Frontline Learning Research Vol.5 No. 2 (2017) 60
- 77
ISSN 2295-3159
University of Potsdam, Germany
Article received 14 February / revised 18 May / accepted 2 August / available online 7 August
Information on social comparison is one of the major factors used to evaluate academic achievement. The presence of Big-Fish-Little-Pond (BFLP) and Basking-In-Reflected-Glory (BIRG) effects of academic achievement on the self-concept have been extensively researched in various observational studies. Recent research suggests that these effects can also be transferred to motivational variables such as task interest. This paper uses an experimental paradigm to take a closer look at the mechanisms expected to be behind BFLP and BIRG effects: contrast and assimilation effects of task performance. The analyses are based on N = 129 primary education students who completed a computer-based learning task. During this task, participants received social comparative feedback that was experimentally manipulated based on 2x2 conditions: social position (high vs. low) and peer performance (high vs. low). Task interest was measured both before the start of the learning task and after it was finished. Results indicate a positive influence of high social position and high peer performance on the development of task interest from pre- to post-test compared to the low social position and peer performance conditions. Further, the hypothesized positive association between self-concept, task interest and performance in the academic learning task could also be shown.
Keywords: social comparison; feedback; interest; contrast effect; assimilation effect
A lot of time during childhood, adolescence, and early adulthood is spent in institutionalized learning situations, such as school, vocational training, or university. Usually, these institutions are group-based and therefore social in nature. While these settings enable the use of group-based learning activities that can have several positive effects on learning achievement and attitudes towards learning (Springer, Stanne, & Donovan, 1999), they also facilitate the use of social comparison information in the evaluation of students' learning progress or achievement. Hence, information on social comparison is one of the major factors used to evaluate academic achievement in many educational contexts, both formally (e.g., grades) and informally (e.g., contact with peers). While social comparison can have desirable effects, it can also have considerable downsides, such as negative affect and lack of self-confidence, especially for lower-achieving students: “In classrooms characterized by frequent grades and public evaluation, students become focused on their ability and the distribution of ability in the classroom group. Many students not only come to believe that they lack ability but this perception also becomes shared among peers” (Ames, 1992, p. 264). This becomes even more relevant considering the increasing heterogeneity of academic learning groups that can be observed lately, due to the implementation of the “Convention of the Rights of Persons with Disabilities” in 2008 and the corresponding move to an inclusive school system in many countries. Processes of social comparison not only present a risk for low-achieving students to develop a negative self-concept, but can also influence academic achievement itself. Studies have shown that motivational (i.e. interest; Möller, Retelsdorf, Köller, & Marsh, 2011; Schiefele, Krapp, & Winteler, 1992) and self-related emotional (i.e. self-concept; Guay, Marsh, & Boivin, 2003; Huang, 2011; Marsh, Trautwein, Lüdtke, Köller, & Baumert, 2005; Möller et al., 2011; Valentine, DuBois, & Cooper, 2004) variables play a very important role in the academic learning process. Further, the relation between academic achievement and both interest and self-concept seems to be bidirectional, as research reveals that low academic achievement in turn leads to a negative academic self-concept and hampers the development of interest in an academic topic (Denissen, Zarrett, & Eccles, 2007; Marsh et al., 2005), which can potentially lead to a vicious circle of negative self-affect, diminished interest and negative performance feedback in academic subjects. Hence, the establishment of motivating and emotionally stabilizing evaluation practices should also be a major goal to ensure successful institutionalized learning for every student.
To be able to do that, we need to analyse how the use of social comparison data in academic evaluation variously influences the interest and self-concept of students with heterogeneous prerequisites. Although there is ample evidence of the impact socially comparative feedback has on the development of an individual's self-concept, the role of interest and motivation has been considerably less well scrutinized. That said, theoretical models, as well as empirical evidence, propose a tight connection between self-concept and interest in an academic context (e.g. Fryer, 2015). Both constructs are domain-specific, and the associations between them always correspond to a specific domain (i.e. math self-concept is associated with math interest). Obviously, self-concept and interest are two distinctive constructs nonetheless: Interest, on one hand, is the motivational orientation of a person toward a certain object, activity, or area of knowledge that is independent of any direct external rewards and can therefore be described as domain-specific intrinsic motivation (Schiefele, 1992). Self-concept, on the other hand, consists of a person’s perceptions of him- or herself. These perceptions are formed based on self-evaluation as well as the evaluation of significant others. That means that self-concept is self-ascribed competence in a specific domain (Marsh & Shavelson, 1985; Shavelson, Hubner, & Stanton, 1976). The Self-Determination Theory (Deci & Ryan, 1985), one of the most popular motivational theories, has identified feelings of competence in a specific domain as crucial for the establishment of intrinsic motivation in that same domain, suggesting a possible explanation for empirical connections between self-concept and interest found in several studies (e.g. Fryer, 2015; Schurtz, Pfost, Nagengast, & Artelt, 2014).
Taking the feeling of competence into focus as a bridge between self-concept and interest, empirical evidence for the effects of social comparison on self-concept will be presented in the following section. According to the Internal/External Frame of Reference Model by Marsh (I/E-Model; 1986) development of the self-concept is based on comparison of perceived competence in a specific domain with both internal (i.e. perceived competence in other domains) and external (i.e. perceived competence of other persons in that same domain) frames of reference. While the former can be expected to have an overall neutral effect on the sum of self-concepts in different domains (e.g. a person perceives himself as good in math but bad in arts), the latter can have predominantly positive or negative effects on the self-concept in various domains. Hence, social comparison processes are a major factor for the development of a positive self-concept, especially when the evaluative atmosphere of the typical classroom makes it hard for students not to compare themselves with each other (Dijkstra, Kuyper, van der Werf, Buunk, & van der Zee, 2008). As stated earlier, social comparison processes can be detrimental to the development of a positive self-concept in low-achieving students. The results of a study by Dickhäuser and Galfe (2004) suggest that the direction of comparison (i.e. comparison with a better or worse performing peer) has an influence on the development of the self-concept. In their study, students were asked to whom they compared themselves after their math test scores were announced. Additionally, math self-concept was assessed before and after the math test. The results showed a negative effect of upward social comparison (i.e. comparison with a better performing peer) on math self-concept, while downward social comparison (i.e. comparison with a worse performing peer) tended to produce more positive effects. Hence, the predominance of upward comparison targets might be one factor responsible for the previously stated negative effects of social comparison on self-concept development in low-achieving students. This notion is further supported by two studies that investigated the effects of ability grouping on self-concept: The first study showed that ability grouping leads to decreased academic self-concept for those in higher ability groups compared to those in lower ability groups (Mulkey, Catsambis, Steelman, & Crain, 2005), while the second study found a similar effect of lowered academic self-concept for academically handicapped special school students who were integrated into mainstream schooling (Strang, Smith, & Rogers, 1978).
Therefore, the influence of social comparison on self-evaluation, and its effects on self-concept, cannot be properly understood without considering the composition of the reference group. The intimate nature of the process connecting self-evaluation and the attributes of social comparison groups have been outlined in the Inclusion/Exclusion Model (IEM; Bless & Schwarz, 2010; Schwarz & Bless, 1992). According to the IEM, evaluation of a particular stimulus is based on the stimulus itself and a standard of comparison. Because humans can never consider all potentially relevant information, both the stimulus itself and the standard of comparison are defined by the subset of information that is most accessible at the time of judgment. The model assumes two possible effects of the standard of comparison on the evaluation of the target: a contrast effect, i.e. the use of the standard of comparison as a reference against which to judge any characteristic of the target, and an assimilation effect, i.e. the transfer of characteristics of the standard of comparison to the target. The IEM covers both situations in which the target is superordinate to the standard of comparison, i.e. a group is judged in relation to an individual that is part of the group, and situations in which the target is subordinate to the standard of comparison, i.e. a member of the group is judged in relation to the group. If we apply the IEM to the process of self-evaluation, the target (e.g., a student) is part of a superordinate category (e.g., the class or school). While there are various sources of information that can be used for self-evaluation, information about other students in one’s class (or school) is very salient and readily available in school contexts and will therefore most likely play a major role in school-based self-evaluations. Depending on which information is actually used to evaluate the self, both contrast effects, i.e. the usage of one’s class (or school) as a reference against which to judge any of one's own characteristics, or assimilation effects, i.e. the transfer of characteristics of the students of one’s class (or school) to oneself, can occur (see Bless & Schwarz, 2010 for a detailed discussion of influencing factors).
The concept of assimilation and contrast effects supposedly lies at the core of one of the most popular phenomena in educational research. Based on their observation that students attending schools with a low average socio-economic status (SES) had higher academic self-concepts than equally performing students attending schools with a higher average SES, Marsh and Parker (1984) came up with the big-fish-little-pond (BFLP) and basking-in-reflected-glory (BIRG) effects. The BFLP effect is the transfer of a contrast effect to an institutionalized learning context; it states that self-evaluation of performance is based on the average performance of other students in the same school, thus resulting in a higher academic self-concept for a student in a lower-performing school compared to an equally performing student in a higher-performing school. The BIRG effect is the transfer of an assimilation effect to an institutionalized learning context and has contrary expectations: A student in a high-performing school will have a higher academic self-concept than an equally performing student in a lower-performing school, because both students identify with their reference group and adapt their academic self-concept based on their group affiliation. However, because of a relatively stronger BFLP effect, the net result of the reference group's strength in terms of academic self-concept is still expected to be negative (Marsh, 1987). Since then, the BFLP effect on academic self-concept has been confirmed in school children from different countries (Marsh et al., 2005; Schurtz et al., 2014; Wouters, De Fraine, Colpin, Van Damme, & Verschueren, 2012), as well as in larger studies covering schools from various culturally and economically diverse countries (Chiu, 2012; Marsh & Hau, 2003; Seaton, Marsh, & Craven, 2009; Wang, 2015). Similar BFLP effects could also be found in young adults (Jonkmann, Becker, Marsh, Lüdtke, & Trautwein, 2012). Further, several studies have also shown empirical evidence for the existence of an - albeit usually smaller - BIRG effect in children and young adults (Huguet et al., 2009; Marsh, Kong, & Hau, 2000; Preckel & Brüll, 2010; Trautwein, Lüdtke, Marsh, & Nagy, 2009).
While the BFLP effect has mainly been investigated regarding academic self-concept, there are also several studies showing that the underlying processes may additionally affect other self-related constructs, as well as emotional variables. More specifically, these include BFLP effects on self-related constructs such as self-esteem (Jansen, Scherer, & Schroeders, 2015) and self-efficacy (Marsh, Trautwein, Lüdtke, & Köller, 2008), on general motivational variables such as educational and occupational aspirations (Marsh, 1991; Marsh & O’Mara, 2010; Nagengast & Marsh, 2011), on control expectations and strategies (Marsh et al., 2008), and on emotional variables such as test anxiety (Goetz, Preckel, Zeidner, & Schleyer, 2008; Zeidner & Schleyer, 1999) and emotionality (Goetz et al., 2008). Additionally, a similar BFLP effect could also be shown in regard to task interest (Köller, Schnabel, & Baumert, 2000; Schurtz et al., 2014; Trautwein, Lüdtke, Marsh, Köller, & Baumert, 2006), possibly reflecting the close connection between self-concept and interest. Trautwein and colleagues (2006) additionally tested for a BIRG effect on task interest. Although they did not find evidence for a BIRG effect on task interest, the failure to reproduce the original BIRG effect on self-concept renders those results somewhat fragile. Even though the presented results provide indirect evidence for the existence of contrast and assimilation effects in institutionalized education contexts, the non-experimental nature of the described studies still leaves some questions in relation to the assumed modes of action unanswered.
In order to be able to investigate the direct influence of social comparative performance feedback (i.e. contrast and assimilation effects) on task interest, experimental designs explicitly manipulating the performance feedback are necessary. To the author’s knowledge, only one study used such an experimental design to show that contrast effects of performance feedback have a direct influence on task interest (Pohlmann & Möller, 2006). A previous experimental study by Wilbert, Grosche, and Gerdes (2010) could also show that social comparison feedback tends to increase task-specific motivation in high-achieving students while undermining low-achieving students’ motivation. However, they did not control for the fact that high-achieving students usually receive positive feedback, while low-achieving students receive negative feedback.
In this paper, we want to further investigate the assumed modes of action behind the BFLP and BIRG effects: the direct influence contrast and assimilation effects in social comparative performance feedback have on students' interest in an academic learning task. The only study directly investigating the influence of social comparative performance feedback on task interest could find evidence for a contrast effect (Pohlmann & Möller, 2006). They did not, however, test for potential assimilation effects. Therefore, we want to investigate the directs effects of social comparative feedback on task interest to shed more light on the modes of action expected to be behind the popular BFLP and BIRG effects.
Hence, the research question is as follows: How do contrast and assimilation effects of social comparative performance feedback directly influence task interest in an academic learning task?
Based on the empirical results and theoretical considerations presented above, we expect both contrast and assimilation effects to influence interest in an academic learning task. That is, we expect a higher social position within a reference group (compared to a lower social position), as well as a higher peer performance (compared to lower peer performance) to be associated with higher interest in an academic learning task.
In order to achieve a clear differentiation between the quality of feedback and actual performance, and therefore to increase the internal validity, feedback related to both social position and peer performance was experimentally manipulated in this study. That means the performance feedback every participant received was independent of actual performance. To directly investigate potential contrast effects the percentile rank of performance feedback given was experimentally manipulated (high and low social position conditions). This was done to simulate the respective performance feedback high and low performing learners usually receive. Further, potential assimilation effects were investigated by experimentally manipulating the percentage of points the reference group scored (high and low peer performance conditions).
Additionally, we want to replicate previously presented positive associations between self-concept, interest and achievement in an academic learning task. Finally, the potential influences of contrast and assimilation effects on task enjoyment, a further variable strongly associated with task-interest, are explored.
Participants were 129 first year pedagogy students at University of Potsdam studying to become elementary school teachers. Seven participants were excluded from further analyses due to missing data or because they expressed doubt about the authenticity of the performance feedback, leaving us with 122 participants. Each participant was randomly assigned to one of 2x2 experimental conditions: 31 participants were in the high social position / high peer performance (SP+/PP+) condition, 30 participants in the low social position / high peer performance (SP-/PP+) condition, 32 participants were in the high social position / low peer performance (SP+/PP-) condition, and 29 participants were in the low social position / low peer performance (SP-/PP-) condition. 110 participants were female, 11 were male, and 1 did not report a gender. The age of the participants ranged from 18 to 37 (M = 22.19, SD = 4.29). All participants attended a lecture on inclusive education and were recruited during the class.
The learning task we used in this study was presented to the participants as the “flag game.” Participants were told that the game consisted of two phases: a learning and a performance phase. During the learning phase, participants were presented with a map of the entire African continent and the country outlines for every country. For each of the 30 learning trials, the outlines of one country were highlighted, accompanied by the corresponding national flag. Each combination was presented for up to 10 seconds. Participants could shorten the presentation of each pair by pressing the space key. After completion of the learning phase, the performance phase was started. During the performance phase, the same 30 items were presented. This time, however, five different country outlines were highlighted with every flag, that is, four distractors in addition to the correct country. Participants were instructed to choose the correct country as quickly as possible by clicking on the corresponding country name. There was no time limit in the performance phase. After the first run, feedback was given depending on the respective feedback condition. Then, a second run comprising of another learning and performance phase with the same 30 items was started. After the second run, no feedback was given.
To avoid sequencing effects between conditions, trial sequences were pseudo-randomized separately for each phase. Distractor items were taken from the item pool, balanced and pseudo-randomized separately for each performance phase. Hence, every participant was presented with the exact same trial sequences and distractor items.
During instruction, participants were told that they would receive information about their performance compared to the last 50 students of the same degree course. However, feedback did not reflect their real performance, but was given based on the respective feedback condition. To test for contrast and assimilation effects on task interest, we manipulated both social position (SP; high vs. low) and peer performance (PP; high vs. low) in a 2x2 design. Participants were told their performance rating could range from 0 to 100%, based on both correctness and the speed of their responses. Response speed was included to make it harder for the participants to predict their own performance. Performance feedback was shown graphically, with 50 stick figures in a graph representing the scores of the reference group and a highlighted stick figure representing the participants’ own score (see Figure 1). Scores for the reference group were approximately normally distributed and ranged between 0 and 50% in the low peer performance conditions. In the high peer performance conditions, the reference group score was simply raised by 30% for every reference group member, resulting in values between 30% and 80%. Based on the respective social position condition, the participant’s individual score was placed either at the top or the bottom of the distribution. This resulted in individual scores of 79% (high social position / high peer performance), 49% (high social position / low peer performance and low social position / high peer performance), or 19% (low social position / low peer performance).
The self-concept scale consists of four items: two positive (e.g., “Usually I have no trouble with learning tasks”) and two negative (e.g., “Dealing with learning tasks isn’t one of my strengths”) statements about the participant’s self-rated ability to resolve learning tasks. Participants rated the items on a Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Negative items were reversed and a mean self-concept score was calculated. Cronbach’s alpha of the scale was α = .84.
The interest scale consists of eight items: four items concerning interest in the task itself (e.g., “I like learning tasks such as this one”) and four items concerning interest in the content of the task (e.g., “I’m interested in country flags”). The items were adapted versions of an interest scale from a German motivation questionnaire (Rheinberg, Vollmeyer, & Burns, 2001). Each item was rated on a Likert-type scale ranging from 1 (strongly disagree) to 7 (strongly agree). Pre-test interest had a Cronbach’s alpha of .87, while post-test interest had a Cronbach’s alpha of α = .90.
Task enjoyment was measured by means of a single item at the end of the study (“Did you enjoy the study?”). The item was rated on a Likert-type scale from 1 (definitely no) to 4 (definitely yes).
Pre-test performance was the proportion of correctly answered multiple-choice items in the first performance phase, while post-test performance was the percentage of correctly answered items in the second performance phase. Both performance phases consisted of the same 30 items.
Participants were tested in groups of four to six persons in our learning laboratory. They were welcomed and informed about the upcoming procedure. Written informed consent was obtained from every participant prior to testing. Each participant was placed in front of a separate computer. Computers were properly shielded to avoid participants seeing each other’s screens. Firstly, the self-concept questionnaire was administered. During the computer-based instruction, participants were informed about the upcoming learning task and how feedback about their performance would be provided. After completion of the instruction, a questionnaire concerning task interest was administered. When all participants finished filling out the questionnaire, the computer-based learning task was begun. Afterwards, the task interest questionnaire was administered a second time to investigate any potential changes. In the end, participants answered two questions concerning whether they were motivated to participate in similar studies and an open question that functioned as a manipulation check. The true nature of the study and the experimental manipulation was then explained to all participants.
To test the previously presented hypotheses that there are contrast and assimilation effects on domain-specific interest, we followed the suggestion by Everitt and Hothorn (2011, p. 232) and used linear mixed models with fixed and random effects. This was done because regular linear regression is unable to deal with data from repeated measures. Maximum likelihood (ML) estimation was used for all models. Based on the suggestions by Nakagawa and Schielzeth (2013), marginal and conditional R² were reported in order to determine variance explained by fixed factors and by both fixed and random factors, respectively. Model 1 contained the subject as a random intercept factor and pre- and post-test interest as Level 1 variables. Contrasts for the measurement time factor were coded 0 for the pre- and 1 for the post-test measurements. For Model 2, we added the Level 2 variable social position and the interaction with measurement time. Then, peer performance, as well as interactions with measurement time and social position were added for Model 3. Finally, we also added the standardized (M = 0, SD = 1) self-concept and interaction terms to the other predictors in order to investigate possible moderating effects in Model 4. The contrasts were effect-coded as -1 for low social position or peer performance and +1 for high social position or peer performance, respectively. To determine the model best fitting our data, likelihood ratio (LR) tests comparing each model to the previous one were calculated. This was done to test whether the subsequent addition of predictors significantly increased variance explained.
To investigate the possible effects of experimental conditions on learning task performance, corresponding models were also calculated with task performance as the dependant variable.
Because task enjoyment was only retrieved at a single point in time at the end of the study, generalized least squares (GLS) models with ML estimation and only one measurement per subject were calculated to investigate BFLP effects on task enjoyment. In order to determine explained variance in the GLS models, Likelihood Ratio R² was calculated as suggested by Magee (1990). This time, an intercept-only model with task enjoyment as the dependent variable was used as the base Model 1. Again, social position was included in Model 2, while peer performance and the interaction between social position and peer performance were added to Model 3 and standardized self-concept and corresponding interactions to Model 4. Once again, LR tests were calculated to compare the four models.
All calculations were carried out using the statistical software R with nlme and MuMin packages.
Table 1 shows intercorrelations and descriptive statistics for all observed variables. As can be seen, self-concept is positively associated with pre- and post-test interest. Associations between self-concept and pre- and post-test performance was also positive, but only significant in the latter case. Further, pre-test interest showed a significant positive association to pre- and post-test performance. Comparable associations could be found between post-test interest and pre- and post-test performance.
Hence, results support the hypothesized associations between interest, self-concept and performance.
Table 1
Descriptive statistics and intercorrelations for all observed variables
Note. ** p < .01, * p < .05; Interest Diff. = difference between interest pre- and post-test, Perf. Pre-Test = performance pre-test, Perf. Post-Test = performance post-test, Perf. Diff. = difference between performance pre- and post-test, Enjoy = enjoyment
Table 2
Descriptive statistics for all observed variables for each experimental condition
Note. SP = social position; PP = peer performance; Performance variables reflect percentage of correct answers
Table 2 shows the descriptive results separately for each experimental condition. One-factorial ANOVAs with the experimental condition as independent variable showed no significant differences between groups for self-concept (F3, 118 = 0.78, p = .51), pre-test interest (F3, 118 = 1.45, p = .23) or pre-test performance (F3, 118 = 0.53, p = .67). Hence, there were no significant pre-test differences between the groups in relevant study variables.
Table 3 shows the model fit and parameter estimates for all models predicting task interest. Model 2 including measurement time, social position and the interaction between social position and measurement time showed significantly better model fit compared to Model 1 only containing measurement time (LR = 9.79, p < .01). Inclusion of peer performance and interactions (Model 3) did increase the model fit, albeit not significantly (LR = 7.72, p = .10). Addition of the self-concept and interactions to all previous predictors (Model 4) further increased the model fit (LR = 38.29, p < .001).
As can be seen in Model 1, measurement time did not significantly predict task interest, showing that there were no differences in task interest between measurement times across all experimental conditions.
Model 2 contained the hypothesized interaction between social position and measurement time, reflecting a more positive development of task interest in the high social position conditions, compared to the low social position conditions.
Model 3 further contained the hypothesized interaction between measurement time and peer performance as a significant predictor, once again reflecting a more positive interest development in the high peer performance compared to the low peer performance conditions.
Model 4 additionally included self-concept as a significant predictor, showing a strong association between self-concept and task interest. The interaction between self-concept and measurement time, however, was small and did not reach significance. Self-concept also did not moderate interactions between social position and measurement time or between peer performance and measurement time.
Hence, even though task interest did not change from pre- to post-test measure across the whole sample, there were interactions between social position and measurement time as well as between peer position and measurement time, reflecting a more positive development of task interest in the high social position and high peer performance conditions compared to the low social position and low peer performance conditions, respectively. Additionally, self-concept was associated with pre-test task interest, but did not influence the effect of experimental conditions on the development of task interest. Therefore, our results support the hypothesized influence of contrast and assimilation effects on task interest.
Table 4 shows the model fit and parameter estimates for all models predicting task performance. Neither Model 2 containing social position and the interaction between social position and measurement time (LR = 1.29, p = .52), nor Model 3 additionally containing peer performance and corresponding interactions (LR = 2.91, p = .57), nor Model 4 further containing self-concept and corresponding interactions (LR = 10.13, p = .26) showed significantly better model fit then the model containing only repeated measurements.
Model 1 did contain a significant predictor for measurement time, showing that our participants performed significantly better in the second compared to the first run. These differences in task performance were influenced by neither social position nor peer performance.
Table 3
Fixed effects for mixed models predicting interest (N = 122)
Note. ** p < .01, * p < .05; MT = measurement time (0/1), SP = social position (-1/+1), PP = peer performance (-1/+1), SC = self-concept (centered), logLik = logarithmized Likelihood of the model, df = degrees of freedom, R²m = marginal R², R²c = conditional R²
Table 4
Fixed effects for mixed models predicting task performance (N = 122)
Note. ** p < .01, * p < .05; MT = measurement time (0/1), SP = social position (-1/+1), PP = peer performance (-1/+1), SC = self-concept (centered), logLik = logarithmized Likelihood of the model, df = degrees of freedom, R²m = marginal R², R²c = conditional R²
Table 5 shows the model fit and parameter estimates for all four models predicting task enjoyment. Model 2 containing the social position showed significantly better model fit then the intercept-only Model 1 (LR = 5.15, p < .05), while inclusion of peer performance and the interaction between social position and peer performance in Model 3 did not further increase the model fit (LR = 1.25, p = .54). Inclusion of self-concept (Model 4) also did not significantly increase model fit (LR = 2.63, p = .62).
Model 2 included social position as a significant predictor, while in Model 3 neither peer performance nor the interaction between social position and peer performance significantly predicted task enjoyment. In Model 4, self-concept was also not a significant predictor.
Hence, while the social position condition did influence task enjoyment in the hypothesized direction (i.e. feedback of a high social position led to higher task enjoyment compared to feedback of a low social position), there was no effect of the peer performance condition on task enjoyment.
Table 5
Fixed effects for mixed models predicting task enjoyment (N = 121)
Note. ** p < .01, * p < .05; SP = social position (-1/+1), PP = peer performance (-1/+1), SC = self-concept (centered), logLik = logarithmized Likelihood of the model, df = degrees of freedom, R²LR = Likelihood Ratio R²
In this study, we tested whether contrast and assimilation effects on task interest, which are thought to be the basis of popular BFLP and BIRG effects, could be replicated in an experimental paradigm. As expected, the feedback of both a high social position and high peer performance had a positive influence on the change in our participants’ interest over the course of this study (compared to low social position and peer performance, respectively). When considering the sample as a whole, there was only a minor, non-significant increase in task interest between pre- and post-test. However, when taking a closer look at group differences in interest development, a pattern matching the hypothesized additive effects of contrast and assimilation effects could be found. While feedback of high social position / high peer performance showed a clear increase, task interest remained virtually unchanged in both intermediate groups (i.e. feedback of high social position / low peer performance and feedback of low social position / high peer performance). Further, feedback of low social position / low peer performance even caused a decrease in task interest. Hence, this study provides additional experimental evidence for the hypothesis that contrast and assimilation effects of performance (as described in the IEM by Schwarz and Bless, 1992), and subsequent positive or negative performance feedback are likely to be at the core of BFLP and BIRG effects in task interest.
However, when it comes to task interest social comparative feedback is a zero-sum game. Not only are there winners, but also a rather consistent group whose interest is systematically undermined by continuing processes of negative social comparative evaluation, which could lead to further decline of academic achievement (Möller et al., 2011; Schiefele, 1992). This tends to be less of a problem in relatively homogeneous groups, where shortcomings in prior knowledge can more easily be overcome by increased effort. In rather heterogeneous groups, however, the influence of effort on academic evaluation is much lower, resulting in continuing negative social-comparative evaluation and therefore diminishing academic interest in slower learners. Hence, the focus on grades as the most popular form of formal feedback in institutionalized learning settings should be re-evaluated in order to facilitate effective learning for all students. Butler (1992), for example, could show that task interest for sixth graders differed depending on whether the instruction focused on task mastery or social comparison as well as on their self-rated performance. Subjects who rated their performance highly showed higher interest in the social comparison condition, while subjects with low self-rated performance showed higher interest in the mastery condition. The results from Harks, Rakoczy, Hattie, Besser, and Klieme (2014) point in a similar direction: Process-oriented feedback was perceived as more useful by ninth graders than grade-oriented feedback and also had a positive effect on the development of mathematics achievement and interest.
Feedback of a high social position did additionally lead to the hypothesized increase in task enjoyment compared to a low social position, suggesting a similar contrast effect of task performance on task enjoyment. Peer performance did not influence task enjoyment in the hypothesized direction, however, not supporting our hypothesis of an assimilation effect in task enjoyment. Hence, an effect of social comparative performance feedback on task enjoyment similar to the one on task interest could only be partially confirmed.
Further, we explored associations between self-concept, task interest and task performance and potential reference group effects on task interest. We were able to confirm the hypothesized positive associations between self-concept, interest and performance in an academic learning task. Hence, a positive self-concept usually coincided with higher interest as well as performance in our learning task.
The experimental paradigm of this study was chosen to increase internal validity and further help to shed light on the exact mechanisms behind the contrast and assimilation effects of academic achievement. The downside is that the learning task in this study was strictly linear and lacked several dimensions of classroom learning (e.g. cooperation, teacher-student interaction, etc.). Therefore, experimental studies with learning situations more closely resembling those of classroom learning are necessary in order to gain a better understanding of the influence of contrast and assimilation effects on interest in actual school settings. Especially the operationalization of the two different peer performance conditions focuses on criterial differences in test score and lacks important aspects of the BIRG effect (e.g. being chosen into a highly valued group; highly visible selection process) that have been suggested (Marsh, 1987; Marsh, Kong, & Hau 2000). Further, pre-/post-differences in interest between the two conditions with individual scores of 49 % (i.e. high social position / low peer performance and low social position / high peer performance) were relatively small, possibly hinting at effects from the criterial score on interest that could not be controlled within the design of this study. Additionally, self-concept was only measured once at the beginning of the study. The addition of a post-learning-task measurement of self-concept would have enabled a deeper investigation of potential interactions between self-concept and task interest.
Ames, C. (1992). Classrooms: Goals, structures, and student
motivation. Journal of Educational Psychology, 84(3),
261–271. https://doi.org/10.1037/0022-0663.84.3.261
Bless, H., & Schwarz, N. (2010). Mental Construal and the
Emergence of Assimilation and Contrast Effects. Advances in
Experimental Social Psychology, 42, 319–373.
https://doi.org/10.1016/S0065-2601(10)42006-7
Butler, R. (1992). What young people want to know when: Effects of
mastery and ability goals on interest in different kinds of social
comparisons. Journal of Personality and Social Psychology,
62(6), 934–943.
https://doi.org/10.1037/0022-3514.62.6.934
Chiu, M.-S. (2012). The internal/external frame of reference
model, big-fish-little-pond effect, and combined model for
mathematics and science. Journal of Educational Psychology,
104(1), 87–107. https://doi.org/10.1037/a0025734
Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation
and self-determination in human behavior. New York: Plenum.
Denissen, J. J. A., Zarrett, N. R., & Eccles, J. S. (2007). I
Like to Do It, I’m Able, and I Know I Am: Longitudinal Couplings
Between Domain-Specific Achievement, Self-Concept, and Interest. Child
Development, 78(2), 430–447.
https://doi.org/10.1111/j.1467-8624.2007.01007.x
Dickhäuser, O., & Galfe, E. (2004). Besser als ..., schlechter
als... Leistungsbezogene Vergleichsprozesse in der Grundschule.
Zeitschrift Für Entwicklungspsychologie Und Pädagogische
Psychologie , 36(1).
Dijkstra, P., Kuyper, H., van der Werf, G., Buunk, A. P., &
van der Zee, Y. G. (2008). Social Comparison in the Classroom: A
Review. Review of Educational Research, 78(4),
828–879. https://doi.org/10.3102/0034654308321210
Everitt, B., & Hothorn, T. (2011). An Introduction to
Applied Multivariate Analysis with R. New York, NY:
Springer New York. https://doi.org/10.1007/978-1-4419-9650-3
Fryer, L. K. (2015). Predicting self-concept, interest and
achievement for first-year students: The seeds of lifelong
learning. Learning and Individual Differences, 38,
107–114. https://doi.org/10.1016/j.lindif.2015.01.007
Goetz, T., Preckel, F., Zeidner, M., & Schleyer, E. (2008).
Big fish in big ponds: A multilevel analysis of test anxiety and
achievement in special gifted classes. Anxiety, Stress &
Coping, 21(2), 185–198.
https://doi.org/10.1080/10615800701628827
Guay, F., Marsh, H. W., & Boivin, M. (2003). Academic
self-concept and academic achievement: Developmental perspectives
on their causal ordering. Journal of Educational Psychology,
95(1), 124–136.
https://doi.org/10.1037/0022-0663.95.1.124
Harks, B., Rakoczy, K., Hattie, J., Besser, M., & Klieme, E.
(2014). The effects of feedback on achievement, interest and
self-evaluation: the role of feedback’s perceived usefulness. Educational
Psychology, 34(3), 269–290.
https://doi.org/10.1080/01443410.2013.785384
Huang, C. (2011). Self-concept and academic achievement: A
meta-analysis of longitudinal relations. Journal of School
Psychology, 49 (5), 505–528.
https://doi.org/10.1016/j.jsp.2011.07.001
Huguet, P., Dumas, F., Marsh, H., Régner, I., Wheeler, L., Suls,
J., … Nezlek, J. (2009). Clarifying the role of social comparison
in the big-fish–little-pond effect (BFLPE): An integrative study.
Journal of Personality and Social Psychology, 97(1),
156–170. https://doi.org/10.1037/a0015558
Jansen, M., Scherer, R., & Schroeders, U. (2015). Students’
self-concept and self-efficacy in the sciences: Differential
relations to antecedents and educational outcomes. Contemporary
Educational Psychology, 41, 13–24.
https://doi.org/10.1016/j.cedpsych.2014.11.002
Jonkmann, K., Becker, M., Marsh, H. W., Lüdtke, O., &
Trautwein, U. (2012). Personality traits moderate the
Big-Fish–Little-Pond Effect of academic self-concept. Learning
and Individual Differences, 22(6), 736–746.
https://doi.org/10.1016/j.lindif.2012.07.020
Köller, O., Schnabel, K. U., & Baumert, J. (2000). Der Einfluß
der Leistungsstärke von Schulen auf das fachspezifische
Selbstkonzept der Begabung und das Interesse. Zeitschrift
Für Entwicklungspsychologie Und Pädagogische Psychologie
, 32(2), 70–80.
https://doi.org/10.1026//0049-8637.32.2.70
Magee, L. (1990). R 2 Measures Based on Wald and Likelihood Ratio
Joint Significance Tests. The American Statistician, 44(3),
250. https://doi.org/10.2307/2685352
Marsh, H. W. (1986). Verbal and Math Self-Concepts: An
Internal/External Frame of Reference Model. American
Educational Research Journal, 23(1), 129149.
https://doi.org/10.3102/00028312023001129
Marsh, H. W. (1987). The big-fish-little-pond effect on academic
self-concept. Journal of Educational Psychology, 79(3),
280–295. https://doi.org/10.1037/0022-0663.79.3.280
Marsh, H. W. (1991). Failure of High-Ability High Schools to
Deliver Academic Benefits Commensurate with Their Students’
Ability Levels. American Educational Research Journal, 28(2),
445. https://doi.org/10.2307/116294
Marsh, H. W., & Hau, K.-T. (2003). Big-Fish--Little-Pond
effect on academic self-concept: A cross-cultural (26-country)
test of the negative effects of academically selective schools. American
Psychologist, 58(5), 364–376.
https://doi.org/10.1037/0003-066X.58.5.364
Marsh, H. W., Kong, C.-K., & Hau, K.-T. (2000). Longitudinal
multilevel models of the big-fish-little-pond effect on academic
self-concept: Counterbalancing contrast and reflected-glory
effects in Hong Kong schools. Journal of Personality and
Social Psychology, 78(2), 337–349.
https://doi.org/10.1037//0022-3514.78.2.337
Marsh, H. W., & O’Mara, A. J. (2010). Long-Term Total Negative
Effects of School-Average Ability on Diverse Educational Outcomes:
Direct and Indirect Effects of the Big-Fish-Little-Pond Effect. Zeitschrift
für Pädagogische Psychologie, 24(1), 51–72.
https://doi.org/10.1024/1010-0652/a000004
Marsh, H. W., & Parker, J. W. (1984). Determinants of student
self-concept: Is it better to be a relatively large fish in a
small pond even if you don’t learn to swim as well? Journal
of Personality and Social Psychology, 47(1),
213–231. https://doi.org/10.1037/0022-3514.47.1.213
Marsh, H. W., & Shavelson, R. (1985). Self-Concept: Its
Multifaceted, Hierarchical Structure. Educational
Psychologist, 20(3), 107–123.
https://doi.org/10.1207/s15326985ep2003_1
Marsh, H. W., Trautwein, U., Lüdtke, O., & Köller, O. (2008).
Social comparison and big-fish-little-pond effects on self-concept
and other self-belief constructs: Role of generalized and specific
others. Journal of Educational Psychology, 100(3),
510–524. https://doi.org/10.1037/0022-0663.100.3.510
Marsh, H. W., Trautwein, U., Lüdtke, O., Köller, O., &
Baumert, J. (2005). Academic Self-Concept, Interest, Grades, and
Standardized Test Scores: Reciprocal Effects Models of Causal
Ordering. Child Development, 76(2), 397–416.
https://doi.org/10.1111/j.1467-8624.2005.00853.x
Möller, J., Retelsdorf, J., Köller, O., & Marsh, H. W. (2011).
The Reciprocal Internal/External Frame of Reference Model: An
Integration of Models of Relations Between Academic Achievement
and Self-Concept. American Educational Research Journal,
48(6), 1315–1346.
https://doi.org/10.3102/0002831211419649
Mulkey, L. M., Catsambis, S., Steelman, L. C., & Crain, R. L.
(2005). The long-term effects of ability grouping in mathematics:
A national investigation. Social Psychology of Education,
8(2), 137–177. https://doi.org/10.1007/s11218-005-4014
Nagengast, B., & Marsh, H. W. (2011). The negative effect of
school-average ability on science self-concept in the UK, the UK
countries and the world: the Big-Fish-Little-Pond-Effect for PISA
2006. Educational Psychology, 31(5), 629–656.
https://doi.org/10.1080/01443410.2011.58641
Nakagawa, S., & Schielzeth, H. (2013). A general and simple
method for obtaining R2 from generalized linear
mixed-effects models. Methods in Ecology and Evolution,
4(2), 133–142.
https://doi.org/10.1111/j.2041-210x.2012.00261.x
Pohlmann, B., & Möller, J. (2006). Vergleichseffekte auf
kognitive, affektive und motivationale Variablen.
Zeitschrift für Entwicklungspsychologie und Pädagogische
Psychologie , 38(2), 79–87.
https://doi.org/10.1026/0049-8637.38.2.79
Preckel, F., & Brüll, M. (2010). The benefit of being a big
fish in a big pond: Contrast and assimilation effects on academic
self-concept. Learning and Individual Differences, 20(5),
522–531. https://doi.org/10.1016/j.lindif.2009.12.007
Rheinberg, F., Vollmeyer, R., & Burns, B. D. (2001). FAM: Ein
Fragebogen zur Erfassung aktuller Motivation in Lern- und
Leistungssituationen. Diagnostica, 47(2),
57–66. https://doi.org/10.1026//0012-1924.47.2.57
Schiefele, U. (1992). Topic Interest and Leveles of Text
Comprehension. In K. A. Renninger, S. Hidi, & A. Krapp (Eds.),
The Role of Interest in Learning and Development (pp.
151–182). Hillsdale, N.J: L. Erlbaum Associates.
Schiefele, U., Krapp, A., & Winteler, A. (1992). Interest as a
Predictor of Academic Achievement: A Meta-Analysis of Research. In
K. A. Renninger, S. Hidi, & A. Krapp (Eds.), The Role of
Interest in Learning and Development (pp. 183–212).
Hillsdale, N.J: L. Erlbaum Associates.
Schurtz, I. M., Pfost, M., Nagengast, B., & Artelt, C. (2014).
Impact of social and dimensional comparisons on student’s
mathematical and English subject-interest at the beginning of
secondary school. Learning and Instruction, 34,
32–41. https://doi.org/10.1016/j.learninstruc.2014.08.001
Schwarz, N., & Bless, H. (1992). Constructing reality and its
alternatives: Assimilation and contrast effects in social
judgment. In L. L. Martin & A. Tesser (Eds.), The
construction of social judgments (pp. 217–245). Hillsdale,
N.J: Lawrence Erlbaum Associates.
Seaton, M., Marsh, H. W., & Craven, R. G. (2009). Earning its
place as a pan-human theory: Universality of the
big-fish-little-pond effect across 41 culturally and economically
diverse countries. Journal of Educational Psychology, 101(2),
403–419. https://doi.org/10.1037/a001
Shavelson, R. J., Hubner, J. J., & Stanton, G. C. (1976).
Self-Concept: Validation of Construct Interpretations. Review
of Educational Research, 46(3), 407.
https://doi.org/10.2307/1170010
Springer, L., Stanne, M. E., & Donovan, S. S. (1999). Effects
of Small-Group Learning on Undergraduates in Science, Mathematics,
Engineering, and Technology: A Meta-Analysis. Review of
Educational Research, 69(1), 21–51.
https://doi.org/10.3102/00346543069001021
Strang, L., Smith, M. D., & Rogers, C. M. (1978). Social
comparison, multiple reference groups, and the self-concepts of
academically handicapped children before and after mainstreaming.
Journal of Educational Psychology, 70(4),
487–497.
Trautwein, U., Lüdtke, O., Marsh, H. W., Köller, O., &
Baumert, J. (2006). Tracking, grading, and student motivation:
Using group composition and status to predict self-concept and
interest in ninth-grade mathematics. Journal of Educational
Psychology, 98(4), 788–806.
https://doi.org/10.1037/0022-0663.98.4.788
Trautwein, U., Lüdtke, O., Marsh, H. W., & Nagy, G. (2009).
Within-school social comparison: How students perceive the
standing of their class predicts academic self-concept. Journal
of Educational Psychology, 101(4), 853–866.
https://doi.org/10.1037/a0016306
Valentine, J. C., DuBois, D. L., & Cooper, H. (2004). The
Relation Between Self-Beliefs and Academic Achievement: A
Meta-Analytic Review. Educational Psychologist, 39(2),
111–133. https://doi.org/10.1207/s15326985ep3902_3
Wang, Z. (2015). Examining big-fish-little-pond-effects across 49
countries: a multilevel latent variable modelling approach. Educational
Psychology, 35(2), 228–251.
https://doi.org/10.1080/01443410.2013.827155
Wilbert, J., Grosche, M., & Gerdes, H. (2010). Effects of
Evaluative Feedback on Rate of Learning and Task Motivation: An
Analogue Experiment. Learning and Disabilities: A
Contemporary Journal, 8(2), 43–52.
Wouters, S., De Fraine, B., Colpin, H., Van Damme, J., &
Verschueren, K. (2012). The effect of track changes on the
development of academic self-concept in high school: A dynamic
test of the big-fish–little-pond effect. Journal of
Educational Psychology, 104(3), 793–805.
https://doi.org/10.1037/a0027732
Zeidner, M., & Schleyer, E. J. (1999). The
Big-Fish–Little-Pond Effect for Academic Self-Concept, Test
Anxiety, and School Grades in Gifted Children. Contemporary
Educational Psychology, 24(4), 305–329.
https://doi.org/10.1006/ceps.1998.0985