Joint work with Jan Skopek
Lifelong learning is a core policy ideal in modern societies, yet participation remains highly unequal. Those with higher qualifications and prior learning experience are far more likely to take part—while others, often with the greatest need, are left behind. This project explores the potential role of metacognitive monitoring—the ability to assess one’s own knowledge and recognise learning needs—in helping to explain this divide. Drawing on the Dunning–Kruger effect, we examine whether individuals with low cognitive competencies and inflated self-assessments are especially unlikely to engage in adult education. We conceptualise this as a micro-cognitive mechanism of the theory of cumulative (dis)advantage, helping to explain how small initial differences compound into persistent inequalities in learning over the life course. Our study focuses on Germany and considers both formal, work-related training and the growing role of informal and digital learning environments. By linking cognitive psychology with sociological theory, we aim to deepen the understanding of how metacognitive competencies influence lifelong learning trajectories. The findings have practical implications for policy, counselling, and curriculum design in adult education.
Joint work with Eileen Peters & Mortimer Schlieker
Access to non-formal continuing vocational training (CVT) is a central mechanism of social stratification, yet participation is stratified by income and shaped by both workplace and family contexts. Previous research has typically treated income as an individual resource, although the same earnings imply different positions depending on workplace wage structure and one’s contribution to household income. We ask: (1) How are employees’ relative income positions within the workplace and within the family associated with participation in employer-supported and in individually initiated CVT? (2) How do the two positions interact in shaping access? Drawing on relational inequality theory and household bargaining perspectives, we conceptualise workplaces and families as contexts of resource allocation in which relative income positions indicate employees’ standing. We expect workplace position to matter especially for employer-supported CVT, where employers provide funding and/or working time (H1); and family position for individually initiated CVT, which draws on personal time and financial resources (H2a), while also radiating into employer-supported CVT through perceived commitment and availability (H2b). We further examine whether a stronger position in one context compensates for a weaker position in the other (H3). We use linked NEPS-SC6-ADIAB survey and administrative data on cohabiting employees aged 23–60 in Germany from 2012 to 2019 (5,622 person-wave observations from 1,686 individuals). Workplace position captures earnings relative to the establishment wage level, while family position captures the individual’s share of household net income. We estimate pooled linear probability models of employer-supported, individually initiated, and overall job-related CVT using lagged income positions and individual- and workplace-level controls. Preliminary results challenge the expected correspondence between each income position and its hypothesised training type. Workplace income position shows no detectable association with employer-supported or overall CVT, whereas family income position is positively associated with both. For individually initiated CVT, which is comparatively rare, workplace position is negatively associated with participation, while family position shows no detectable association. We find no evidence of compensation between the two positions. The study reframes income-related training inequality by treating income as a relational position across workplace and family contexts rather than as an individual resource. It extends relational inequality theory to the family as a second inequality regime, models two distinct CVT types side by side, and leverages linked employer–employee data to model workplace and family positions within a single estimation framework. Preliminary results suggest that aggregating forms of CVT obscures heterogeneous associations.
Joint work with Johanna Binnewitt
Computational text analysis offers new opportunities for studying symbolic dimensions of social inequality in naturally occurring communication. Yet measuring occupational stereotypes in text raises a prior measurement problem: contextual language must first be interpreted as evaluative and heterogeneous interpretations aggregated into computational targets. Using 225 paragraphs from German parliamentary debates, we annotate occupational references along eight warmth- and competence-related semantic differentials and examine disagreement and consensus. Disagreement arises particularly between no attribution and evaluative labels, indicating uncertainty about when contextual information warrants a trait attribution. Consensus aggregation yields a more selective representation, with fewer activated dimensions and greater prominence of competence-related evaluations. Our findings show that interpretation and aggregation are constitutive parts of contextual stereotype measurement rather than neutral preprocessing steps. Text-based measures are therefore better understood as source-specific measures of discursive occupational valuation that can complement, but not substitute for, survey-based occupational stereotype measures.