See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/266151132

The role of interpersonal relationships with peers and with teachers in

students' academic achievement

Article · January 2007

CITATIONS READS

15 4,358

3 authors:

Katja Košir Gregor Socan

University of Maribor University of Ljubljana
47 PUBLICATIONS   266 CITATIONS    34 PUBLICATIONS   479 CITATIONS   

SEE PROFILE SEE PROFILE

Sonja Pečjak
University of Ljubljana
57 PUBLICATIONS   293 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

PYD-SI-MODEL: Positive Youth Development in Slovenia: Developmental Pathways in the Context of Migration View project

Evaluation of processes of care View project

All content following this page was uploaded by Sonja Pečjak on 29 September 2014.

The user has requested enhancement of the downloaded file.

 Review of Psychology,
2007, Vol. 14, No. 1, 43-58 UDC 159.9

The role of interpersonal relationships with peers and with teachers in

students’ academic achievement

KATJA KOŠIR, GREGOR SOČAN and SONJA PEČJAK

The purpose of this study was to examine the relation between social and academic variables in different school
periods and to investigate, whether affective and motivational factors can be regarded as mediating variables in the
relation between social and academic variables in school. 1159 students from three different periods of schooling,
covering the age range from late childhood through early to middle adolescence, participated in the study. Differ-
ent models of relations between social and academic variables were tested using structural equation modeling. The
results show that the included mediating variables (well-being in school and academic engagement) do not explain
the relation between social relations and academic achievement. In younger students, peer relations are related
to students’ academic achievement, which does not hold true for both older age groups. Relations to teachers are
related to students’ academic outcomes in all periods of schooling. The results also suggest that the method of data
assessment is a very important factor in establishing the relations between variables.

Key words: peer relations, teacher-student relations, academic achievement, psychological assessment, structural
equation modeling

Social and academic variables in educational research Recently, the researchers in the field of academic moti-
vation started to be increasingly oriented to the investiga-
In the past, academic and social variables were regarded tion of the effect of social and emotional factors on aca-
as two completely separated aspects of motivation. Recent- demic motivation and achievement. Numerous authors (e.g.
ly, these beliefs have been changed radically; namely, aca- Dweck, 1996; Weiner, 1996; Wentzel, 1996a) stress that in
demic and social variables can be intertwined in a number the past, academic achievement and social motivation were
of ways, as for example (Weiner, 1996): studied separately and that the reciprocal interaction be-
tween the two constructs was not taken into consideration.
(1) Feelings of rejection, the lack of social support and This also holds true for the research in school context where
dissatisfaction of social needs influence academic the relation between academic and social factors in students
motivation and performance in school; students feel- has been investigated only recently.
ing lonely in the classroom are more likely to give up Two most frequent and important forms of social rela-
in learning situations and have a higher rate of school tions that students form and maintain in school are relations
drop-out. to teachers and to peers. In the next part of the article, the
(2) Students’ academic achievement and their choice importance of the two forms of social relations is discussed
of peer groups are related; students having similar in detail.
achievement strivings tend to form friendships and
peer groups. The role of teacher support and involvement
(3) Peers also act as models of appropriate academic, as
well as social behavior. Contemporary studies (e.g. Kindermann, 1993; Ryan,
Stiller, & Linch, 1994; Wentzel, 1998) mostly investigate
the influence of relations to peers and to teachers on stu-
dents’ academic motivation, and consequently on their aca-
demic achievement. Studies examining the role of relations
Katja Košir, Department of Psychology, Faculty of Arts, University of to teachers in students’ academic achievement variables
Ljubljana, Aškerčeva 2, 1000 Ljubljana, Slovenia. E-mail: katja.ko- originate mainly from the self-determination theory (Con-
sir@ff.uni-lj.si (the address for correspondence); nell & Wellborn, 1991; Deci & Ryan, 2000-2001) which
Gregor Sočan, Department of Psychology, University of Ljubljana; stresses the need for relatedness as one of the basic psycho-
Sonja Pečjak, Department of Psychology, University of Ljubljana. logical needs. According to this theory, teachers’ involve-

43

KOSIR SOCAN.indd 43 22.1.2008 13:10:55

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

ment is crucial for the satisfying the need for relatedness. and motivational factors (Kupersmidt, Buchele, Voegler, &
This refers to the quality of interpersonal relations with stu- Sedikides, 1996; Wentzel, 1996b, 1998). However, the na-
dents and is manifested through teachers having time for ture of the relation between social and academic variables in
students, expressing positive feelings towards them, being students still remains unclear.
flexible to their needs, etc. Teacher’s involvement seems to
be the strongest predictor of students’ academic motivation,
Students’ social acceptance in relation to their aca-
amongst all of the other presumably important dimensions
demic outcomes
of teachers’ behavior. The students of highly involved teach-
ers perceive their teachers not only as involved, but also as
giving more structure and support to student’s autonomy, Peer relations have already been studied in relation to
independently of the teacher’s actual behavior on these two various academic outcomes. The results of such studies
dimensions (Furrer & Skinner, 2003; Skinner & Belmont, consistently show that popular students are those usually
1993). The sense of relatedness tapped by the measures of achieving better outcomes and that rejected students are
school climate and the quality of teacher-student relations, frequently those having learning difficulties (e.g. Wentzel,
as well as the feelings of belonging, acceptance, importance, 1991). These findings were most consistent when academic
and interpersonal support, are related to important academic outcomes were measured by students’ grades (Hatzuchtis-
outcomes, including self-efficacy, expectations of success, tou & Hopf, 1996; Wentzel, 1991; Wentzel & Cadwell,
achievement values, positive affect, effort, engagement, in- 1997), although peer acceptance was also found to be re-
terest in school, task goal orientation, and grades (see Fur- lated to standardized test achievements (Austin & Draper,
rer & Skinner, 2003). Feeling related in the school context 1984), as well as to students’ intelligence (Wentzel, 1991).
gives students a sense of importance and thereby enhances Students of lower sociometric in-class status, especially
their activity (Furrer & Skinner, 2003). rejected students, represent a group of higher risk for
Wentzel (1993) reported that teachers’ liking of students learning difficulties and school drop-out (Hatzuchtistou &
is positively related to students’ academic achievement. Hopf, 1996; Ollendick et al., 1992). This holds especially
Students forming positive relations to their teachers have true for aggressive rejected students (Wentzel & Asher,
more positive attitudes towards school and like school more 1995).
(Skinner & Belmont, 1993). In classrooms where teachers
report that they respond to students’ academic, as well as Students’ friendships in relation to their academic
social needs, students report that they ask for help more fre- outcomes
quently (Ryan, Gheen, & Midgley, 1998). Also, the relation
to teachers plays an important role in some inappropriate
behaviors in school such as cheating at tests; students that Wentzel and Caldwell (1997) conducted a longitu-
report to have a good relation with their teacher also report dinal study of the relation between students’ friendships
less cheating (Murdock, Hale, & Weber, 2001; Murdock, and their academic achievement. The results of this study
Miller, & Kohlhardt, 2004). Similarly, Ryan, & Patrick showed that the relation between having a friend and aca-
(2001) found that 8th grade students that perceiving teach- demic outcomes held stable in the period of two years. In
ers’ support also report about having more self-regulative another longitudinal study, Wentzel, McNamara Barry and
learning and less misbehavior compared to their behavior Caldwell (2004) investigated the influence of friendships
in the 7th grade. on motivation and school adjustment of early adolescents.
Students without reciprocal friendships were less proso-
cial, had lower academic outcomes and reported more
The relation between peer relations and students’
emotional distress compared to their peers with reciprocal
school performance
friendships.
The second research direction in the investigation of the
effect of social factors on academic variables is represented Perceived peer support and students’ academic vari-
by studies of the relation between peer relations and academic ables
achievement. The results of these studies (e.g. Hatzuchtistou
& Hopf, 1996; Ollendick, Weist, Burden, & Greene, 1992; The perception of peer and teacher support is regarded
Wentzel, 1991; Wentzel & Asher, 1995; Wentzel & Cad- as an especially important factor in the students’ achieve-
well, 1997) indicate positive relation between peer relations ment of learning goals. The students believing their peers
and students academic performance. The results (e.g. Chen, support and care for them are usually more engaged in posi-
Chang, & He, 2003; Guay, Boivin, & Hodges, 1999; Went- tive classroom behaviors compared to the students which
zel, 1993; Wentzel & Caldwell, 1997) mostly indicate that do not perceive such a support (Wentzel, 1994, 1997). The
students feeling more accepted by their peers also achieve latter group of students represents a group having a higher
more. It is likely that this relation is mediated by emotional risk to develop learning difficulties (Goodenow, 1993).

44

KOSIR SOCAN.indd 44 22.1.2008 13:10:55

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Relations to teachers versus relations to peers cially critical period for students’ learning beliefs and be-
haviors (Eccles & Midgley, 1989). For some early adoles-
In younger students, Birch and Ladd (1997) found a cents, the increase in self-reflection, autonomy and identity
positive relation between students’ peer relations and their exploration leads to new academic interests, an increase
relations to teachers. Teachers provide unpopular students in self-regulative learning and commitment to education
with more corrective feedback and popular students with (Goodenow, 1993). But for many adolescents this is also
more positive reinforcement (White, Sherman, & Jones, a period of doubts in their abilities to succeed in academic
1996). Teacher’s behavior towards students exerts an espe- activities, questioning the value and meaning of school
cially strong influence on other students’ perceptions of this work and consequently the decline of academic effort. The
student in preschoolers and younger school children. At this social environment in the classroom that includes perceived
age, students form more positive relations with peers who teacher, as well as perceived peer support, is therefore cru-
also have more supportive and less conflictive relations with cial for this period. Thus, early adolescence represents an
teachers (Ladd, Birch, & Buhs, 1999; Taylor, 1989). Birch especially sensitive period. For many adolescents, this is a
and Ladd (1996) argue that relations to peers and relations period of decrease in their academic achievement (Ryan &
to teachers probably predict different aspects of school ad- Patrick, 2001).
justment. In some situations, the relation between these two
In transition to adolescence the conflict between social
forms of relations can also be compensatory. The results of
and academic goals is also more frequent than in childhood.
the study investigating the characteristics of early adoles-
cents without friends (Wentzel & Asher, 1995) show that Such incongruities occur when students try to simultane-
teachers’ liking of students can soften the negative effects of ously satisfy the expectations of teachers and of peers which
peer rejection on students’ school adjustment. In any case, can sometimes collide. Also, the incongruities between stu-
the (in)congruity of student’s popularity by peers and by dents’ academic values and those of their friends can appear.
teachers represents an important piece of information about As emphasized by Juvonen (1996), such situations have still
the student’s position in the classroom. not been sufficiently investigated and deserve more research
attention.
As indicated by Wentzel (2003), the strength of peer in-
fluence in relation to the influence of adults is one of the
crucial research questions in the field of school adjust- Mediating variables in relation between measures of
ment. However, both social factors of students’ academic students’ interpersonal relationships and their aca-
achievement – i.e. relations to teachers and to peers – were demic achievement
rarely examined simultaneously. Likewise, when tak-
ing into account the social development in late childhood Mediating processes between social relations in school
and adolescence, it is reasonable to expect that the rela- and students academic outcomes represent another research
tive importance of both forms of relationships for students’ area that has not been explained sufficiently. With the refer-
academic achievement change during different periods of ence to the relation between social acceptance and academic
schooling. Wentzel (2003) emphasized that school adjust- variables, Wentzel (1991; also Wentzel & Asher, 1995) hy-
ment and factors which influence it should be investigated pothesized four possible explanations of this connection:
from a developmental perspective – taking into account
students developmental abilities and tasks. Nevertheless, - Positive peer relations influence students’ sense of re-
the studies in this field mostly focus on one single period latedness which has a positive effect on students’ learn-
of schooling and less on the possible changes in relations ing motivation, as predicted by the self-determination
between social, emotional, and motivational factors and theory. Consequently, students having a better accepted
academic achievement, which might result due to devel- of their peers also achieve more.
opmental changes. - The relation between social and academic variables can
be explained by students’ self-regulatory skills. More
Some developmental changes in transition to adoles- mature and independent students are more self-confi-
cence dent, having a better impulse control, and probably out
of this reason are more successful in school, as well as
better accepted by their peers.
In late childhood and adolescence, peer relations be-
come increasingly important. During the childhood, peer - Students accepted by their peers are usually also more
groups enlarge and become less supervised by adults (Gif- accepted by their teachers, as well. In contrast to this,
ford-Smith & Brownell, 2003). In early adolescence, the teachers are more critical towards rejected students and
role of peers as a source of emotional and instrumental offer them less help, which can lead to lower academic
support becomes even more important than it was in the outcomes of these students.
childhood. Peers also become an important factor in organ- - It is also possible that students’ academic reputation is
izing spare time and act as a factor of sexual satisfaction directly related to their social acceptance; the students
and identity formation. Also, early adolescence is an espe- achieving more are also more liked by their peers.

45

KOSIR SOCAN.indd 45 22.1.2008 13:10:55

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

METHOD

Participants

All together 1159 students from 49 classrooms of Slov-

Figure 1. The presumed model of relations between social, emo- enian elementary and secondary schools participated in the
tional, motivational, and academic variables study. This includes 424 4th grade elementary school students
(mean age 9.80 years; 51% male), 404 7th grade elementary
school students (mean age 12.93 years; 54% male), and 331
2nd year high school students (mean age 15.98 years; 48%
male). From the developmental point of view, participants
Wentzel (2003) hypothesized that the relation between covered the age range from late childhood to early and mid-
students’ interpersonal relations and their academic be- dle adolescence. Also 49 teachers assessed their students on
havior is mediated by some affective variables. The stu- different measures.
dents that do not perceive teacher and peer support can In the sample of secondary school students all forms of
experience psychological discomfort or distress which in secondary school education were represented. Sampling
turn increases their orientation towards themselves and re- was proportional, considering the data of the Ministry of
duces the possibility of their positive orientation towards Education and Sport about the enrolment structure of the
learning and social interactions. Author also stresses the secondary school programs.
need for the development and testing of theoretical mod-
els explaining the relation between social motivation and Variables and instruments
academic variables. Likewise, Covington (2000) empha-
sizes the importance of multivariate studies that longitu- The model tested in the study consists of the following
dinally track the joint effect of cognitive, motivational, variables: peer relations, relations to teachers, well-being
social, and emotional factors on academic achievement. in school, academic engagement, and academic achieve-
The nature of the relation between social and academic ment. The data was gathered using three different sources:
variables is still quite unclear, but the results of some students’ self-report, teacher assessment, and peer nomina-
studies show that the assumption of affective mediating tions. For each variable included in the model at least two
variables could be correct (see e.g. Wentzel, 1998; Went- sources were used.
zel & Caldwell, 1997).
The main purpose of the study is to establish the pre- Measures of peer relations
dictive value of the quality of peer relations and relations
with teachers for students’ academic achievement in differ- Social preference. Social preference was assessed by
ent school periods. It is hypothesized that the relations with sociometric nominations (like most and like least). It was
peers and with teachers independently influence students’ defined as a difference between standardized positive and
academic achievement, and that this influence is mediated negative nominations and is a measure of student’s relative
by students’ well-being in school and academic engage- likableness (Coie, Dodge, & Coppotelli, 1982). Students
ment. The causal model based on the models of Connell and were asked to name three schoolmates they liked most and
Wellborn (1991) and Wentzel (1998) is presented in Figure three they liked least.
1. Peer perceived popularity. Students were asked to name
In this model, it is hypothesized that students with more three most popular classmates in their classroom. The
positive relations to peers and to teachers experience school number of nominations for each student was standardized
as more pleasant and are therefore more engaged in school within classrooms. These standardized values were used as
work. Higher school engagement results in better academic a measure of perceived peer popularity.
achievement. The relation between separate constructs in Number of friendships. Students were asked to name three
the model was already established in various studies (Furrer members of their classroom that were their best friends. The
& Skinner, 2003; Hatzuchtistou & Hopf, 1996; Ollendick et standardized number of nominations each student received
al., 1992, Wentzel, 1991, 1994; Wentzel & Caldwell, 1997; was used as a measure of student’s friendships.
Wentzel et al., 2004). Number of reciprocal friendships. The number of re-
In addition, we wanted to establish the relation between ciprocal friendships was determined for each student. After
the quality of relationships with teachers and with peer in that, this variable was standardized within classrooms.
different school periods and to evaluate different methods Peer academic and personal support. Two subscales
for assessing these variables. from the Classroom Life Instrument (Johnson, Johnson, &

46

KOSIR SOCAN.indd 46 22.1.2008 13:10:56

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Anderson, 1983) were used to assess the perceived academ- school. The scale was translated into Slovene and adapted
ic (e.g. Other students care about how much I learn) and for the purpose of the study. The internal consistency coef-
personal peer support (e.g. Other students like me the way ficient was .90.
I am). These two scales were translated into Slovene and
adapted for the purposes of this study. The Peer Academic Measures of academic engagement
Support Scale consists of four items; alpha coefficient was
.68. The Peer Personal Support Scale consists of five items,
Self-reported academic engagement. The behavioral and
alpha coefficient was .78. Students rated the items using a 5-
cognitive engagement in schoolwork scale (Assor, Kaplan,
point scale (1- never true for me; 5 – always true for me). & Roth, 2002) was used to assess students’ perceptions of
their academic engagement. The scale was translated into
Measures of relations to teachers Slovene and adjusted for this study. The scale consists of
six items (e.g. I do more than what I am required). Students
Teacher’s liking of the students. Teachers received the rated the items on a 5-point scale. The internal consistency
following instruction: “With regard to the numerous dif- coefficient was .74.
ferences between the students, it is normal and human that Teacher report of students’ academic engagement. The
the teacher does not like all students in the same way, al- scale described above was adjusted for the teachers, as well.
though he/she behaves equally fair to all of them. For all This version consisted of five items. The item In classes I
of the students in the class please indicate how much you try to look busy, but I really do not pay attention was missed
like would to have the particular student in your class next out. Teachers rated the items for every single student using
year”. The rating scale was a 5-point scale (1- not al all; 5 a 5-point scale (1 – never; 5 – always). The alpha coefficient
– very much). for this scale was .91.
Peer perceived relation with teachers. Students were Peer perceived academic engagement. Students named
asked to name three or less classmates that have the best three or fewer classmates that were most engaged in lessons.
relations with teachers. The within classrooms standardized The within classrooms standardized number of nominations
number of nominations that each student received was used for each student was used as a measure of peer perceived
as a measure of peer perceived relation with teachers. academic engagement.
Teacher academic and teacher personal support. Teach-
er academic support and Teacher personal support scales Measures of academic achievement
which form a part of the already mentioned Classroom
Life Instrument were used to assess students’ perceptions GPA in the previous school year.
of both, teacher academic (e.g. My teacher likes to help me Peer perceived academic achievement. Students were
learn) and personal support (e.g. My teacher really cares instructed to name three or less best students in their class.
about me). Both scales consisted of four items. The rating The within classroom standardized number of nominations
scale is the same as for peer support scales. Students were for each students was regarded as a measure of peer per-
instructed bear in mind the majority of teachers. Alpha coef- ceived academic achievement.
ficient of internal consistency is .70 for Teacher Academic
Support Scale and .74 for Teacher Personal Support. Procedure

Measures of students’ well-being in school The data was collected at the beginning of the school
year. The questionnaires for students were administered col-
Teacher report of students well-being in school. Teach- lectively in the classrooms. Students were told they should
ers assessed the degree to which every single student felt report about their relations with other students and teachers,
well/was satisfied in school using a 5-point rating scale (1 the characteristics of their classroom, their well-being in
– very unsatisfied/feels very bad; 5 – very satisfied/feels school and about how they engage in schoolwork. Students
very well). first answered the peer nomination procedures and then the
Self-reported well-being in school. Students’ well-be- self-report scales. During the classroom testing teachers
ing in school was assessed using the Scale of Well-Being in were asked to complete the items in which they assessed
students in their class.
School (Keller, Moser, & Rhyn, 1996), a shorter version of
the Scale of Subjective Well-Being (Moser, Bless, & Hae-
berlin, 1989). Students with a high score on this scale feel RESULTS
good in school and like going to school (Moser, Bless, &
Haeberlin, 1989). The shorter version consists of five items Descriptive statistics and the correlations between mani-
with a 5-point rating scale (1- never true for me; 5 – always fest variables are presented in Table 1. Evidently, the source
true for me). The scale consists of items like I like going to of assessing the variable is an important factor of the size of

47

KOSIR SOCAN.indd 47 22.1.2008 13:10:56

 Table 1

48
Descriptive statistics and correlations between varables included in the model for all pupils together and in different school periods

KOSIR SOCAN.indd 48
M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
1. Social preference 4th grade 0.00 1.63
7th grade 0.00 1.61 -
2nd grade 0.00 1.60
Together 0.00 1.59
2. Popularitya 4th grade 1.80 2.66 .49
7th grade 2.00 3.03 .41
-
2nd grade 2.02 3.23 .28
Together 1.92 2.95 .40
3. Friendship 4th grade 2.55 1.85 .70 .39
7th grade 2.52 1.84 .73 .47 -
2nd grade 2.30 1.51 .60 .31
Together 2.47 1.76 .68 .39
4. Reciprocal friendship 4th grade 1.12 1.00 .42 .29 .54
7th grade 1.27 1.06 .50 .37 .64 -
2nd grade 1.60 1.02 .35 .22 .57
Together 1.18 1.02 .43 .30 .58
5. Student academic support 4th grade 3.55 0.86 .14 .12 .06 .08
7th grade 3.16 0.71 .22 .11 .14 .10 -
2nd grade 3.28 0.69 .17 .03 .10 .12
Together 3.34 0.78 .18 .09 .10 .10
6. Student personal support 4th grade 3.78 0.80 .13 .11 .03 .06 .73
7th grade 3.40 0.74 .28 .18 .26 .17 .65 -
2nd grade 3.49 0.65 .18 .09 .12 .16 .70
Together 3.56 0.75 .19 .12 .13 .12 .71
7. Teacher’s liking 4th grade 4.48 0.89 .30 .16 .14 .11 -.01 .00
7th grade 4.12 1.06 .22 .02 .16 .13 .04 -.01 -
2nd grade 4.28 0.89 .15 .06 .03 .07 .02 .00
Together 4.30 0.96 .23 .08 .11 .11 .04 .02
8. Relations to teachers (peer 4th grade 2.14 3.28 .33 .46 .24 .24 .06 .07 .18
report)a 7th grade 2.35 4.22 .11 .21 .13 .14 .01 -.01 .33 -
2nd grade2.35 2.35 4.78 .06 .29 .03 .15 .09 .15 .26
Together 2.27 4.08 .20 .34 .15 .18 .04 .05 .23
KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

22.1.2008 13:10:57
 M SD 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
9. Teacher academic 4th grade 4.35 0.62 -.05 .00 -.04 .02 .41 .34 .01 .01

KOSIR SOCAN.indd 49
support 7th grade 3.69 0.68 -.08 -.05 -.11 -.08 .40 .11 .00 .12 -
2nd grade 3.54 0.68 .09 -.04 .00 -.02 .36 .24 .10 .08
Together 3.89 0.74 -.01 -.04 -.04 -.02 .43 .29 .08 .02
10. Teacher personal 4th grade 4.12 0.82 .00 .08 -.04 -.08 .41 .40 .12 .01 .56
support 7th grade 3.22 0.80 -.06 -.12 -.10 -.07 .37 .21 .13 .22 .63 -
2nd grade 3.14 0.67 .00 -.03 -.05 -.06 .37 .30 .08 .16 .61
Together 3.53 0.90 -.01 -.02 -.06 -.06 .43 .37 .16 .09 .69
11. Well-being in school 4th grade 4.24 0.87 .36 .34 .19 .18 .06 .07 .53 .32 -.07 .09
(teacher report) 7th grade 3.79 0.96 .27 .19 .25 .16 .08 .11 .50 .28 -.05 -.01
-
2nd grade 3.95 0.82 .25 .12 .02 .06 .05 .08 .53 .28 .11 .16
Together 4.00 0.91 .27 .22 .17 .14 .11 .13 .53 .27 .08 .16
12. Well-being in school 4th grade 3.94 0.96 .11 .06 .08 .05 .39 .39 .16 .15 .37 .37 .15
(self-report) 7th grade 2.93 0.99 .07 .03 .02 .00 .31 .20 .13 .15 .32 .34 .06 -
2nd grade 3.16 0.84 .09 .11 .08 .12 .25 .28 .08 .20 .28 .29 .14
Together 3.37 1.04 .10 .05 .05 .04 .38 .35 .17 .12 .45 .47 .19
13. Engagement (teacher 4th grade 3.92 0.88 .37 .34 .18 .15 .01 .06 .53 .45 -.08 .13 .64 .11
report) 7th grade 3.54 0.99 .21 .05 .17 .11 -.02 .01 .54 .49 .03 .13 .44 .14 -
2nd grade 3.44 0.83 .07 .06 .04 .10 .09 .04 .60 .41 .09 .15 .49 .19
Together 3.65 0.93 .27 .15 .14 .12 .07 .08 .56 .41 .12 .23 .54 .22
14. Engagement (self- 4th grade 4.22 0.62 .15 .13 .02 .02 .22 .31 .22 .21 .28 .29 .24 .38 .29
report) 7th grade 3.59 0.65 -.05 -.09 -.11 -.12 .20 .03 .22 .28 .44 .41 .08 .49 .30 -
2nd grade 3.43 0.55 -.05 .03 -.05 .01 .05 .00 .16 .31 .40 .36 .20 .33 .27
Together 3.78 0.70 .03 .01 -.04 -.03 .24 .21 .23 .18 .52 .51 .23 .52 .36
15. Engagementa (peer- 4th grade 2.42 3.71 .37 .43 .27 .21 ,01 .04 .16 .67 -.02 .06 .29 .09 .45 .20
report) 7th grade 2.40 4.25 .09 .16 .10 .12 .01 .01 .31 .77 .12 .20 .39 .12 .50 .29 -
2nd grade 2.42 4.66 .08 .26 .06 .13 -.04 -.02 .33 .66 .06 .14 .32 .17 .45 .35
Together 2.42 1.18 .21 .30 .16 .16 -.01 0.00 .24 .70 .00 .07 .28 .08 .43 .19
16. Academic achievement 4th grade 4.41 0.79 .37 .33 .20 .18 -.04 .01 .41 .42 -.11 .09 .44 .03 .62 .25 .44
7th grade 4.14 0.90 .24 .11 .21 .14 -.05 .03 .51 .50 -.06 .08 .45 .05 .71 .26 .53 -
2nd grade 3.27 0.76 .13 .12 .14 .20 .09 .08 .43 .42 .19 .21 .27 .14 .56 .27 .45
Together 3.97 0.95 .23 .16 .17 .16 .03 .08 .42 .35 .18 .27 .38 .16 .64 .40 .37
17. Academic achievement 4th grade 2.41 4.15 .36 .47 .31 .19 .04 .02 .20 .64 -.04 .08 .36 .05 .46 .12 .68 .44
(peer report)a 7th grade 2.47 4.92 .14 .20 .14 .12 -.02 .00 .27 .74 .11 .20 .28 .07 .53 .21 .73 .53 -
2nd grade 2.43 5.46 .16 .21 .08 .14 .06 .02 .34 .64 .19 .17 .24 .21 .39 .32 .69 .52
Together 2.44 4.82 .24 .32 .20 .15 .02 -.02 .23 .67 -.01 .05 26 .03 .41 .10 .70 35
Note. Variables are arranged with regard to belonging latent variables; for measures of interpersonal peer assessment the standardized values are used in correlational analysis; correlations sig-
nificant at p≤ .01 are marked with bold; N for 4th grade is 428, for 7th grade 404, for 2nd grade 333, and for all students together 1165; a – normalized variable.
KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

49

22.1.2008 13:10:58
 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

the correlations. The correlations between measures gath- Table 2

ered from the same source (self-report, teacher report, peer Fit indices for the originally presumed model with mediating variables
report) are higher than the correlations between measures RMSEA
gathered from different sources, although they might meas- Model df χ2 GFI AGFI 90% confidence
ure the same construct. interval
Model 1: model with mediating variables
Testing the models about the relations between social together 95 1071.33** .89 .82 .103 (.097; .110)
and academic variables 4th grade 95 339.99** .90 .83 .087 (.077; .097)
7th grade 95 457.97** .86 .78 .106 (.096; .120)
2nd grade 95 369.27** .88 .80 .099 (.085; .110)
The originally presumed model described above and
some alternative models were tested using structural equa- Note. The maximum likelihood method was used; GFI – goodness-of-fit-
index; AGFI – adjusted goodness-of-fit-index; RMSEA – root mean
tion modeling. Therefore, the program Lisrel 8.50 was used.
square error of approximation.
*p ≤ .05; **p ≤ .01.
Model 1: Originally presumed model with mediating
variables

First the model in which relations to peers and teachers Also, the factor of the method of data assessment was
effect academic outcomes via students’ well-being in school taken into consideration in the model. The correlations be-
and their school engagement was tested. This model con- tween the manifest variables, gathered from the same source
sists of five latent and seventeen manifest variables and is (self-report, peer nominations, teacher report) were set as
presented in Figure 2. free parameters.
For variables that do not distribute normally the normal- In Table 2 some fit indices for this originally presumed
ized values were used. These variables are some of the peer model tested on all students together and separately for each
interpersonal assessment measures: (1) peer perceived pop- age group are presented. As suggested by Raykov and Mar-
ularity; (2) peer perceived relations with teachers; (3) peer coulides (2000) the value of RMSEA or the lower value of
perceived academic engagement and (4) peer perceived aca- the RMSEA confidence interval of less than .05 is indicative
demic achievement. of the model being a reasonable approximation of the data.
The originally presumed model did not provide a good
fit to the data. The fit for this model is somewhat better when
testing it separately for different school periods; it fits best
for the students of the 4th grade. Because of the inappropri-
ateness of the fit for this model, the model coefficients are
not presented.
The adequacy of the parts of the model was established
for independent, as well as for dependent variables using
confirmatory factor analysis. The fit for the model with rela-
tions to peers and to teachers is appropriate (RMSEA= .021).
The same is also true for the part of the model with depend-
ent variables (RMSEA= .037). However, if directed relations
between dependent variables are predicted (i.e. the influence
of well-being in school on academic engagement and aca-
demic achievement) the fit is very poor (RMSEA= .246).

Testing the influence of each kind of relations sepa-

rately

In addition, the fit of the models with each independ-

ent variable separately was tested, i.e. how the relations to
peers and to teachers can independently predict academic
Figure 2. Model 1: model of relations between social and aca-
outcomes, considering the presumed mediating variables.
demic variables with mediating variables
Note. Definitions of manifest variables are given in Method sec-
The fit indices for these two versions of the originally pre-
tion. The manifest variables assessed by the same source (self-re- sumed model are presented in Table 3.
port, peer nominations, teacher report) are correlated (separately Testing the influence of each independent variable sepa-
for dependent and for independent variables). rately shows that relations to peers and relations to teach-

50

KOSIR SOCAN.indd 50 22.1.2008 13:10:58

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Table 3 independent variable fitted to the data for the students of the
The fit indices for the versions of the originally presumed model 7th grade and for the 2nd grade of secondary school. For these
RMSEA two older age groups, the model coefficients are presented
Model df χ2 GFI AGFI 90% confidence in Figure 3. As is evident in Figure 3, the strongest relation
interval appears between students’ academic engagement and their
Model 1a: model with mediating variables - peers academic achievement.
together 52 314.13** .95 .92 .071 (.063; .079)
4th grade 52 146.50** .94 .89 .073 (.059; .087) Model 2: Model without mediating variables
7th grade 52 113.82** .95 .91 .059 (.044; .074)
2nd grade 52 97.92** .95 .91 .056 (.038; .072)
Because the originally presumed model did not prove to
Model 1b: model with mediating variables - teachers
be appropriate, the model without mediating variables was
together 37 893.70** .86 .75 .153 (.144; .180)
tested, i.e. the model in which the direct influence of rela-
4th grade 37 244.70** .89 .80 .127 (.110; .140)
7th grade 37 346.83** .84 .72 .156 (.140; .170)
2nd grade 37 176.06** .90 .82 .114 (.097; .130)
Table 4
Note. The maximum likelihood method was used; GFI – goodness-of-fit-
Fit indices for the model without mediating variables
index; AGFI – adjusted goodness-of-fit-index; RMSEA – root mean
square error of approximation. RMSEA
*p ≤ .05; **p ≤ .01. Model df χ2 GFI AGFI 90% confidence
interval
Model 2: model without mediating variables
together 35 245.56** .96 .91 .078 (.069; .088)
4th grade 35 96.25** .96 .90 .071 (.054; .088)
7th grade 35 34.56 .99 .97 .000 (.000; .038)
2nd grade 35 51.47* .97 .94 .041 (.011; .063)
Note. See note to Table 3.
*p≤ .05; **p≤ .01.

Figure 3. Parameter estimates for a variant of originally presumed

model (Model 1a) for the students of the 7th grade and the 2nd
grade of secondary school
Note. Coefficients in the first row refer to the students of the 7th
grade and those in the second row to the students of the 2nd grade
of secondary school. See also note to Figure 2.

ers differently relate to well-being, academic engagement

and academic achievement. None of the models meets the Figure 4. Parameter estimates for the model without mediating
criteria for fitting the model. However, the fit of the model variables (Model 2) for students of the 7th grade and 2nd grade of
with relations to peers as an independent variable is much secondary school
better than the fit of the model with relations to teachers as Note. Coefficients’ in the first row refer to the students of the 7th
an independent variable. When testing the models for dif- grade and those in the second row to the students of the 2nd grade of
ferent age groups, the model with relations to peers as an secondary school. See also note to Figure 2.

51

KOSIR SOCAN.indd 51 22.1.2008 13:10:59

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

tions to peers and to teachers on academic achievement is

hypothesized.
The fit indices for the model without mediating variables
are presented in Table 4.
The model without mediating variables fits better than
the originally presumed model assuming that the relation
between social and academic variables is mediated by well-
being in school and academic engagement. However, the fit
is only appropriate for the students of the 7th grade and the
students of the 2nd grade of secondary school. The model
coefficients for these two age groups can be observed in
Figure 5. Parameter estimates for the model without mediating
Figure 4.
variables with peer relations as independent variables (Model 2a)
As evident from Figure 4, the influence of relations to for all students together and for the each age group
teachers on students’ academic achievement is for both Note. Coefficients above the lines refer to all students together;
age groups much larger than the influence of peer relations coefficients below the lines refer to the students of the 4th grade
which is close to zero and negative, respectively. (first row), the students of the 7th grade (second row), and the
students of the 2nd grade of secondary school (third row). See also
Testing the influence of both kinds of relations sepa- note to Figure 2.
rately

As for the Model 1, two versions of the model with each

independent variable separately were tested for the Model
2 as well. The fit indices of these two models are presented
in Table 5.
The fit of the model without mediating variables with
peer relations as an independent variable is appropriate
for all age groups of students, as well as for all students
together. The coefficients for this model are presented in
Figure 5.
Figure 6. Parameter estimate for the model without mediating
As presented in Figure 5, the influence of peer relations variables with relations to teachers as an independent variable
on students’ academic achievement is the largest in 4th grade (Model 2b) for the students of the 2nd grade of secondary school
students, whereas in older students this influence is very low Note. See note to Figure 2.
and zero, respectively. The model with relations to teachers

as an independent variable (Model 2b) fits the data only for

Table 5 the older group of students, i.e. students of the 2nd grade of
Fit indices for the versions of the model without mediating variables secondary school. The model coefficients for this age group
RMSEA are presented in Figure 6.
Model df χ2 GFI AGFI 90% confidence
interval
Model with latent variables with regard to the method
Model 2a: model without mediating variables – peers of data assessment
together 12 34.40** .99 .97 .044 (.027; .061)
4th grade 12 26.30** .98 .94 .059 (.028; .089)
7th grade 12 17.32 .99 .96 .036 (.000; .071)
The analysis of the correlations between manifest vari-
2nd grade 12 12.66 .99 .97 .014 (.000; .063)
ables indicated that variables assessed from the same source
(self-report, peer nominations, teacher assessment) were
Model 2b: model without mediating variables – teachers
more strongly correlated. Therefore, also the model in which
together 7 160.18** .95 .85 .149 (.130; .170)
latent variables are formed with regard to the method of data
4th grade 7 48.01** .96 .87 .129 (.096; .170)
assessment was tested. In this model, the correlations with
7th grade 7 58.74** .95 .84 .147 (.110; .180)
regard to the construct (i.e. peer relations, relations with
2nd grade 7 17.42* .98 .94 .072 (.030; .110)
teachers, well-being, academic engagement, and academic
Note. See note to Table 3. achievement) were set as free parameters. The fit indices for
*p≤ .05; **p≤ .01. this model are presented in Table 6.

52

KOSIR SOCAN.indd 52 22.1.2008 13:10:59

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Table 6 For all students together, the model with latent variables
Fit indices for the model with latent variables with regard to the method with regard to the method of data assessment does not fit to
of data assessment the data. However, the RMSEA value is quite close to .05.
RMSEA The fit is appropriate for the students of the 2nd grade of sec-
Model df χ2 GFI AGFI 90% confidence ondary school. The model coefficients for this age group are
interval shown in Figure 7. In any case, the fit for all age groups is
Model 3: model with latent variables with regard to the method of data better than for the originally presumed model with mediat-
assessment ing variables. This is surprising considering the fact that the
together 90 521.76** 0.94 .90 .070 (.064; .076) variables in the originally presumed model are formed with
4th grade 90 326.23** 0.93 .87 .069 (.058; .079) regard to their contents, whereas in this model all that vari-
7th grade 90 251.57** 0.92 .87 .072 (.061; .083) ables within a specific latent variable have in common is the
2nd grade 90 194.03** 0.93 .87 .064 (.049; .076) method of data assessment.
Note. See note to Table 3. As can be observed in Figure 7, the relations between
*p≤ .05; **p≤ .01. measures of peer nominations and teacher assessment are
stronger compared to the measures of self-report that weak-
ly correlate with other measures.

DISCUSSION
The main purpose of the study was to test the model of
relations between social and academic variables in school
context. Therewith, we wanted to establish the relation
between social and academic variables in different school
periods and test whether affective and motivational factors
act as mediating variables in the relation between social and
academic variables in school.
The model that hypothesizes that the relation between
social (i.e. peer and teacher relations) and academic vari-
ables is indirect via students’ well being in school and their
academic engagement did not provide the appropriate fit
to the data. Additional assumptions about the nature of
the relations between social and academic variables were
therefore made. The model without mediating variables that
assumes the direct influence of peer and teacher relations
on students’ academic achievement proved to be more ad-
equate. The fit indices of this model were better than for
the originally presumed model with mediating variables.
However, the fit was appropriate only for the students of
Figure 7. Parameter estimates for the model with latent variables the 7th grade and the students of the 2nd grade of secondary
with regard to the method of data assessment (Model 4) for the
school. In these two age groups, the influence of relations to
students of the 2nd grade of secondary school
teachers on students’ academic achievement is much larger
Note. The manifest variables for the latent variable self-report
are the following: S1 – teacher academic support; S2 – teacher compared to the influence of peer relations, which is close
personal support; S3 – peer academic support; S4 – peer personal to zero and slightly negative, respectively.
support; S5 – self-reported well-being in school; S6 – self-report-
ed academic engagement. Manifest variables of the latent variable The relation between social and academic variables in
peer nominations are: PN1 – social preference; PN2 – peer per- students
ceived popularity; PN3 – number of friendships; PN4 – number
of reciprocal friendships; PN5 – peer perceived relations with
teachers, PN6 – peer perceived academic engagement; PN7 – peer
The results indicate that the originally presumed mod-
perceived academic achievement. The latent variable teacher el is not adequate because of the mediating variables. The
assessment consists of the following variables: TA1 – teacher’s model without mediating variables fits better than the model
liking of the student; TA2 – teacher report of students’ well-being with mediating variables1. For the model that presumes only
in school; TA3 – teacher report of students’ academic engagement;
TA4 – GPA. Manifest variables that represent the measures of the
same construct (peer relations, relations to teachers, well-being, 1 This can also be inferred from the following results that we do not refer
academic engagement, academic achievement) are correlated. here in detail because of the lack of space: (1) If the latent variables

53

KOSIR SOCAN.indd 53 22.1.2008 13:11:00

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

peer relations as an independent variable, the fit is also bet- and academic variables, it is reasonable to assume that
ter if mediating variables are not included in the model. The the relation between social and academic variables de-
inclusion of mediating variables decreases the explanatory pends on the social norms that are formed in the peer
value of the model. This indicates that the presumed mediat- group. These norms are developmentally specific and
ing variables are not adequate. are especially prone to changes when students enter ad-
olescence. In younger students, it is usually the teacher
The finding that emotional and motivational variables
that influences the norms, rules, and values that are
are not the factors mediating the influence of social relations
valid in the peer group. Consequently, students’ norms
students form in school on their academic outcomes is not regarding school work and learning are very similar
congruent with the findings of some other studies. Some au- to the expectations that teachers and other adults hold
thors report motivational variables as those that mediate the toward students. However, when entering adolescence
relation between students’ relations in school/family con- the influence of the peer group becomes increasingly
text and their academic achievement (e.g. Grolnick, Ryan, important. Adolescents often believe that two different
& Deci, 1991; Wentzel & Caldwell, 1997). Alternative as- sets of behavioral and academic attributes are needed
sumptions about the relation between social and academic to be popular with peers versus preferred by teachers
variables should thus be made. Some possible interpreta- (Hopmeyer Gorman, Kim, & Schimmelbusch; 2002).
tions of this relation are mentioned below: In adolescents, the peer group presents a very influen-
(1) The relation between social and academic variables is tial context for the development of academic beliefs
direct. and behaviors (Ryan, 2001). It is possible that the so-
(2) Some third variable (e.g. self-regulatory skills or self cial norms that are formed in students’ peer group are
variables as self-concept or self-efficacy) is behind the the reason for a stronger connection between peer rela-
relation between both groups of variables. The results tions and academic achievement in the period of late
of the Buhs’ (2005) study speak in favor of this as- childhood as compared to adolescence.
sumption. These results showed that the influence of Also, the possibility that the model does not fit the data
peer relations on academic outcomes is mediated by because they are coming from different sources should be
students’ perceptions of their own competence. Went- taken into consideration.
zel and Asher (1995) also suggest that students’ self-
regulatory skills are those influencing students’ effec-
The relation between social and academic variables in
tive performance in social, as well as academic field.
different school periods
(3) The positive relation between students’ peer relations
and their academic achievement is due to the relations
students form with their teachers. Teachers usually Considering the social development in late childhood
prefer students which perform better in school. It is and adolescence it is reasonable to assume that the impor-
possible that because of the teacher preference, these tance of both kinds of relations (i.e. relations to peers and to
students could also be more accepted by their peers. teachers) for the academic achievement changes during the
different school periods. For that reason, the relation of both
(4) The relation between social and academic variables kinds of relationships that students form in school to their
depends on developmentally specific norms which academic achievement was investigated in students in dif-
form among peers with regard to learning and learning ferent school periods. The comparison between the students
behavior. Namely, the interpretations mentioned above of different ages is possible only in the model that assumes
can not explain the fact that in older students the corre- the direct influence of peer relations to students’ academic
lations between peer relations and academic variables achievement (Model 2a). This is namely the only model that
are much lower compared to younger students. For ex- fits the data in all of the three age groups. The model’s coef-
ample, it is not very likely that a decrease in students’ ficients between the peer relations and students’ academic
self-regulation skills occurs when entering adolescence achievement are highest for the students of the 4th grade
(actually, the research shows the opposite; see e.g.
and lower in both of the older age groups. Thus, the crucial
Paris & Newman, 1990). Because of the age depend-
change in the influence of peer relations obviously occurs in
ence on the strength of the connection between social
the period between ten and thirteen years of age. Most like-
ly, a decrease in the influence of peer relation on academic
outcomes coincides with the transition to adolescence.
peer relations and relations to teachers are excluded from the model
and only the relation between those variables that represent dependent In adolescence, the influence of peer relations on stu-
variables in originally presumed model is tested, and directed links dents’ academic achievement is low. The students that form
between those variables are presumed (i.e. well-being – engagement more positive relations with peers are not those students that
– academic achievement), the fit of such model is very poor (RM-
SEA=0.246). (2) If the influence of both independent variables is
are also academically better (this holds true for the students
tested separately for each mediating variable (well-being and engage- of the 4th grade). Some models of relations between social
ment), none of those models fits to the data. and academic variables fit only to the two older age groups

54

KOSIR SOCAN.indd 54 22.1.2008 13:11:00

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

of students. These models enable the comparison between ables develops gradually – the teacher’s liking positively
the students of the 7th grade and the students of the 2nd grade influences students’ academic achievement, which in turn
of secondary school. There are no essential differences in strengthens the positive relation that these students form
the model’s coefficients between different variables in the with their teachers.
models of these two age groups. Apparently, there are no
differences in the relation between measures of peer rela- Evaluation of the methods for assessing the data about
tions and academic variables between early and middle ado- students’ social and academic characteristics
lescents. Nevertheless, it should not be forgotten that we do
not posses longitudinal data and is, therefore, impossible to
It this study at least two different measures of each con-
draw conclusions about developmental changes.
struct were used to assess the latent variables included in the
models. On the one hand, the purpose of such assessment
The relation between both kinds of social relations in was to validly assess each latent variable, but at the same
different school periods time we were also interested in the relations between the
different measures of the constructs with each other. The
For the most part, the correlations between the measures analysis of the correlations between all manifest variables
of peer relations and relations to teachers decrease with stu- already shows that the correlations between the measures
dents’ increasing age. This is true only for the measures that of the same construct, assessed from the different sources,
are assessed using peer and teacher report. The self-report are mostly lower than the correlations between the meas-
measures relate very weakly to other measures. ures of different constructs, assessed from the same source.
In late childhood, students that form more positive rela- Therefore, the model with latent variables taking into the
tions with peers are also in better relations with their teach- account the method of data assessment was tested. Since
ers. But in adolescence (i.e. students of the 7th grade and the latent variables in this model are designed only with regard
2nd grade of secondary school) the relations which students to the method of data assessment and include manifest vari-
form with peers and with teachers become less related to ables from different constructs it would be reasonable to
each other. In this period, the correlations between both con- expect that the model does not fit to the data. However, the
structs are still positive, but low. Thus, the findings of some results show a different picture. The fit indices are appropri-
studies which show that students who form more positive ate only for the oldest group of students, whereas for both
of the younger age groups these indices are quite close to
relations with peers are also more liked by the teachers (e.g.
the recommended limit values. In any case, the fit is much
Birch & Ladd, 1997; Donohue, Perry, & Weinstein, 2003;
better than for the originally presumed model with mediat-
Taylor, 1989) can not be generalized to adolescents without
ing variables that consists of latent variables designed with
some caution.
respect to the construct they are supposed to measure. Such
Also the differences in the importance of both kinds of results seem quite alarming and have some important im-
relations for students’ academic achievement were found. plications for psychological research. Namely, especially
The results of the structural equation modeling indicate that for the research in the past it was common that only one
the effect of peer relations on students’ academic achieve- source of data assessment was used in the study. According
ment depends on the students’ age. For students in late child- to the findings of our study the relations between variables,
hood, peer relations are important factors of their academic as found in such studies, can also be interpreted as an arti-
achievement, whereas in older students the model’s coef- fact of the method of data assessment and not as the actual
ficients between peer relations and academic achievement (content) relation between variables. The possibility of such
are low. In contrast to this, the importance of relations to interpretation questions the findings of numerous psycho-
teachers for students’ academic achievement is not age-de- logical studies that are based upon one single source of data
pendent. The correlations between the measures of students’ assessment.
teacher relations and academic achievement are low to mod- The analysis of the correlations between all manifest
erate in all periods of schooling. Peer relations to teachers variables indicates that the measures of self-report are es-
are important for students in all periods of schooling. Furrer pecially weakly connected with other variables, whereas the
& Skinner (2003) explain this relation with the sense of re- peer and teacher report variables are more strongly corre-
latedness that is crucial for students’ active engagement in lated. This is true for all three age groups of students and,
school activities, which leads to better academic outcomes. therefore, can not be attributed to the factors of students’
On the other hand, it is also possible that the direction of cognitive development (e.g. insufficiently developed meta-
that relation is quite the opposite – in all periods of school- cognitive skills in younger students).
ing teachers might prefer the students that are more engaged Similar discrepancies between measures of self-report
and achieve more. and other measures of data assessment were also reported in
However, irrespective of the original direction of the re- other studies (Bierman, 2004; Humphrey, 1984; Rohrbeck,
lation it is likely that the reciprocal link between both vari- Azar, & Wagner, 1991). Nevertheless, Meyers, Cohen, and

55

KOSIR SOCAN.indd 55 22.1.2008 13:11:00

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Schleser (1991) suggest that this lack of relationship among J. Juvonen & K. R. Wentzel (Eds.). Social motivation.
the raters should not be interpreted as an error variance. It Understanding children’s school adjustment (pp. 199-
simply reflects different conceptualizations of the student’s 225). New York: Cambridge University Press.
everyday context. Therefore, on the basis of the inconsist- Birch, S.H., & Ladd, G.W. (1997). The teacher-child rela-
ency between measures of self-report on the one hand, and tionship and children’s early school adjustment. Jour-
peer- and teacher-report measures on the other hand, it is not nal of School Psychology, 35, 61-79.
justified to make conclusions about which source of data is
Buhs, E. (2005). Peer rejection, negative peer treatment, and
more accurate or more valid. Nevertheless, the mere use of
school adjustment: Self-concept and classroom engage-
self-report measures, which is quite frequent in psychologi-
ment as mediating processes. Journal of School Psy-
cal studies, seems quite controversial.
chology, 43, 407-424.
Chen, X., Chang, L., & He, Y. (2003). The peer group as a
CONCLUSIONS context: Mediating and moderating effects on relations
between academic achievement and social functioning
Investigating the social factors of students’ academic in Chinese children. Child Development, 74, 710-727.
outcomes contributes to a better understanding of students’
Coie, J.D., Dodge, K.A., & Coppotelli, H.A. (1982). Di-
academic (under)achievement or failure and enables a more
mensions and types of social status: A cross-age per-
complete intervention in case of academic failure and es-
spective. Developmental Psychology, 18, 557-569.
pecially preventive functioning. Relating social, emotional
and motivational factors and academic achievement is a Connel, J.P., & Wellborn, J.G. (1991). Competence, au-
step towards a conceptualization of school adjustment as the tonomy, and relatedness: A motivational analysis of
most holistic educational goal. Such simultaneous investi- self-system processes. In M.R. Gunnar & L.A. Sroufe
gation of different factors of school adjustment and relations (Eds.). Self processes and development: The Minnesota
between them can represent a basis for the formation of the symposia on child development. Vol. 23 (pp. 43-78).
school adjustment theory that is, as emphasized by Wentzel Hillsdale: Erlbaum.
(2003), not well developed at the moment. The implications Covington, M.V. (2000). Goal theory, motivation, and
that arise from understanding the relations between social school achievement: An integrative review. Annual Re-
relations and students academic achievement in different view of Psychology, 51, 171-200.
school periods enable the adaptation of the interventions to Deci, E. L., & Ryan, R.M. (2000-2001). Self-determination
students’ developmental needs and qualification of teachers theory. An approach to human motivation and personal-
for such work. ity. Retrieved July 22, 2002 from Rochester University
From the methodological point of view, the findings Web site: http://www.psych.rochester.edu/SDT/index.
of this study indicate the necessity of a methodological html
triangulation. Also, they emphasize the need for a critical Donohue, K.M., Perry, K.E., & Weinstein, R.S. (2003).
evaluation of the existing studies that have investigated the Teachers‘ classroom practices and children‘s rejection
relations between different variables and are based on one by their peers. Journal of Applied Developmental Psy-
single method of data assessment. chology, 24, 91-118.
Dweck, C.S. (1996). Social motivation: Goals and social-
REFERENCES: cognitive processes. A comment. In J. Juvonen & K.
R. Wentzel (Eds.). Social motivation. Understanding
Assor, A., Kaplan, H., & Roth, G. (2002). Choice is good, children’s school adjustment (pp. 181-195). NY: Cam-
but relevance is excellent: Autonomy-enhancing and bridge University Press.
suppressing teacher behaviors predicting students’ en- Eccles, J.S., & Midgley, C. (1989). Stage-environment fit:
gagement in schoolwork. British Journal of Education- Developmentally appropriate classrooms for young
al Psychology, 72, 261-278. adolescents. In C. Ames & R. Ames (Eds.). Research
Austin, A.B., & Draper, D.C. (1984). The relationship on motivation in education. Vol. 3. (pp. 139-186). New
among peer acceptance, social impact, and academic York: Academic Press.
achievement in middle school. American Educational Furrer, C., & Skinner, E. (2003). Sense of relatedness as a
Research Journal, 21, 597-604. factor in children’s academic engagement and perfor-
Bierman, K. (2004). Peer rejection. Developmental process- mance. Journal of Educational Psychology, 95, 148-
es and intervention strategies. New York: The Guilford 162.
Press. Gifford-Smith, M.E., & Brownel, C.A. (2003). Childhood
Birch, S.H., & Ladd, G.W. (1996). Interpersonal relation- peer relationships: social acceptance, friendships, and
ships in the school environment and children’s early peer networks. Journal of School Psychology, 41, 235-
school adjustment: The role of teachers and peers. In 284.

56

KOSIR SOCAN.indd 56 22.1.2008 13:11:00

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Goodenow, C. (1993). Classroom belonging among early Meyers, A.W., Cohen, R., & Schleser, R. (1989). A cog-
adolescents students: relationships to motivation and nitive-behavioral approach to education: Adopting a
achievement. Journal of Early Adolescence, 13, 21-43. broad-based perspective. In J.N. Hughes & R.J. Hall
Grolnick, W.S., Ryan, R.M., & Deci, E.L. (1991). Inner re- (Eds.). Cognitive behavioral psychology in the schools
sources for school achievement: Motivational media- (pp. 62-84). New York: The Guilford Press.
tors of children’s perceptions of their parents. Journal Moser, U., Bless, G., & Haeberlin, U. (1989). Fragebogen
of Educational Psychology, 83, 508-517. zur Erfassung von Dimensionen der Integration von
Guay, F., Boivin, M., & Hodges, E. (1999). Predicting Schulern (FDI 4-6) [Questionnaire for assessing the
change in academic achievement. A model of peer ex- dimensions of students' integration]. Psychologie in Er-
periences and self-system processes. Journal of Educa- ziehung und Unterricht, 36, 19-26.
tional Psychology, 91, 105-115. Murdock, T., Miller, A., & Kohlhardt, J. (2004). Effects of
Hatzichristou, C., & Hopf, D. (1996). A multiperspective classroom context variables on high school students‘
comparison of peer sociometric status groups in child- judgements of the acceptability and likelihood of cheat-
ing. Journal of Educational Psychology, 96, 765-777.
hood and adolescence. Child Development, 67, 1085-
1102. Murdock, T.B., Hale, N.M., & Weber, M.J. (2001). Predic-
tors of cheating among early adolescents: Academic
Hopmeyer Gorman, A., Kim, J., & Schimmelbusch, A.
and social motivations. Contemporary Educational
(2002). The attributes adolescents associate with peer
Psychology, 26, 96-115.
popularity and teacher preference. Journal of School
Psychology, 40, 143-165. Ollendick, T.H., Weist, M.D., Burden, M.C., & Greene,
R.W. (1992). Sociometric status and academic, behav-
Humphrey, L.L. (1984). Children’s self-control in relation ioral, and psychological adjustment: A five-year longi-
to perceived social environment. Journal of Personality tudinal study. Journal of Consulting and Clinical Psy-
and Social Psychology, 46, 178-188. chology, 60, 80-87.
Johnson, D.W., Johnson, R., & Anderson, D. (1983). Social Paris, S.G., & Newman, R.S. (1990). Developmental as-
interdependence and classroom climate. Journal of Psy- pects of self-regulated learning. Educational Psycholo-
chology, 114, 135-142. gist, 25, 87-102.
Juvonen, J. (1996). Self-presentation tactics promoting Raykov, T., & Marcoulides, G.A. (2000). A first course in
teacher and peer approval: The function of excuses and structural equation modeling. Mahwah: Lawrence Erl-
other clever explanations. In J. Juvonen & K. R. Went- baum Associates.
zel (Eds.). Social motivation. Understanding children’s Rohrbeck, C.A., Azar, S.T., & Wagner, P.E. (1991). Child
school adjustment (pp. 43-65). New York: Cambridge self-control rating scale: Validation of a child self-re-
University Press. port measure. Journal of Clinical Child Psychology, 20,
Keller, C., Moser, U., & Rhyn, H. (1996). Evaluation der 179-183.
Sekundarstufe I im Kanton Zürich. Methodisches Vor- Ryan, A.M. (2001). The peer group as a context for the de-
gehen, Instrumente und Kennwerte zum Bericht Schul- velopment of young adolescent motivation and achieve-
systemvergleich [Evaluation of the secondary level in ment. Child Development, 72, 1135-1150.
canton Zuerich. Methodical procedures, instruments, Ryan, A.M., Gheen, M., & Midgley, C. (1998). Why do
and characteristics to the report The comparation of some students avoid asking help? An examination of
the school systems] (unpublished report). Bern: Institut the interplay among students’ academic efficacy, teach-
für Pädagogik, Abteilung Allgemeine Pädagogik. er’s social-emotional role and classroom goal structure.
Kindermann, T. A. (1993). Natural peer groups as contexts Journal of Educational Psychology, 90, 528-535.
for individual development: The case of children‘s mo- Ryan, A.M., & Patrick, H. (2001). The classroom social en-
tivation in school. Developmental Psychology, 29, 970- vironment and changes in adolescents’ motivation and
977. engagement during middle school. American Educa-
Kupersmidt, J.B., Buchele, K.S., Voegler, M.E., & Se- tional Research Journal, 38, 437-460.
dikides, C. (1996). Social self-discrepancy: a theory Ryan, R.M., Stiller, J.D., & Lynch, J.H. (1994). Represen-
relating peer relations problems and school maladjust- tations of relationships to teachers, parents and friends
ment. In J. Juvonen & K. R. Wentzel (Eds.). Social mo- as predictors of academic motivation and self-esteem.
tivation. Understanding children’s school adjustment Journal of Early Adolescence, 14, 226-249.
(pp. 66-93). New York: Cambridge University Press. Skinner, E. A., & Belmont, M.J. (1993). Motivation in the
Ladd, G.W, Birch, S.H., & Buhs, E.S. (1999). Children’s so- classroom: Reciprocal effects of teacher behavior and
cial and scholastic lives in kindergarten: related spheres student engagement across the school year. Journal of
of influence? Child Development, 70, 1373-1400. Educational Psychology, 85, 571-581.

57

KOSIR SOCAN.indd 57 22.1.2008 13:11:01

 KOŠIR, SOČAN and PEČJAK, Interpersonal relationships and academic achievement, Review of Psychology, 2007, Vol. 14, No. 1, 43-58

Taylor, A. (1989). Predictors of peer rejection in early el- Wentzel, K.R. (1997). Student motivation in middle school:
ementary grades: Roles of problem behavior, academic The role of perceived pedagogical caring. Journal of
achievement, and teacher preference. Journal of Clini- Educational Psychology, 89, 411-419.
cal Child Psychology, 18, 360-365.
Wentzel, K.R. (1998). Social relationships and motivation in
Weiner, B. (1996). Foreword. In J. Juvonen & K. R. Went- middle school: The role of parents, teachers, and peers.
zel (Eds.). Social motivation. Understanding children’s Journal of Educational Psychology, 90, 202-209.
school adjustment (pp. xiii-xv). New York: Cambridge
University Press. Wentzel, K.R. (2003). School adjustment. In W.M. Reyn-
olds & G.E. Miller (Eds.). Handbook of psychology.
Wentzel, K. R. (1991). Relations between social compe-
Volume 7. Educational psychology (pp. 235-258). New
tence and academic achievement in early adolescence.
York: John Willey & Sons.
Child Development, 62, 1066-1078.
Wentzel, K.R. (1993). Does being good make the grade? So- Wentzel, K.R., & Asher, S.R. (1995). Academic lifes of
cial behavior and academic competence in middle school. popular, average, and controversial children. Child De-
Journal of Educational Psychology, 85, 357-364. velopment, 66, 76-97.
Wentzel, K.R. (1994). Relations of social goal pursuit to Wentzel, K.R., McNamara Barry, C., & Caldwell, K.A.
social acceptance, classroom behavior, and perceived (2004). Friendships in middle school: Influences on
social support. Journal of Educational Psychology, 86, motivation and school adjustment. Journal of Educa-
173-182. tional Psychology, 96, 195-203.
Wentzel, K.R. (1996a). Social goals and social relationships Wentzel, K.R., & Caldwell, K. (1997). Friendships, peer ac-
as motivators of school adjustment. In J. Juvonen & K. ceptance, and group membership: Relations to academ-
R. Wentzel (Eds.). Social motivation. Understanding ic achievement in middle school. Child Development,
children’s school adjustment (pp. 226-247). New York: 68, 1198-1209.
Cambridge University Press. White, K.J., Sherman, M.D., & Jones, K. (1996). Children’s
Wentzel, K.R. (1996b). Social and academic motivation in perceptions of behavior problem peers: Effects of teach-
middle school: Concurrent and long-term relations to aca- er feedback and peer-reputed status. Journal of School
demic effort. Journal of Early Adolescence, 16, 390-406. Psychology, 1, 53-72.

58

KOSIR View
SOCAN.indd
publication stats 58 22.1.2008 13:11:01