫹ 12 9
3. Unsure ⌵ 4 7
4. Not cope academically — 1 12
5. Not cope socially — 1 10
6. Easily led — 0 7

Frequency of Teachers’ Statements


Self-Concept Positive Neutral Negative Total N ␹2 (df ⫽ 2)

High 38 4 2 44 55.84***
Low 11 7 29 47 17.54***
␹2 (df ⫽ 1) 38.55*** 12.54*** 22.29*** 91 Total ␹2 73.38***
Classification Key: ⫹⫽ positive; ⫺ ⫽ Negative; N ⫽ Neutral.
***p ⬍ .001. ** p ⬍ .01. * p ⬍ .05.
398 Hay, Ashman, and van Kraayenoord

Table 3
Comparison Between Students With Highest or Lowest Self-Concept for Teachers’ Ratings
of Students’ Popularity, Cooperation, Persistence, Leadership, and Anxiety
High Self-Concept (n ⫽ 44) Low Self-Concept (n ⫽ 47)
Univariate F test
M SD M SD (df: 1,89)

Popularity with Peers 7.9 1.63 5.0 1.82 64.02***


Cooperation 8.1 1.76 5.8 2.08 35.20***
Persistence 8.0 1.72 4.9 2.23 51.31***
Leadership 7.5 1.63 3.3 1.84 135.68***
Anxiety 4.7 1.95 6.4 1.96 18.08***
***p ⬍ .001.

Looking at the response data, teachers reported that students with high self-concept were aca-
demically orientated and were more interested in school than their peers with low self-concept.
Teachers observed that students with low self-concept were considered quiet and withdrawn in class,
while those with high self-concept were talkative, bossy, and dominating with peers. These obser-
vations lend support to the research of Campbell (1990) and Baumeister et al. (1989) that self-
concept scores, in part, reflect self-presentation styles, and low scorers are more cautious individu-
als. In the current study, teachers also observed that students with high self-concept received greater
parent support than low self-concept students. This is consistent with Harter’s (1990, 1996) view that
high support from parents increased students’ self-concept.
Teachers suggested a link between low self-concept and suspected sexual or physical abuse in
the home. This would need confirmation. However, as the SDQ-1 identifies a range of academic and
nonacademic self-dimensions, the test may have value as a screening instrument that school coun-
sellors might use to identify possible “at risk” students. Finally, conflicts within the family (such as
separation and divorce), were not associated with high or low self-concept. This is consistent with
previous research which has suggested that separation and divorce do not necessarily produce long-
term negative influences on self-concept (Marsh, 1990; Partridge & Kotler, 1987).
Although Coopersmith’s research has been criticized on methodological grounds the thrust of
his investigation, that students with high self-concepts are associated with more positive psycho-
logical and educational variables is generally supported by this research. The teachers’ comments
do, however, recommend possible directions for intervention based programs for students with low
self-concept. The lack of popularity, poor peer interaction skills, and expected failure to cope socially
with future education suggest that social skills enhancement could be considered, along with activ-
ities that develop friendships and cooperative networks within the classroom. The findings that stu-
dents with low self-concept were easily led, withdrawn, and showed little leadership indicate their
assertion and decision making skills need development. Furthermore, the reduced persistence with
school work, expected difficulty with future academic schooling, and higher levels of anxiety are
concerns that signal that this population will need greater access to counselling, guidance, and sup-
port services within the school and community.

References
Australian Council for Educational Research. (1979). ACER Mathematics Test (AM Series) Revised Manual.
Melbourne: Author.
Bachman, J. G., & O’Malley, P. M. (1984). Black-White differences in self-esteem: Are they due to differences in
response style? American Journal of Sociology, 90, 624–639.
High or Low Self-Concept 399

Baumeister, R. F., Tice, D. T., & Hutton, D. G. (1989). Self-presentational motivations and personality differences in
self-esteem. Journal of Personality, 57, 547–579.
Black, S. (1991). Self-esteem sense and nonsense. The American School Board Journal, July, 179, 27–29.
Borkowski, J. G. (1996). Metacognition: Theory or chapter heading? Learning and Individual Differences, 8, 391–402.
Boyle, G. J. (1994). Self-Description Questionnaire II: A review. Test Critiques, 10, 632–643.
Campbell, J. D., & Fehr, B. (1990). Self-esteem and perceptions of conveyed impressions: Is negative affectivity associ-
ated with greater realism? Journal of Personality and Social Psychology, 58, 122–133.
Carlton, R. A. (1981). Self-concept research: The gap between social theory and educational practice. Canadian Journal
of Education, 6, 77–91.
Chapman, J. W., Lambourne, R., & Silva, P. A. (1990). Some antecedents of academic self-concept: A longitudinal study.
British Journal of Educational Psychology, 60, 142–152.
Coopersmith, S. (1967). The antecedents of self-esteem. San Francisco, CA: W. H. Freeman.
Daniel, A. (1983). Power, privilege and prestige: Occupations in Australia. Melbourne: Longman Cheshire.
Frisby, C. L., & Tucker, C. M. (1993). Black children’s perception of self: Implications for educators. The Educational
Forum, 57, 146–156.
Gecas, V. (1982). The self-concept. Annual Review of Sociology, 8, 1–33.
Gibbons, F. X., & McCoy, S. B. (1991). Self-esteem, similarity, and reactions to active versus passive downward compar-
isons. Journal of Personality and Social Psychology, 60, 414–424.
Harter, S. (1990). Processes underlying adolescent self-concept formation. In R. Montemayor, G. R. Adams, & T. P. Gul-
liton (Eds.), From childhood to adolescence: A transition period (pp. 205–239). Newbury Park, CA: Sage.
Harter, S. (1996). Historical roots of contemporary issues involving self-concept. In Bracken B. A. (Ed.), Handbook of
self-concept: Developmental, social, and clinical considerations (pp. 1–38). New York: Wiley.
Hattie, J. A. (1992). Self-concept. Hillsdale, NJ: Erlbaum.
Hattie, J. & Marsh, H. W. (1996). Future directions in self-concept research. In Bracken B. A. (Ed.), Handbook of self-
concept: Developmental, social, and clinical considerations (pp. 421–462). New York: Wiley.
Hay, I. (1995). Enhancing the learning of students with learning difficulties through an understanding of the interactions
between students’ effort, self-perception, achievement, and teacher actions. Australian Journal of Remedial Education,
27, 22–25.
Hay, I., Ashman, A., & van Kraayenoord, C. (1994). Children’s self-perception and thinking: How are they related? In
J. Edwards (Ed.), Thinking: International interdisciplinary perspectives (pp. 259–268). Cheltenham, Victoria: Hawker
Brownlow Education.
Hay, I., Ashman, A., & van Kraayenoord, C. (1997). Investigating the influence of achievement on self-concept using
an intra-class design and a comparison of the PASS and the SDQ-1 self-concept tests. British Journal of Educational
Psychology, 67, 311–321.
Hays, W. L. (1994). Statistics (5th ed.). Fort Worth, TX: Harcourt Brace College.
Marsh, H. W. (1988). Self-Description Questionnaire-1 SDQ-1: Manual and research monograph. San Antonio, TX: The
Psychological Corporation.
Marsh, H. W. (1990). Two-parent, step-parent and single-parent families: Changes in achievement, attitudes and behav-
iors during the last two years of high school. Journal of Educational Psychology, 82, 327–340.
Marsh, H. W., Parker, J. W., & Smith, I. D. (1983). Preadolescent self-concept: Its relation to self-concept as inferred by
teachers and to academic ability. British Journal of Educational Psychology, 53, 60–78.
Morvitz, E., & Motta, R. W. (1992). Predictors of self-esteem: The roles of parent–child perceptions, achievement, and
class placement. Journal of Learning Disabilities, 25, 72–80.
Mossenson, L., Hill, P., & Masters, G. (1987). Tests of Reading Comprehension (TORCH). Melbourne: Australian Coun-
cil for Educational Research.
Osborne, R. E. (1996). Self: An eclectic approach. Boston: Allyn & Bacon.
Partridge, S., & Kotler, T. (1987). Self-esteem and adjustment in adolescents from bereaved, divorced and intact fami-
lies: Family type versus family environment. Australian Journal of Psychology, 39, 223–234.
Paulhus, D. L. (1984). Two-component models of socially desirable responding. Journal of Personality and Social Psy-
chology, 46, 596–609.
Raskin, R., & Novacek, J. (1989). An MMPI description of the narcissistic personality. Journal of Personality Assessment,
53, 66–80.
Raskin, R., & Terry, H. (1988). A principal components analysis of the Narcissistic Personality Inventory and further
evidence of its construct validity. Journal of Personality and Social Psychology, 54, 890–902.
Robinson, W. P. (1990). Academic achievement and self-esteem in secondary school: Muddles, myths, and reality. Edu-
cation Research and Perspectives, 17, 3–21.
400 Hay, Ashman, and van Kraayenoord

Short, E. J., & Weissberg–Benchell, J. A. (1989). The triple alliance for learning: Cognition, metacognition, and moti-
vation. In C. B. McCormick, G. Miller, & M. Pressley (Eds.), Cognitive strategy research: From basic research to edu-
cational applications (pp. 33–63). New York: Springer–Verlag.
Swann, W. B. (1996). Self-traps: The elusive quest for higher self-esteem. New York: Freeman & Co.
Weber, R. (1985). Basic content analysis. Beverly Hills, CA: Sage.
Westwood, P. S. (1979). Helping children with spelling difficulties. Adelaide: South Australian Department of Education.
Wylie, R. C. (1974). The self-concept: Vol. 1. A review of methodological considerations and measuring instruments.
Lincoln, NE: University of Nebraska Press.

You might also like