Synthetic Panels for Methods: From Data Engineering to Estimation Strategy

This chapter introduces three fully synthetic case studies that mirror administrative panels while adhering to the Five Safes. It explains how each file is generated, documented, and reshaped, and walks readers through the construction of essential derived variables, most notably a Wage Gap Indicator and interaction terms (e.g., Sector × Promotion). This chapter positions each case within this book’s modelling arc: fixed-effects linear and conditional-logit analyses for wage inequality; random-effects models for education–mortality gradients; and population-averaged generalised estimation equations for intergenerational mobility. By linking data engineering to estimation strategy, it clarifies why time-invariant predictors vanish under fixed effects yet reappear in RE/GEE, preparing readers for Chapters 5 – 7 .