Problem:
Suppose I gave you the following kinds of data. The data come from interviews conducted with employees before and after these employees voted in union representation elections. The data are from employees who voted in 100 different representation elections. The data include:
A. The demographic characteristics of the employees such as their race, sex and age;
B. Responses to questions asking the employees about their attitudes toward unions (such as whether they think unions in general benefit society, and whether they think most union leaders are corrupt);
C. Responses to questions asking the employees their attitudes toward their employer (such as whether they feel satisfied by their pay, and how they feel towards their supervisor);
D. The employees' voting intentions--for or against union representation. This was asked one month prior to the election; and
E. The employees' actual vote--for or against union representation. This was asked right after the representation election. How could you use these data to understand why unions have been doing so poorly recently in representation elections?