Revisiting the Globalisation-Welfare Nexus: the Dependent and Independent Variable in Quantitative Comparative Research

Quantitative comparative research, particularly in the form of pooled time series cross-sectional regression, has been scrutinized ever since it first contributed to the comparative analysis of social policies. Thereby, model specifications and indicator operationalizations have largely been influenced by in-depth studies representing small samples of European cases. As data quality and availability for Asian countries is improving, there is an immediate temptation to ease the 'small-N problem' of quantitative comparative research by expanding country samples. Equally, there is a temptation for researchers to revert to well-established dependent and independent variables in the Western literature to explore Asian countries. This paper critically assesses the limitations of both approaches. It will do so by tracing the qualitative origins of the most important theoretical traditions accounting for the development and change of national social policies in Western context. Utilizing regression diagnostics, we highlight changes in the explanatory power of quantitative models and identify multivariate outliers for country samples with increasing diversity. Thus, we show traditional Western theories and indicators of social policy development, on their own, are hardly able to represent the development trajectories of Asian countries. We call, therefore, for a new and independent surge of in-depth case studies of Asian countries to inform a new round of more comprehensive, truly comparative, deductive research in the future.