Investigating Global Self-Esteem by Integrating Theory and Methods

Abstract

One major challenge in psychological research is the integration of theory and methods. A successful integration (i.e., the harmonious consideration of theoretical ideas and methodological implementations) is crucial for drawing conclusions from empirical studies back to theoretical phenomena and has the potential to generate synergies for theoretical and methodological progress in science (Greenwald, 2012; Marsh & Hau, 2007). However, long-cherished assumptions in well-established research fields, accompanied by growing analytical complexity, have often limited the potential for substantive-methodological integrations. Therefore, the present dissertation was aimed at integrating theory and methods for one of the most well-studied constructs in psychology, namely, the global self, oftentimes represented as the construct global self-esteem. Global self-esteem describes individuals’ overall subjective feelings of worth and has attracted the interest of many researchers due to its rele-vance in the context of mental health (e.g., James, 1890/1963; Orth & Robins, 2014; Rosen-berg, 1989). Studying self-esteem is of particular concern during adolescence and young adulthood because, in this phase, individuals have to face many developmental and environ-mental challenges (Harter, 1998; Rosenberg, 1986). Global self-esteem has been described as a construct that is unidimensional (e.g., Rosenberg, 1989), trait-like (e.g., Orth & Robins, 2019), and socially manifested through parents and peers (e.g., Cooley, 1902; Harris & Orth, 2019; Leary & Baumeister, 2000). At the same time, however, there are deviations from and extensions of these assumptions such as conceptualizations of multidimensional, hierarchical global self-concept (Shavelson, Hubner, & Stanton, 1976), a consideration of state-like self-esteem (Leary & Baumeister, 2000), and the incorporation of other social contexts beyond parents and peers (e.g., students’ interactions with teachers in classrooms). Despite the theoretical relevance of these deviations and extensions, they have received only a little empirical attention in research on global self-esteem. One reason for this gap could be that sophisticated methodological implementations for these research foci have been missing. The present dissertation was thus aimed at integrating theory and methods in research on self-esteem. Thereby, this dissertation pursued two overarching objectives. The first objective was to improve the understanding of self-esteem in adolescence and young adulthood. For this purpose, this dissertation (a) addressed different conceptualizations of global self-concept as the apex of a multidimensional hierarchy and brought them together with global self-esteem, (b) investigated state and trait self-esteem and the consequences for their rela-tions with depressive symptoms, and (c) examined reciprocal relations between self-esteem and student-teacher relationships and examined these relations over time. The second objec-tive was to improve the understanding of different methodological implementations (namely, the operationalization of higher order constructs, states and traits, and reciprocal relations) and their unique consequences for the aforementioned research questions and beyond. The two overarching objectives of this dissertation were addressed in three empirical studies. In the first study (Rethinking the Elusive Construct of Global Self-Concept: A Latent Composite Score as the Apex of the Shavelson Model), different conceptualizations of the global self (i.e., global self-concept and global self-esteem) were examined. As the focus of the study, two different conceptualizations of global self-concept as the apex of the multidi-mensional self-concept hierarchy were compared by applying different analytical models to represent higher order constructs. Using three independent large-scale studies (N1 = 8,068; N2 = 3,876; N3 = 2,095) of adolescents and young adults, we (a) applied second-order factor models, which assume that global self-concept affects lower order domain-specific self-concepts, and (b) compared them with a model-based latent composite scores, which reflect processes by which global self-concept is formed on the basis of domain-specific self-concepts. Our results indicated advantages of the latent composite scores as indicated by higher variances and a more plausible pattern of stabilities and correlations with external crite-ria, such as unidimensional global self-esteem. Therefore, we propose that global self-concept—the apex of the multidimensional hierarchy of self-concept—be modeled as a latent composite score. Over and above this, the study indicated that the conceptualization of mul-tidimensional hierarchical global self-concept was more aligned with unidimensional global self-esteem when nonacademic self-concepts were included in comparison with academic self-concepts. In the second study (How State and Trait Versions of Self-Esteem and Depressive Symptoms Affect Their Interplay: A Longitudinal Experimental Investigation), the stability of self-esteem was addressed by using a two-fold operationalization of states and traits (meas-urement and modeling approach). Using exploratory (N1 = 683) and preregistered confirmatory (N2 = 1,087) experimental longitudinal designs, university students were randomly as-signed to state and trait measures of self-esteem (and depressive symptoms), which were operationalized by using different time frames in the questionnaires (“In general…” vs. “During the last 2 weeks…”). The results indicated that, first, consistently across the two studies, the trait time frames revealed higher proportions of trait variance and lower proportions of state residual variances than the state time frames. Second, across the two studies, the cross-lagged relations between self-esteem and depressive symptoms depended on the time frames used to assess the constructs. Third, when controlling for stable trait differences, the cross-lagged results were least consistent when trait time frames were used, which highlighted the interdependency involved in measuring and modeling state and trait self-esteem. In the third study (Is Teacher Attachment Prospectively Related to Self-Esteem? A 10-Year Longitudinal Study of Mexican-Origin Youth), the reciprocal relation between student-teacher relationships and students’ self-esteem was investigated by using two different cross-lagged panel models. The study used data from N = 674 Mexican-origin students followed annually from age 11 to 21 and tested eight preregistered hypotheses about reciprocal rela-tions between students’ perceived teacher attachment (i.e., teacher support and teacher rejec-tion) and students’ global self-esteem. The results indicated (a) positive prospective reciprocal relations between teacher support and self-esteem, based on cross-lagged panel models (CLPMs; focus on overall between-person differences) as well as random-intercept cross-lagged panel models (RI-CLPMs; focus on differences in within-person deviations), and (b) negative prospective reciprocal relations between teacher rejection and self-esteem, based only on CLPMs but not on RI-CLPMs. Overall, the results suggested that transactional processes underlie reciprocal relations between teacher attachment and self-esteem, whereas the results were more consistent in the CLPMs than in the RI-CLPMs. From a theoretical perspective, this dissertation refines the understanding of (a) the relation between unidimensional global self-esteem and multidimensional, hierarchical global self-concept, (b) trait and state self-esteem, as well as (c) individual and environmental predictors and consequences of self-esteem. From a methodological perspective, across the three studies, this dissertation observed important empirical differences from different methodological implementations. Thereby, this dissertation points to the consequences of cross-sectional and longitudinal higher order factor models and emphasizes the importance of integrating theory, methods, and data.