Self-Concept and Academic Achievement of Central and Western European Groups of Adolescents.

Full text

SELF-CONCEPT AND ACADEMIC ACHIEVEMENT OF CENTRAL AND WESTERN
EUROPEAN GROUPS OF ADOLESCENTS
Darja Kobal-Palčič and Janek Musek
Abstract
According to the accumulated research evidence, the self-concept is mutually connected
with academic achievement: it is influenced by different aspects of school life including academic
success, and vice versa, the self-concept itself may substantially contribute to the academic
aspirations and accomplishments. In teen- and adolescent years, the interdependence between
academic successfulness and self-concept becomes increasingly important due to the fact that
both, academic achievement and the formation of the self-concept play a growing role in this
period of individual's life. Interestingly, a number of experimental studies suggested that the
relationship between academic success and self-concept could significantly differ according to the
national differences (with possible background influence of specific curricula).
In the present study, we tried to verify the hypothesis that academic achievement affects
different components of self-concept including self-esteem. Further, we tried also to investigate
the possible role of national differences in the relationship between academic achievement and
self-concept. For this reason, we included into investigation a Slovene and a French sample of
subjects. The results of two-factor (academic achievement x nationality) analyses of variance and
discriminant analyses showed significant effects on self-concept for both factors or groups. The
French subjects exceeded Slovenes in some domains of self-concept (verbal, academic, relations
with same sex peers, relations with parents, religion and spirituality, and general self-concept),
while Slovene subjects exceeded French subjects in the domain of problem solving and creativity.
There was no significant difference between both national samples in self-esteem. Also, the French
subjects exceeded Slovene pupils in general academic achievement. The results were interpreted
on the grounds of theoretical expectations related to the formation of self-concept and academic
achievement as well as on the basis of national differences in the school system and personality
structure.
1

SELF-CONCEPT AND ACADEMIC ACHIEVEMENT OF CENTRAL AND
WESTERN EUROPEAN GROUPS OF ADOLESCENTS
INTRODUCTION
Over the last twenty years teachers and researchers have become increasingly
interested in the role of adolescents' self-concept in institutional education. Numerous
psychological studies have shown that the formation of a stable and a positive self-concept is one
of the major developmental challenges of adolescence (Bohan, 1973; Keltikangas-Järvinen, 1990;
Watkins, McCreary Juhasz, Walker & Janvlaitiene, 1995). Undoubtedly, the school as a very
fundamental agent of socialisation has an important influence on the development of the self-
concept in the age of adolescence.
The formation of self-concept in the context of institutional education is a rather complex
process evoking several questions interesting for psychological research. According to many
observations and a considerable amount of experimental evidence, the academic achievement
could be a prominent variable connected to the formation of the self-concept in adolescence.
Therefore, we should concentrate on the relationship between self-concept and academic
achievement as a relationship, which is considered as salient for individual and for social
development of each adolescent.
Self-concept
Self-concept as a theoretical term has numerous synonyms and numerous definitions. In
the literature it is also identified sometimes as self-schema (Cross & Markus, 1994; Lewicki,
1984; Markus, 1977; Markus & Wurf, 1987), self-representation (Cross and Markus, 1994), self-
image (Offer et al., 1988), self-perception (Evans & Poole, 1991), self-esteem (Rosenberg,
1965), self-evaluation etc. In practice, these terms have been inconsistently used, sometimes
distinctively in scope or extension and sometimes synonymously. It is really hard to distinguish
between different psychological self-generalisations building the comprehensive conception of
2

ourselves. For example, we can not say that what we generally feel in a particular moment is our
self-image, but what we feel in the other moment is our self-evaluation. For avoiding
confusion, we will use the term self-concept in its general, prevailingly cognitive notion, keeping
some other terms in a more narrow sense referring to the cognitive or emotional or both aspects
of the self-concept (e.g. self-schema, self-representation, self-image self-evaluation, self-esteem).
Broadly defined, self-concept consists of our perceptions of ourselves (Shavelson &
Bolus, 1982). According to Burns, self-concept is a psychological entity which includes our
feelings, evaluations and attitudes as well as descriptive categories of ourselves (Burns,
1979). It is manifested outwardly by our behavioral and personality traits and inwardly by how
we feel about ourselves and the world around us (Maccoby, 1980). As a psychological whole,
it has an effect on one's particular self-perceptions and on perceptions of other people. It also
regulates social cognition, academic achievement, attitudes to school etc. (Keltikangas-
Järvinen, 1990).
It could be further concluded that self-concept is a cognitive generalisation about the
self (Cross & Markus, 1994) which mostly includes the self-descriptions of neutral values. On
the other hand, the evaluative and emotional generalisations about the self could be defined as
self-esteem (Lamovec, 1994). According to Rosenberg, self-esteem is a positive or negative
attitude of individual towards himself. It is "... closely connected to feeling of life
satisfaction…" (Rosenberg, 1985, pp. 212).
Self-concept researchers have long been interested whether the structure of self-
concept is mono- or multidimensional. To date, many theoretical models of self-concept have
been proposed.
Shavelson and Bolus (1982) postulated a model which has "undergone the most
rigorous examination in both cross-sectional and longitudinal designs," (Byrne & Shavelson,
1986, pp. 474). Their basic hypothesis was that the structure of self-concept is multidimensional
and hierarchical. General self-concept is at the highest and most general level of the structure. On
the next level, it splits into conceptions about self in academic and nonacademic areas. These
areas could be further divided into corresponding subareas - for example English, History,
peers, significant others self-concepts etc. Finally, at the most basic level, the model contains
3

specific evaluations of behaviour in specific situations.
In 1984, the Self-Description Questionnaire was constructed based on Shavelson’s
model, a psychodiagnostic device aimed to measure different areas of self-concept (Marsh &
O'Neill, 1984). The instrument, specially designed for the adolescents, contains the following self-
concept areas: (1) mathematics, (2) verbal, (3) academic, (4) problem solving/creativity,
(5) physical abilities/sports, (6) physical appearance, (7) relations with same sex peers, (8)
relations with opposite sex peers, (9) relations with parents, (10) religion, (11)
honesty/reliability, (12) emotional stability /security, and (13) general self-concept.
According to the results of many cross-cultural studies of self-concept in adolescence
(Robinson, Tayler & Piolat 1990) the structure of students' self-concept may vary both
intraculturally (from low to high academic achievers within one national educational system) and
interculturally (from low to high academic achievers within a frame of different countries).
Academic achievement
Two large groups of definitions could be mentioned concerning the academic
achievement. The first group could be denoted as an objective one because it refers to marks of
pupil's knowledge, which measure the degree of pupil's adaptation to school work and school
system (Gbati, 1988). The second group is supposed to be a subjective or a psychological one.
According to Khadivi-Zand (1982) academic achievement is defined as self-perception and
self-evaluation of objective academic success. Therefore, academic achievement consists of
student's attitudes towards his academic achievement and himself as well as attitudes of his
significant others (parents, teachers,...) towards his success and himself.
Academic achievement and self-concept
Somewhat surprisingly, very substantial correlations between academic achievement and
self-concept were rarely reported, and from the majority of relevant studies it could only be
concluded that self-concept is related to academic achievement (Marsh, 1990). There is no clear
and convincing scientific evidence of causative relations between self-concept and academic
achievement. Nevertheless, it seems plausible to hypothesise that the relationship between both
4

variables could be mutual one: the academic achievement could influence the self-concept
(especially self-esteem and self-evaluation) and vice-versa, the differences in self-concept could
stimulate or destimulate the achievement efforts.
A number of investigations indicated lower correlation between academic achievement and
self-esteem (or general self-concept) and also between academic achievement and domain-
specific self-concept (Marsh, 1990; 1994; Watkins & Gutierrez, ). House (1993) examined the
relationship among five areas of academic self-concept and academic achievement of students at
the start of study at an U.S. university and their continued enrollment at university for four
years. The results showed that the "students' self-concept of their overall academic ability was
the single most significant predictor of subsequent school withdrawal," (House, 1993, 127).
Marsh and O'Neill (1984) founded that achievement in mathematics of Australian
students was most highly correlated with math self-concept (0.58; p<0.01), less highly
correlated with academic self-concept (0.27; p<0.01), and it wasn't correlated with verbal and
nonacademic self-concept. School certificate scores in English were correlated with verbal self-
concept (0.42; p<0.01), academic self-concept (0.24; p<0.01), math self-concept (0.19;p<0.01)
and with problem solving self-concept (0.17; p<0.01). They were non correlated with other
self-concept scales of Self Description Questionnaire SDQ III.
In a cross-cultural study, Robinson and Breslav (unpublished paper) founded no
significant differences between English and Latvian groups of students related to academic
achievement and self-concept. They speculated that 13 or 14 year old Latvian participants
did not appear to have been chronically affected by the years of Soviet domination. In a study
of low academic achievers in three countries Robinson and Tayler (1989) founded that
Japanese pupils showed the culturally expected signs of modesty on self-concept areas,
especially on self-esteem (Markus & Kitayama, 1991).
We can speculate that the unconvincing correlation between academic achievement and
general self-concept may be due to the differential and possibly counteracting relationships
between academic achievement and different components of self-concept. For instance, it seems
imaginable that positive correlation between academic achievement and academic self-concept
could be counterbalanced by negative correlation between academic achievement and sexual self-
5

concept leading thus to the zero effect on general self-concept. In the other hand, the differences
between different cultures with respective academic tradition and organisation remained
remarkable. Thus, despite the fact that the relationship between self-concept and academic
achievement has been the subject of a number of psychological studies, a further systematic
research on this topic is still needed. We designed therefore a cross-cultural study in order to
clarify some unresolved questions concerning the relationship between academic achievement and
self-concept. The main purpose of the present study was to examine on the multicultural level
the connection between academic achievement and self concept in its general and specific
domains. Consequently, we hypothesised that (a) differences in the levels of academic
achievement would be reflected in general and domain specific components of self-concept and
(b) that the national differences would also influence the various components of self-concept.
METHOD
Experimental Design
The entire investigation comprised a combination of a classical experimental (factorial)
design and appropriate multivariate approaches. First, a multivariate analysis of variance model
(two-factor /2x3/ MANOVA) was performed exploring differences in variables of self-concept
and self-esteem (including 13 SDQ-III subscales and Rosenberg Self-esteem Scale) as a function
of nationality and academic achievement. Nationality (Slovenian or French) and academic
achievement (low, medium, high) were treated as independent (main) factors, the variables of self-
concept and self-esteem being dependent. The sex as possible influential factor was included in the
analysis as the control variable.
Complementary to the analyses of variance, three discriminant analyses were also
programmed in order to discover more precisely the role of particular self-concept domains in
discriminating between nationalities and academic success. In the first analysis, nationality was
6

computed as a group factor (Slovenian and French group), in the second analysis, the academic
achievement was the group factor (low, medium and high achievement group), while in the final
analysis a group factor was defined as a composite of nationality by academic achievement (low-
achievement Slovenian students, medium-achievement Slovenian students, high-achievement
Slovenian students, low-achievement French students, medium-achievement French students and
high-achievement French students). In all cases of discriminant analyses, the self-concept variables
(representing domains of self-concept and self-esteem) were included as independents.
Subjects
A total of 230 high school students (120 women, 110 men) from Slovenia and France
participated in the study. They were selected on the ground of comparable educational programme
and approximately the same age (16-17). The Slovenian sample (169 participants) was drawn
from three high schools (Gymnasium) in Ljubljana. All of them were attended to third class of the
school. The sample of 61 French participants was drawn from members of class 1 from one
school (college) in Strasbourg.
Measuring instruments and variables
Self-concept. Two psychological instruments were applied to measure general self-concept
and specific domains of self-concept. The first instrument was the Self-Description-Questionnaire
III (SDQ III), constructed by Marsh and O'Neill (1984) The questionnaire is based upon the
Shavelson model of self-concept (Shavelson & Bolus, 1982). SDQ III is specially designed for
adolescents at age 15 and over (Marsh, 1989) and it consists of 13 self-concept areas
mentioned above (Marsh & O'Neill, 1984). Besides, the questionnaire includes several academic
components of self-concept (mathematical, verbal and general academic self-concept). Therefore,
it is particularly appropriate for the investigations in the context of academic achievement.
The second instrument was the Rosenberg Self-esteem Scale (Burns, 1979). This scale
measures participants' global self-esteem. It is a scale which is very frequently used in self-esteem
investigations; the special reason for the application of this scale was its frequent use in other
multicultural studies (Robinson, Tayler & Piolat, 1990), what only increases the comparative
7

value of the results of the present study.
Academic achievement. Academic achievement was computed by the sum of general
academic achievement, achievement in mathematics and in native language (Slovene for
Slovenian students and French for French students) of the previous and current school year. It
was measured by average school marks (1-unsufficient, 2-sufficient, 3-good, 4-very good, 5-
excellent). For the use in the analyses of variance and discriminant analyses the marks were
grouped in three categories: low achievement (1, 2), medium achievement (3) and high
achievement (4, 5).
The question might be raised, however, in regard of usefulness of the school marks as
measures of academic achievement. Certainly, more objective achievement tests could be used
instead. Undoubtedly, the external application of standardized achievement tests would
substantially raise the objectiveness and therefore the comparability of data. On the other hand,
the achievement tests are the measures of the scholastic knowledge, not the measures of the
academic success in full sense. Besides, the aim of our study was the examination of the
relationship between academic achievement and self-concept on the multinational level; the
investigation of eventual national differences in academic knowledge was not purported at all. In
the realm of school situation, the school marks still dominantly determinate the subjective sense of
the academic achievement and academic success. School marks are therefore highly relevant
indices of academic success as related to self-concept. According to the problem of the present
study, it is far more appropriate to choose school marks as indicators of academic success that
should be reasonably related to the various aspects of self-concept.
Additional data. General information (age, gender, socioeconomic variables, academic
achievements etc.) was collected by an additional questionnaire.
All questionnaires were appropriately translated and prepared for administration in
respective languages (Slovenian and French).
Procedure and statistical treatment of data
The gathering of data took place in February 1993 in three high schools in Ljubljana and
in February 1994 in a college in Strasbourg. The participants obtained and completed a packet of
8

questionnaires (general questionnaire, SDQ III and Rosenberg Self-esteem Scale). The results
were then analysed with statistical programme SPSS (Hull & Nie, 1984).
RESULTS AND DISCUSSION
Multivariate Analysis of Variance (MANOVA)
Two-factor MANOVA, exploring differences in self-concept domains as a function of
nationality and academic achievement yielded many significant results. All multivariate effects
were significant, namely for both main effects, nationality (multivariate F = 6.44, p < .001), and
academic achievement (F = 3.09, p < .001), and for interaction affect also (F = 1.6, p = .029).
The univariate effects of nationality, academic achievement and interaction on different self-
concept domains are shown in Table 1.
Table 1
Univariate effects of nationality, academic achievement and their interaction on different self-
concept domains
Self-concept
domains
nationality academic
achievement
interaction:
nationality x
academic achievement
F p F p F p
SDQ1
(mathematical)
2,21997 ,138 24,33395 ,000 *** 2,50675 ,084
SDQ2
(verbal)
17,59138 ,000*** ,88187 ,416 2,25800 ,107
SDQ3
(academic)
13,65658 ,000*** 6,62682 ,002 ** 1,34235 ,264
SDQ4
(creativity)
4,22091 ,041* 3,19174 ,043 * 3,53623 ,031*
SDQ5
(physical abilities)
2,96735 ,087 1,27404 ,282 ,47546 ,622
SDQ6
(physical appearance)
1,55195 ,214 ,35072 ,705 ,34987 ,705
SDQ7
(same sex peer
relations)
6,66337 ,011* ,14585 ,864 ,40679 ,666
SDQ8
(opposite sex peer
relations)
2,48782 ,116 1,55789 ,213 ,45483 ,635
SDQ9 9,50349 ,002** ,83663 ,435 1,53161 ,219
9

(relations with
parents)
SDQ10
(religion)
15,08923 ,000*** 2,15407 ,119 ,21264 ,809
SDQ11
(honesty/reliability)
3,08458 ,081 ,58139 ,560 ,12818 ,880
SDQ12
(emotional stability)
1,09167 ,297 1,45250 ,236 2,81432 ,062
SDQ13
(general)
4,23618 ,041* 1,20174 ,303 1,37265 ,256
Self-esteem ,25816 ,612 ,25228 ,777 ,22073 ,802
* p < .05 ** p < .01 *** p < .001
For the nationality, significant effects were found for 7 subscales of SDQ III. Table 2
shows the respective mean values of self-concept domains and self-esteem for different academic
achievement groups from both samples (Slovene and French). Thus, we can see that French
students obtained higher values in the following self-concept domains: verbal (SDQ2), academic
(SDQ3), relations with same sex peers (SDQ7), relations with parents (SDQ9), religion and
spirituality (SDQ10), and general self-concept (SDQ13). Slovenian students had higher self-
concept in the area of problem solving and creativity (SDQ4). There were no statistically
significant differences in self-esteem between Slovenian and French groups.
Academic achievement is significantly related to mathematical (SDQ1), academic (SDQ3)
and creativity (SDQ4) domain of self-concept. This result is concordant with the evidence, that
academic success correlates mostly with the academic or scholastic components of self-concept
(Marsh, 1990, 1994; Marsh & O'Neill, 1984). It is obvious that academic success contributes
very little to the other self-concept domains and vice-versa. The incompatibility between some
major components of self-concept (e.g. the sexual component) and academic achievement is well-
known. In the present study, the heterosexual component of self-concept (SDQ8) also showed
insignificant correlation with general academic success (r = -.109) and even significant negative
correlation with the success in mathematics (r = -.159).
The interaction between nationality and academic success proved to be significant in the
case of creativity domain of self-concept (SDQ4). As we can see from Table 2, the creativity
scores increased with the academic success in French sample, while they remained the same or
10

even slightly decreased in Slovene sample. It seems that French teachers more properly rewarded
the creative students than Slovene teachers. It is interesting, that regarding the creativity
component, the low-achievement groups are equal or even better than medium-achievement
groups, and for the Slovene sample, they are on the same or even higher level than high-
achievement group (see Table 2). It could be that creative subjects invest more interest and energy
in non-scholastic activities, while their able but less creative peers concentrate upon school work
and thus outperformed them on the academic field. Being so, the question could be raised, how
appropriate are the methods of selection in the schools and in the universities which favorize
almost exclusively previous academic success. It is quite possible that some substantial number of
creative candidates would be excluded by those selective procedures.
Table 2
Arithmetic means of self-concept domains for respective academic achievement groups in
Slovene and French samples.
Self-concept Slovene French
domains Academic achievement Academic achievement
low medium high low medium high
SDQ1
(mathematical)
30,70 32,80 38,50 22,00 30,29 41,41
SDQ2
(verbal)
38,11 37,33 36,96 42,50 42,96 46,38
SDQ3
(academic)
33,68 37,41 41,23 43,25 40,64 44,97
SDQ4
(creativity)
42,55 41,48 41,39 37,00 36,61 42,38
SDQ5
(physical abilities)
43,50 45,67 41,70 47,50 46,54 45,97
SDQ6
(physical appearance)
38,19 37,80 37,82 36,50 39,14 41,41
SDQ7
(same sex peer relations)
42,22 42,02 41,67 45,00 48,14 47,41
SDQ8
(opposite sex peer relations)
42,51 41,51 37,73 43,75 43,36 42,52
SDQ9
(relations with parents)
40,02 41,71 42,20 50,50 44,61 44,97
SDQ10
(religion)
18,13 25,57 21,00 23,50 36,68 36,97
SDQ11 48,91 49,92 51,86 53,50 53,21 53,76
11

(honesty/reliability)
SDQ12
(emotional stability)
39,71 35,59 38,28 29,75 37,89 39,52
SDQ13
(general)
49,11 48,47 48,98 46,00 54,68 55,00
Self-esteem 28,71 27,88 28,83 25,75 27,46 28,14
Discriminant Analyses
Discriminant analyses were computed additionally to the MANOVA in order to estimate
the relative contribution of self-concept domains to the discrimination between two national
groups (Slovene and French; see Table 3), to the discrimination between three groups of academic
achievement (low, medium and high achievement; see Table 4), as well as to the discrimination
between six groups with combined values of nationality and academic achievement (low
achievement Slovene - S1, medium achievement Slovene - S2, high achievement Slovene - S3,
low achievement French - F1, medium achievement French - F2, high achievement French - F3;
see Table 5).
Table 3 shows the correlations between self-concept domains as discriminating variables
and canonical discriminant function for two national groups (French and Slovene). The
discriminant function being extracted was highly significant (p < ,001). The values of group
centroids were 0,664 for the Slovene group, and 1,567 for the French. Positive correlations with
discriminant function thus indicate higher results of French subjects. The results are very
concordant with the results of MANOVA. In general, the largest contributions to the
differentiation between groups were produced by the same components of the self-concept which
also showed the most significant differences between the national groups in MANOVA: the verbal
(SDQ2), religious (SDQ10), academic (SDQ3), same sex peer relations (SDQ7), general
(SDQ13), and relations with parents (SDQ9) component. In all these cases the French subjects
demonstrated higher values. On the other side, Slovene subjects again are higher on creativity
component (SDQ4).
12

Table 3
Correlations between discriminating variables and canonical discriminant function for two
national groups (French and Slovene). Positive correlations indicate higher results of French
subjects.
Variables Discriminant
function
SDQ2 ,44645
SDQ10 ,37681
SDQ3 ,35396
SDQ7 ,30299
SDQ13 ,26343
SDQ9 ,21681
SDQ4 -,19792
SDQ11 ,19791
SDQ6 ,11708
SDQ8 ,10337
SDQ5 ,09356
SS -,05240
SDQ1 ,04236
SDQ12 ,01385
Table 4 shows the correlations between self-concept variables and two canonical
discriminant functions for three groups differing in the academic achievement (low-, medium-, and
high-achievement group). Both discriminant functions were significant. The values of group
centroids for the first discriminant factor were -0,893 for the low-achievement group, -0,124 for
medium-achievement group, and 0,931 for high-achievement group. The respective values of
13

group centroids for the second discriminant function were 0,371 for the low-achievement group,
--0,561 for medium-achievement group, and 0,243 for high-achievement group. Obviously, the
first discriminant function differentiated almost linearly between the lowest and the highest
academic achievement group, the medium group being in between, while the second function
discriminates between the medium achievers on the one side and both low and high achievers on
the other side.
Table 4
Correlations between discriminating variables and two significant canonical discriminant
functions for three academic achievement groups. Positive correlations with first discriminant
function indicate higher results of high achievement group, and negative correlations indicate
higher results of low achievement group. Positive correlations with second discriminant function
indicate higher results for both low- and high-achievement group, while negative correlations
indicate higher results in medium-achievement group..
Variables Discriminant function
Function 1 Function 2
SDQ3 ,63782* -,13638
SDQ1 ,63599* ,29676
SDQ11 ,23619* -,09242
SDQ9 ,18221* -,02445
SDQ8 -,17408* -,14225
SDQ13 ,13568* -,10610
SDQ2 ,13538* ,03121
SDQ6 ,08697* -,02845
SDQ10 ,22704 -,49150*
SDQ4 -,04760 ,47032*
SDQ12 -,00154 ,29037*
SDQ5 -,06093 -,26984*
14

SS ,00963 ,14153*
SDQ7 ,09204 -,12581*
The data obtained from discriminant analysis of academic achievement again confirm some
previously presented results of MANOVA. The first discriminant function which contributed
almost linearly to the difference between low and high achievement group correlates with general
academic (SDQ3) and mathematical self-concept (SDQ1). Remarkable is the negative correlation
with the opposite sex peer relations component of self-concept (for the difference between
respective groups see Table 2). The heterosexual component of self-concept is also in negative
relationship with academic achievement. This result is in accordance with expectation that more
expressed sexual interests and motives could interfere with the success in the school.
The second discriminant function delineating medium achievers from both low- and high-
achievement groups is also interesting. For example, it shows that medium achievers exceed both
other groups in religious (SDQ10) and physical domain of the self-concept (SDQ5). On the other
hand, they tend to be lower in the creativity domain of the self-concept (SDQ4), in emotional
stability (SDQ12), and in the mathematical component (SDQ1).
The next discriminant analysis was performed on the 6 groups defined by nationality and
academic success. The groups were designed as follows: low achievement Slovene subjects (S1),
medium achievement Slovene subjects (S2), high achievement Slovene subjects (S3), low
achievement French subjects (F1), medium achievement French subjects (F2), high achievement
French subjects (F3). Table 5 shows the correlation between self-concept variables and three
significant discriminant functions.
Table 5
Correlations between discriminating variables and three significant canonical discriminant
functions for six groups defined by combination of nationality and academic achievement.
Variables Discriminant function
Function 1 Function 2 Function 3
15

SDQ2 ,55992* -,00786 ,03749
SDQ13 ,35704* ,00249 ,05087
SDQ10 ,32501* ,18034 -,12813
SDQ7 ,32239* ,02954 -,08379
SDQ8 ,22111* -,19668 -,13816
SDQ6 ,17659* ,00526 ,08304
SDQ3 -,00368 ,68609* ,06631
SDQ9 ,00767 ,26614* -,13434
SDQ11 ,05106 ,25035* -,03414
SDQ1 ,01610 ,41674 ,66139*
SDQ4 ,04299 -,20882 ,43286*
SDQ12 ,23785 -,17967 ,25221*
SDQ5 ,09675 -,03679 -,15200*
SS ,01582 -,04668 ,12804*
According to the values of group centroids for all three discriminant functions (Table 6) it
is obvious that the first function differentiates between two nationalities irrespective of academic
success. The high- and medium-achievement French group remarkably exceeded all other groups.
The first discriminant function correlated for the most part with verbal (SDQ2), religious
(SDQ10) and general (SDQ13) self-concept components. That is in good accordance with the
results showing the dominance of French subjects in those aspects of the self-concept (see Table 2
too). The second discriminant function vary with the academic achievement but also with the
nationality (the values of group centroids being the largest in F3 and smallest in S1, but all French
group have higher values than Slovenes). The function correlates most positively with general
academic (SDQ3) and mathematical academic (SDQ1) self-concept component and most
negatively with the creativity (SDQ4) and emotional stability component (SDQ12). This result
16

could be explained by data seen in Table 2: discriminant loadings probably reflect high values
obtained by French high achievement group on general and mathematical academic self-concept
and high values of Slovene low-achievement group on creativity and emotional stability self-
concept component. Finally, the third discriminant function delineates mainly between academic
achievement (both high achievement groups having the greatest positive values of centroids). The
highest correlations with the third function has mathematical (SDQ1) component of self-concept
(both high-achievement groups have clearly highest means on this component, see Table 2).
Nevertheless, the nationality plays a certain role within the third discriminant function too. It
could be demonstrated by the relatively high position of the group centroid for the low-
achievement Slovene group. Thus, the substantial correlation of third discriminant function with
the creativity component of self-concept could be explained on the grounds of the high results of
Slovene low-achievers on the SDQ4.
Table 6
Group centroids for three discriminant functions.
Group Func 1 Func 2 Func 3
S1 -,41163 -1,10251 ,16803
S2 -,51775 -,22208 -,17312
S3 -,68373 ,48539 ,62544
F1 -,55152 ,78944 -2,24341
F2 1,25980 ,60239 -1,14031
F3 1,44937 1,20323 ,44700
17

General discussion and conclusions
Before proceeding, some general limitations and precautions regarding our results should
be mentioned. The results were drawn from the rather narrow national samples, consisting of
the subjects from three secondary schools in Ljubljana, Slovenia, and from one secondary school
in Strasbourg, France. The representativeness of the French sample is especially questionable. In
this respect, a replication of the study on larger, intranationally more balanced samples would be
recommended.
We have also been strongly aware of another limitation, dealing with some well-known
problems of cross-cultural research. The general assumption of metric equivalence of applied
instruments is rather dubious when confronting in multicultural or multinational studies (Watkins
and al. 1995). The question is, whether the same instruments have equivalent meaning for the
subjects belonging to different cultural backgrounds. Nevertheless, we assume that cultural
distance do not represent an insurmountable obstacle, especially in the case when different
national samples share the same basic culture (both French and Slovene subjects belong to the
common European or Western cultural tradition). If it would not be so, every intercultural or
international comparison would be senseless or even impossible.
In general, the results of both Slovenian and French participants on self-concept and self-
esteem scales fall in the range of normal average (Lamovec, 1994). But additionally to this global
perspective, we found the following differences between French and Slovene subjects:
1. The French subjects exceeded Slovenes in some domains of self-concept (verbal,
academic, relations with same sex peers, relations with parents, religion and spirituality, and
general self-concept), while Slovene subjects exceeded French subjects in the domain of problem
solving and creativity.
2. There was no significant difference between both national samples in self-esteem.
3. The French subjects exceeded Slovene pupils in academic achievement in general. They
have significantly better school marks in average (mean 3.66 for French and 3.05 for Slovene
sample, p < 0.001).
18

The results obtained in our study could be interpreted partially in the light of cultural and
cross-national differences. French and other Latin nations are characterized by more extraverted
traits, including a higher degree of verbal expressiveness in comparison to Central, Western and
North European nations, which tend to be more reserved and introverted (Musek, 1994).
Moreover, the stimulation of verbal competence is very pronounced in educational tradition in
France. All this may substantially contribute to higher results of French sample in verbal
component of self-concept, and maybe to some other differences in self-concept. Higher degree of
peer-orientation in social relationships could also be expected from more extroverted culture and
cause henceforth the corresponding difference in the same-sex-peer-relations domain of self
concept. On the other hand, the perfection and creativity were traditionally admired in Slovenian
culture: there is sufficient empirical evidence for strong involvement of those traits in Slovene
psychological autostereotype, and for factual existence of high degree of achievement motivation
in Slovenian population (Musek, 1994). Higher results in problem-solving and creativity domain
of self-concept in Slovene subjects are thus not surprising.
In our study, French subjects turned out to be more successful in school than Slovenian.
Of course, this information is an ambiguous one, because it could be an artifact of different
assessment standards applied by teachers in the schools in both countries. But there are some
others possible explanations too.
According to many studies who founded a significant correlation between academic
achievement and academic self-concept (House, 1993; Marsh, 1987) it could be supposed that
also in this study higher academic achievement of French students is mostly related to their
higher academic self concept. It is interested that Watkins and al. (1995) founded higher
academic self-concept among groups of Chinese, Nepalese, Nigerian and Filipino female school
students and lower self-concept among Lithuanian school students. We may speculate similarly
with the authors that the changes in educational school system in countries in transition from
socialistic versus democratic societies could affect the academic self-concept of students.
On the other hand, our results don't support the results of Watkins and al. (1995) who
founded also lower self-esteem among Lithuanian students. It is interesting, indeed, that in our
study French subjects exceeded the Slovenes in both academic and general self-concept, but not in
19

the global self-esteem. It is even more surprising, if we consider that the actual correlations
between self-esteem and both other measures of self-concept are significant. The correlation
between self-esteem and academic self-concept is 0.22 (p < .001), and the correlation between
self-esteem and general self-concept is quite high (0.72, p < .0001). Two possible explanations
could be found for this interesting result: the first possibility is that the results of Rosenberg Self-
esteem scale are simply unsufficiently reliable, and the obtained unsignificant difference between
both national samples is only casual (in academic as well as in general self-concept both samples
differ significantly). The second possible explanation is that greater satisfaction with themselves
(higher self-concept) moves French students to the greater self-esteem, but that self-esteem of
Slovene subjects remained unaffected by lower academic self-concept and lower actual academic
achievement. We can only guess, why it is so.
Further, we can compare our results with the results found in the study School
attainment, self-esteem, and identity by Robinson and his collaborators (1990). In spite of the
slight difference in instruments and in the age of French and English subjects participating in
their study, they found similar results as we did. French subjects were more succesful than their
British counterparts.
In Robinson’s study as well as in ours French students are in some way more satisfied
with themselves: the French participants in Robinson's research have higher self-esteem
whereas the French participants in our research have higher general self-concept. In the study
of Robinson et al. (1990), low French achievers showed a higher self-esteem than low English
achievers. As already mentioned, we didn't find any differences in self-esteem among high and
low achieving French and Slovenian groups of students.
According to Robinson and his collaborators (1990) it seems that the French educational
system deals quite democratically with the problems of academic achievement, self-esteem and
creativity. They suggest "that French pupils in their study offered a portrait of more satisfied
conformity to cultural norms both in and out of school, and it is particulary noteworthy that
their low achievers remain more strongly integrated within the educational system,"
(Robinson, Tayler & Piolat, 1990, pp. 401). If the data reported in this contribution are
representative, “it seems that the French educational system enables pupils' faith in their
20

status as agents responsible for their actions, it enables their regard for themselves as problem-
solvers, and it also enables each pupil to achieve his or her potential. Consequently,
organisation of the French educational system seems to be friendlier to the their pupils, they are
more satisfied with themselves, they have higher aspirations and higher academic achievement
than their English peers.” (Robinson, Tayler & Piolat, 1990, pp. 401).
Similar conclusions could be made in regard of differences between French and Slovene
subjects obtained in our study. It could be supposed that the Slovenian educational system is still
based on premises which were characterised by less democratic (although maybe more
egalitarian) conceptions, and a school subject(???)-centred instead of more person-centred
approach (Piciga, 1993). These premises could be also related with rather surprising data
concerning the relationship between the academic achievement and problem solving/creativity
self-concept in Slovene sample. There is almost no difference in problem solving and creativity
domain between low-, medium- and high-achievers in Slovene group (in fact, low-achievers
obtained the highest values). In the French group, the high-achievers grossly overscored the
others in respect of creativity domain of self-concept. It seems that Slovenian students have
problems how to invest the creativity for better academic achievement, or they even have to
renounce their critical thinking and creativity for the sake of academic promotion.
In this regard, our results may stimulate particular efforts in the advancing of educational
system in some countries. If the strategic aim of the improvement of educational system is to
became more democratic and creativity-stimulating on the one side, but efficient in knowledge
gaining in the same time, the psychological factors in education should be taken into consideration
very seriously. The optimal models of education should respect the general or even universal
conditions for academic effectiveness, but also consider specific factors including cultural and
national characteristics.
In order to stimulate the strategic aim of improvement of educational system, we find very
important to realise some necessary conditions: (1) to know different modern educational
systems and consider their deficiencies which could cause eventual psychological damage on
students, (2) on the basis of the above mentioned research to elaborate the own national
21

curriculum based especially on the knowledge of personality characteristics of the own nation,
(3) to try to make the government and its ministry consider not only their beliefs, but also the
results of scientific research dealing with national schooling. If we consider the relationship
between each student's needs and values and the school environment and if we enable
critical thinking and the development of creative abilities and creativity self-concept of
students, we will already make a step in the direction of democratisation of the educational
system in Slovenia.
LITERATURE
Cross, S.E., & Markus, H.R. (1994). Self-schemas, possible selves, and competent performance.
Journal of Educational Psychology, 86 (3), 423-438.
House, J.D. (1993). The relationship between academic self-concept and school withdrawal.
The Journal of Social Psychology, 133, (81), 125-127.
Hull, C.H., & Nie, N.H. (1984). SPSS-X. New York,: McGraw-Hill.
Keltikangas-Järvinen, L. (1991). The stability of self-concept during adolescence and early
adulthood: a six-year follow up study. The Journal of General Psychology, 117(4), 361-368.
Maccoby, E.E. (1980). Social development: Psychological growth and the parent-child
relationship. New York:Harcourt Brace Jovanovich.
Marsh, H.W. (1987). The Big-Fish-Little-Pond Effect on Academic Self-Concept. Journal of
Educational Psychology. 79 (3), 280-295.
Marsh, H.W. (1989). Age and sex effects in multiple dimensions of self-concept: preadolescence
to early adulthood. Journal of Educational Psychology, 81 (3), 417-430.
Marsh, H.W. (1990). A multidimensional, hierarchical model of self-concept: theoretical and
empirical justification. Educational Psychology Review, 2(2), 77-171.
Marsh, H.W. (1994). Using the national longitudinal study of 1988 to evaluate theoretical
models of self-concept: the self description questionnaire. Journal of Educational Psychology,
86 (3), 439-456.
Marsh, H.W., & O'Neill, R. (1984). Self description questionnaire III: The construct validity of
multidimensional self-concept ratings by late adolescents. Journal of Educational Measurement,
21 (2), 153-174.
Piciga, D. (1993). Psychology and the innovation of science education in Slovenia. III. Alps-
22

Adria Symposium of Psychology. Department of Psychology, Faculty of Philosophy, University
of Ljubljana, 159-163.
Robinson, W.P. (1990). Academic achievement and self-esteem in secondary school: muddles,
myths and reality. Education Research and Perspectives. 17 (1), 3-21.
Robinson, W.P., & Tayler, C.A. (1989). Correlations of low academic attainment in three
countries. International Journal of Educational Research, 13, 581-594.
Robinson, W.P., Tayler, C.A., & Piolat, M. (1990). School attainment, self-esteem, and
identity: France and England. European Journal of social psychology, Sept-Oct 20 (5), 387-
403.
Robinson, W.P., & Breslav, G. (unpublished). The relations of language spoken and academic
achievement to the perceptions and evauations of adolescent pupils in Latvia.
Rosenberg, M. (1965). Society and adolescent Self-Image. Princetone, NJ: Princetone
University Press.
Rosenberg, M. (1985). Self-concept and psychological well-being in adolescence. In: R.L.
Leahy (Edit.). The development of the self. Academic Press, Inc, 205-246.
Shavelson, R.J., & Bolus, R. (1982). Self-concept: The interplay of theory and methods. Journal
of Educational Psychology. 74 (1), 3-17.
Watkins, D., & Gutierrez, M. () Causal Relationships among self-concept, attributions, and
achievement in Filipino students. The Journal of Social Psychology, 130 (5), 625-631.
Watkins, D., McCreary Juhasz, A., Walker, A., & Janvlaitiene, N. (1995). The self-description
questionnaire-1: a Lituanian application. European Journal of Psychological Assessment, 11
(1), 41-51.
23