To measure educational mismatch, the research team analysed nationally-representative cross-sectional survey data from the 2014 School to Work Transition Survey (STWTS) for youth aged 15 to 29. The team examined whether there are differences between the subjective (self-reported survey) and empirical (computed based on average education in a given occupation) effects of working while studying, through which they also verified the robustness of their results. To estimate the causal effect of working while studying on educational mismatch, the researchers used a range of Extended Ordered Probit regression models, taking into account possible self-selection and sample-selection biases, and controls for youth characteristics.