Recent studies have challenged the conclusion Getman, Goldberg, and Herman reached in their 1976 study that company campaigns have little if any effect on how workers vote in union certification elections. This study attempts to reconcile these conflicting results by re-analyzing data collected by Getman, Goldberg, and Herman for their study. The author estimates a probit model of voting that controls for several variables not considered in other studies. He finds that employer threats and actions taken against union supporters, some written communications, and captive-audience speeches all have statistically significant effects on voting. Election simulations based on the voting model also show that the effects of these employer tactics can be important in determining election outcomes.
Stored only when both papers are in the corpus — 80 citations exist upstream.