3.2 Antecedents
To address reverse causality, we collected all antecedent variables by using a one-year lag (t − 1) and updated them for each year from 1997 through 2006 (cf. Hambrick 2007, p. 338). Most of our measures and all of our data sources have been used previously.