Exercise 7: Attitude on Medicare Vouchers and the Presidential Vote
Another reason for introducing a control variable into the analysis is to see if the relationship between the independent and dependent variable is stronger for some groups of people than it is for others. If we find that such is the case, we commonly refer to that as a conditional relationship. An example of a conditional relationship involves the association between attitudes toward Medicare and the presidential vote. Obama and Romney differed in their positions on Medicare, with Romney proposing to move to a voucher system for Medicare and Obama opposing such a change. If voters cast their ballots on the basis of this issue, at least in part, then we would expect voters who were more opposed to privatizing Medicare to be more likely to vote for Obama. To see if this is so, generate a two-variable table that examines the relationship between attitude on Medicare vouchers (H04) and presidential vote (A02). To simplify the table, use the recoded version of A02 that excludes the minor party.
Table 7A shows that overall there is a strong relationship between attitude on privatizing Medicare and the presidential vote. This does not mean that all groups of voters were equally influenced by this issue. We can hypothesize about the types of voters for whom this relationship would be stronger and the types for whom it would be weaker. One possible hypothesis is that the relationship will be stronger for those who are older because they are more likely to think that it is an issue of high importance, due to the fact that they currently rely on Medicare or expect to in the near future. Younger voters, in contrast, probably are less concerned about this issue, so they should be less likely to vote on the basis of that issue. You can test this hypothesis by introducing age as a control variable (R03) into the analysis. To keep your table simple and to ensure that there are enough respondents in each column of each subtable, recode R03 so that it has just three categories (younger, middle-aged, and older voters).
In this example, our original two-variable relationship becomes somewhat stronger among older voters and somewhat weaker among younger voters. Among all age groups — younger, middle-aged, and older voters — those who oppose a voucher program are more likely to vote for Obama than are those who favor it, but the extent of the difference is the greatest among older voters and the least among younger voters. Therefore, age is a conditional variable in this situation, as we hypothesized.