Dynamic Bayesian Adjustment of Educational Gradients in Divorce Risks: Disentangling Causation and Misclassification

Parfait Munezero, Stockholm University
Gebrenegus Ghilagaber , Stockholm University

We address a problem in causal inference from retrospective surveys where the value of a covariate is measured at the date of the survey but is used to explain behaviour that has occurred long before the survey. This causes bias because the anticipatory covariate does not follow the temporal order of events. We propose a Bayesian dynamic modelling approach that allows effects of the anticipatory covariate to vary over time and, thereby, restore its value at the event of interest. The issues are illustrated with data on the effects of anticipatory educational level on divorce risks among Swedish men. The overall results show that failure to adjust for the anticipatory nature of education leads to underestimation of the relative risks of divorce across educational levels. The results build, in part, on previous analyses of the same data set but also reveal that the degree of underestimation varies over marriage durations.

See extended abstract

 Presented in Session 185. Statistical Demography