Her dissertation is titled “Skill Bundles and Labour Market Outcomes: Identifying Different Types of Skills in Curriculum Texts by Applying Natural Language Processing”.

In her thesis, she shows how occupational skill bundles (i.e., types of skills and their combinations) can be measured by applying natural language processing (NLP) to curriculum texts. She also investigates the relationship between different types of occupational skill bundles and the labor market outcomes of graduates of the respective occupations. In a first chapter, she focuses on IT training and argues that IT skills (e.g., CNC) are not taught “in isolation” but rather in combination with other skills (e.g., reading manufacturing plans). Thus, efficient IT training consists of skills packages that not only include IT skills but also other complementary technical and nontechnical skills. Using an unsupervised topic modelling algorithm, she identifies four different skills packages with IT in Swiss occupational training curricula. Subsequently, she conducts empirical analyses showing that the skills packages are associated with positive wage returns and employment probabilities. However, the positive wage returns and employment probabilities vary across the different skills packages with IT (get Working Paper).

In a second chapter, she investigates noncognitive skills in occupational training curricula and their labor market returns. She examines how self-competence—i.e., the ability to act responsibly on one’s own—is associated with labor market returns (wages). First, she shows how to measure self-competence in occupational training curricula by NLP methods; second, based on these measures, she conducts empirical labor market analyses. She finds that self-competence is related to wages, but not in a linear way. Individuals trained in occupations with a medium level of self-competence receive the highest wage returns. Furthermore, she finds that these results also depend on the cognitive requirement level of the occupation: Individuals trained in an occupation with higher cognitive requirements also receive higher wages with a very high level of self-competence. Thus, cognitive requirement levels and self-competence seem to be complementary (get Working Paper).

In a third chapter, she studies whether and how specialized skill bundles (i.e., bundles with (few) very dominant single skills) in combination with social skills affect labor market outcomes. She particularly looks at episodes of rapid progress in information technologies (IT) and how that affects returns to specialized occupations with or without social skills. She develops a theoretical model and empirically analyzes its implications. The model shows how IT progress can, but does not have to, lead to increasing returns to specialization and social skills. To build measures for specialized occupations and to identify the extent of social skills in a curriculum, she again applies NLP methods to Swiss occupational training curricula. Her empirical analyses show that specialized workers receive higher wage returns during episodes with larger IT progress. Moreover, she finds that workers with high social skills receive increasing wage returns with IT progress only if they are also specialized.