Jan Paul Heisig

Who Does More Housework: Rich or Poor? A Comparison of 33 Countries

This article studies the relationship between household income and housework time across 33 countries. In most countries, low-income individuals do more housework than their high-income counterparts; the differences are even greater for women’s domestic work time. The analysis shows that the difference between rich and poor women’s housework time falls with economic development and rises with overall economic inequality. I use a cross-national reinterpretation of arguments from the historical time-use literature to show that this is attributable to the association between economic development and the diffusion of household technologies and to the association between economic inequality and the prevalence of service consumption among high-income households. Results for a direct measure of technology diffusion provide striking evidence for the first interpretation. The findings question the widespread notion that domestic technologies have had little or no impact on women’s housework time. On a general level, I find that gender inequalities are fundamentally conditioned by economic inequalities. A full understanding of the division of housework requires social scientists to go beyond couple-level dynamics and situate households and individuals within the broader social and economic structure.

Secondary Education Systems and the General Skills of Less- and Intermediate-educated Adults A Comparison of 18 Countries

We investigate the impact of external differentiation and vocational orientation of (lower and upper) secondary education on country variation in the mean numeracy skills of, and skills gaps between, adults with low and intermediate formal qualifications. We use data on 30- to 44-year-olds in 18 countries from the 2011–12 round of the Program for the International Assessment of Adult Competencies. We find that higher levels of external differentiation (tracking) amplify skills gaps between less- and intermediate-educated adults. This is mainly due to lower mean skills achievement of less-educated adults. By contrast, greater emphasis on vocational skills in upper-secondary education is positively related to numeracy skills for both less- and intermediate-educated adults. Gains are larger for the less educated, so the gap in numeracy skills tends to fall with the degree of vocational orientation. We discuss implications of our findings for research on educational and labor market inequalities.

The Costs of Simplicity: Why Multilevel Models May Benefit from Accounting for Cross-Cluster Differences in the Effects of Controls

Context effects, where a characteristic of a higher-level unit or “cluster” (e.g., a country) affects outcomes and relationships at a lower level (e.g., that of the individual), are a primary object of sociological inquiry. During recent decades, sociologists have increasingly investigated context effects using quantitative methods. We show that quantitative multilevel studies in leading sociology journals nearly always assume that the effects of lower-level control variables do not vary across clusters. This “invariant coefficients assumption” is often implausible. Based on analytical reasoning and Monte Carlo evidence, we argue that this assumption can have severe consequences for the estimation of context effects. Comparing mixed effects (random intercept and slope) models, cluster-robust pooled OLS, and two-step approaches, we show that models neglecting cluster differences in the effects of control variables can be dramatically less efficient than models that allow for such heterogeneity. That is, they tend to miss the true strength of the relationship of interest by larger amounts. Efficiency losses are largest when there is pronounced cross-cluster heterogeneity in the coefficients of control variables, when there are marked compositional differences among clusters, and when the number of clusters is small. Thus, the costs of assuming invariant control slopes should be highest in country comparisons. We end with concrete recommendations for applied researchers.