Self-esteem, typically measured by the Rosenberg Self-Esteem Scale (RSE), is one of the most widely studied constructs in psychology. Nevertheless, there is broad agreement that a simple unidimensional factor model, consistent with the original design and typical application in applied research, does not provide an adequate explanation of RSE responses. However, there is no clear agreement about what alternative model is most appropriate-or even a clear rationale for how to test competing interpretations. Three alternative interpretations exist: (a) 2 substantively important trait factors (positive and negative self-esteem), (b) 1 trait factor and ephemeral method artifacts associated with positively or negatively worded items, or (c) 1 trait factor and stable response-style method factors associated with item wording. We have posited 8 alternative models and structural equation model tests based on longitudinal data (4 waves of data across 8 years with a large, representative sample of adolescents). Longitudinal models provide no support for the unidimensional model, undermine support for the 2-factor model, and clearly refute claims that wording effects are ephemeral, but they provide good support for models positing 1 substantive (self-esteem) factor and response-style method factors that are stable over time. This longitudinal methodological approach has not only resolved these long-standing issues in self-esteem research but also has broad applicability to most psychological assessments based on self-reports with a mix of positively and negatively worded items.

Many classroom climate studies suffer from two critical problems. They: (a) treat climate as a studentlevel (L1) variable in single-level analyses instead of a classroom-level (L2) construct in multilevel analyses; and (b) rely on manifest-variable models rather than latent-variable models that control measurement error at L1 and L2, and sampling error in the aggregation of L1 ratings to form L2 constructs. Based on an analysis of 2541 students in Grades 5 or 6 from 89 classrooms, we demonstrate doubly-latent multilevel structural equation models that overcome both these problems. The results show that L2 classroom climate (a higher-order factor representing classroom mastery goal orientation, challenge, and teacher caring) had positive effects on self-efficacy and achievement. We conclude with a discussion of related issues (e.g. the meaning of L2 constructs versus L1 residuals, the dimensionality of climate constructs at L2) and guidelines for future research.

Expectancy-value theory (EVT) is a dominant theory of human motivation. Historically, the Expectancy × Value interaction, in which motivation is high only if both expectancy and value are high, was central to EVT. However, the Expectancy × Value interaction mysteriously disappeared from published research more than 25 years ago. Using large representative samples of 15-year-olds (N = 398,750) from 57 diverse countries, we attempted to solve this mystery by testing Expectancy × Value interactions using latent-variable models with interactions. Expectancy (science self-concept), value (enjoyment of science), and the Expectancy × Value interaction all had statistically significant positive effects on both engagement in science activities and intentions of pursuing scientific careers; these results were similar for the total sample and for nearly all of the 57 countries considered separately. This study, apparently the strongest cross-national test of EVT ever undertaken, supports the generalizability of EVT predictions--including the "lost" Expectancy × Value interaction.

No abstract