Using Multiple Q-Q Plots to Measure Comparative Change from Pseudo-cohort Studies

Abstract Although International Large-scale Assessment Surveys (ILSA) produce a wealth of policy-relevant information, as cross-sectional surveys they don’t yield measures of change for specific cohorts. This severely limits not only their ability to describe cohort learning over time, but also their utility for supporting causal inferences regarding policy efficacy. In this article, we describe a new methodology that, by employing multiple Q-Q plots in the context of pseudo-cohort designs, enables the estimation of comparative change between populations Such comparisons can be made between subpopulations within a country, among different countries, or for specific subpopulations across countries. These Q-Q plot-based descriptions of comparative change complement the performance level statistics currently available. The methodology, termed the ‘Multiple Q-Q Plot’, inherits the advantages of Q-Q plots and generates graphics that are easy to interpret. The plots are intended to be used as an exploratory tool to identify differences and/or similarities of interest to education stakeholders and education researchers, to stimulate further investigations and, in some settings, to contribute to causal analysis. Using data from Trends in International Math and Science Study (TIMSS), the article presents a number of illustrative examples, concluding with a discussion of related methodological issues, as well as some policy implications.