The gender wage gap is accounted for to a substantial degree by the sex composition of occupations. The present thesis examines the mechanisms that produce this pattern. In particular, the theory of devaluation, currently the most widely accepted sociological explanation, is tested. The empirical findings, reported in three self-contained essays, question this line of explanation. All results are based on Swedish data: the Level of Living surveys (LNU; essays I and II) and administrative labour market registers (essay III). In Essay I the association between occupational prestige and occupational sex composition is examined. The association is non-linear, with gender mixed occupations having the highest prestige. Further, care work does not have lower prestige than other kinds of work. These results are inconsistent with expectations derived from devaluation theory. The analysis also shows that the wage returns to occupational prestige are lower for women than for men. Essay II examines why women receive relatively low returns to prestige. Family related factors are shown to be crucial. The gender difference in pay-off to prestige is thus marked among married/cohabiting employees with children but insignificant among singles as well as among childless married/cohabiting women and men. The gender wage gap in high-prestige occupations is largely due to differences between women and men in work characteristics difficult to reconcile with family duties. In essay III the functional form of the relation between wages and occupational sex composition is investigated. In the cross-section gender mixed occupations have the highest wages. Panel data tend to confirm this pattern: mobility from strongly male or female dominated occupations to more gender mixed occupations is associated with relatively high rates of wage growth. Further, there is a wage premium for care work but a wage penalty for other service work. These findings do not support devaluation theory. / <p>At the time of the doctoral defense, the following papers were unpublished and had a status as follows: Paper 2: Manuscript. 3: Accepted.</p>

Cognitive theories of depression, including Beck’s Cognitive Theory, suggest that depressed individuals hold negative schemas about themselves and their environment. These negative schemas may influence the extent to which depressed individuals process positivity. Reward Devaluation Theory posits that depressed individuals avoid and devalue positivity. This suggests that depressed individuals may be less likely to hold positive schemas, or may be more likely to associate positivity with negativity. Previous meta-analytic reviews suggest that this is potentially the case, but have not assessed for self-referential stimuli. Self-referential encoding and recall tasks assess for self-schemas and may give further insight into how depressed individuals process self-referential positivity. The aim of this systematic review and meta-analysis was to examine the extent to which depressed individuals differ in processing self-referential positivity and negativity. Results indicate that depressed individuals recalled fewer positive words than negative words, with severely depressed individuals also endorsing fewer positive words than negative words, in line with Reward Devaluation Theory. In addition, depressed individuals endorsed fewer positive words and more negative words as self-referential than other-referential. In comparison to nondepressed individuals, depressed individuals demonstrated endorsed and recalled fewer positive words and more negative words. These findings suggest that treatments targeting both reduced positive biases and increased negative biases may be most beneficial for depressed individuals, particularly those exhibiting more severe symptoms of depression