Frame-of-Reference Effects on Academic Self-Concept: Addressing Unresolved Issues with New Designs

Abstract

Students’ self-perceived competence in academic domains, also referred to as academic self-concept (Marsh et al., 2016), is assumed to be a central motivational factor that affects academic effort, achievement, aspirations, and choices (e.g., Guo et al., 2015; Valentine et al., 2004). Consequently, the formation of a positive academic self-concept is regarded as an essential requirement for successful learning processes, and researching its antecedents is of high theoretical and practical relevance. Social comparison processes, in which students evaluate their academic achievement in relation to their classmates (e.g., Huguet et al., 2009; Marsh, Kuyper, et al., 2014), are assumed to be one focal determinant of academic self-concept. Empirical evidence for this assumption stems from the finding that, on average, equally able students have a higher academic self-concept in low-achieving schools and classes than in high-achieving ones. In the former scenario, students feel like big fish in little ponds; thus, this phenomenon has been labeled the big fish little pond effect (BFLPE). The BFLPE has been extensively investigated in the last three decades. For instance, numerous studies have demonstrated its robustness to individual and contextual level moderators, the fact that frame-of-reference effects also impact other educational outcomes, and its existence across cultures (for an overview, see Marsh & Seaton, 2015). Despite the massive number of studies investigating the BFLPE, an in-depth understanding of the frame-of-reference effect is still lacking with respect to mechanisms (e.g., To which reference groups do students tend to compare themselves?), implications (e.g., What does the BFLPE mean for the design of educational systems?), and interdisciplinary integration (e.g., How can the BFLPE theory be embedded in other social science disciplines that focus on social comparison processes?). In the present dissertation, I argue that one of the reasons for this lack of in-depth understanding is that previous research on the BFLPE has partly been characterized by homogeneity in terms of the research designs used (Dai et al., 2013; Dai & Rinn, 2008). More specifically, researchers have investigated the BFLPE using education-specific cluster sampling data in which either a random sample of students within schools or a random sample of intact classrooms was drawn. This cross-sectional data was then analyzed using multilevel models in which the school or classroom represented the higher level. The overarching aim of the present dissertation is to address unresolved issues in research on the BFLPE by extending the range of research designs used and consequently providing new insights. Specifically, in the present dissertation, I discuss four unresolved issues in research on the BFLPE: multiple class environments as frames of reference for academic self-concept formation, the association between grading on a curve and the BFLPE, the effects of tracking on academic self-concept, and neighborhood effects on academic self-concept. For each of these unresolved issues, I will describe design-based challenges of previous research on a conceptual level and clarify what is needed for new designs. Finally, in four empirical studies, I address the presented issues with innovative research designs. Alongside this overarching aim, this dissertation pursues two subordinate aims that were addressed with two empirical studies each. The first subordinate aim is to use extensive large-scale data (comprehensive educational monitoring data and interdisciplinary large-scale data) for an in-depth investigation of the mechanisms and interdisciplinary integration of the BFLPE. Specifically, Study 1 used Austrian educational monitoring data to investigate multiple class environments as pivotal frames of reference for academic self-concept formation in systems with course-by-course tracking. Study 2 used interdisciplinary large-scale data to explore the neighborhood as a frame of reference for academic self-concept formation. The second subordinate aim is to use natural experiments to investigate the mechanisms and implications of the BFLPE in greater detail. For this purpose, Study 3 used a school reform that abolished formal grades to investigate the association between grading on a curve and the BFLPE. Additionally, Study 4 used two school reforms in which students were detracked to test the BFLPE’s predictions concerning tracking. The first study (Which Class Matters? Juxtaposing Multiple Class Environments as Frames of Reference for Academic Self-Concept Formation) was based on the 2012 Austrian Educational Standard Assessment (BIFIE, 2016; Schreiner & Breit, 2012), a comprehensive survey of all Austrian eighth-grade students in the domain of mathematics that contains identifiers for the multiple educational environments students experience. This extensive dataset made it possible to investigate the pivotal frames of reference for academic self-concept formation in school systems with course-by-course tracking. Secondary school students were tracked according to ability in the core subjects (mathematics, German, English) but attended all other subjects in the same mixed-ability class. When regressing math self-concept on math achievement aggregates on all levels in which students were nested, the math class BFLPE was most negative. The regular class BFLPE was less negative, and the school BFLPE was the least negative. These results are in line with local dominance theory, which argues that more local comparative information matters the most for self-evaluations. The second study (Living in the Big Pond: How Socioeconomic Neighborhood Composition Predicts Students’ Academic Self-Concept) benefited from data from Starting Cohort 3 of the German National Educational Panel Study (NEPS; Blossfeld et al., 2011), an interdisciplinary large-scale assessment that contains information on students’ educational outcomes as well as neighborhood characteristics. This extensive dataset made it possible to investigate the neighborhood as a frame of reference for academic self-concept formation. Better neighborhood socioeconomic conditions did not or negatively affected students’ academic self-concept. Our results stand in supposedly contrast to neighborhood effects research in sociology, which has found that better neighborhood socioeconomic conditions positively impact a broad range of educational outcomes. The third study (Can Grades Move the Big Fish? The Consequences of Receiving Report Cards for Frame-of-Reference Effects on Academic Self-Concept) was based on data from the Swedish Evaluation Through Follow-Up Study (ETF; Härnqvist, 2000). The data were collected during a reform period in which Swedish municipalities were free to decide whether or not to abolish formal grading in elementary school, enabling us to compare non-graded and graded students regarding the BFLPE in a natural experiment. We found no differences between non-graded and graded students regarding the BFLPE. The results are in line with an evolutionary approach to social comparisons, in which social comparison processes present an innate human drive that exists independent of grade provision. The fourth study (The Dark Side of Detracking: Mixed-Ability Classrooms Hurt Low Achievers’ Math Motivation) uses data coming from the Austrian National Educational Standard Assessment from 2012 and 2017 (Schreiner et al., 2017; Schreiner & Breit, 2012) as well as the Additional Study Thuringia from the German National Educational Panel Study (NEPS; Blossfeld et al., 2011). These data measured students before and after detracking school reforms, enabling us to investigate such reforms by employing a cohort-control design, in which student cohorts before and after detracking are compared. The BFLPE predicts that detracking decreases low achievers’ motivation in terms of academic self-concept, as this group of students is exposed to high-achieving classmates in mixed-ability classrooms. In line with this prediction, we found that low-achieving students’ academic self-concept was negatively impacted by detracking, whereas this was not the case for high achievers. Our results stand in contrast to detracking proponents, who argue that abolishing ability grouping will increase low-achieving students’ motivation, while highlighting the practical implications of the BFLPE. Finally, at the end of the dissertation, all four studies’ findings are embedded in a broader research context. There is also a final assessment of how successfully the design-based challenges raised have been addressed. Additionally, strengths and limitations are presented, and implications for practice and future research are discussed.