Better Than Approval? Direction of the Country Questions as Predictors of Incumbent Races

By Larry Hugick and Stacy DiAngelo

The direction of the country question has become a staple of the national media polls. Every major polling organization asks a question to gauge people’s perceptions of how things are going in the country, or if they believe it is headed in the right direction. In nonpresidential election years, only the presidential job approval question is asked and reported more often.

While the value of approval ratings as a tool in forecasting elections is easy to grasp, interpreting the results of direction questions presents a real challenge to professional pollsters, not to mention more casual consumers of national media polls. For instance, researchers and commentators often characterize direction of the country results as a barometer of an incumbent’s chances of winning reelection. Generally missing from this analysis, however, are precision and a demonstrable track record.

So what are we to make of direction of the country ratings as a useful source of information for forecasting incumbent races? Generally, higher ratings are better and lower ratings are worse for an incumbent’s reelection prospects, but while a job approval rating significantly above the 50 percent mark usually signals reelection and one significantly below 50 percent usually signals defeat, there is no specific threshold for direction ratings that is commonly associated with electoral success or failure.

Indeed, some people argue that direction ratings do not provide information useful in assessing an incumbent’s electoral prospects. Response can be influenced by events that may not be in a political officeholder’s power to affect, such as a natural or manmade disaster. Some concerns that translate to negative views of the direction of the country, like worries about declining moral values, may be more strongly associated with an incumbent’s most ardent supporters than those inclined to vote against him or her.

Here we attempt to shed some light on the direction of the country question and the information it provides to poll consumers about electoral prospects for incumbent officeholders. To do this, we compared preelection satisfaction ratings and presidential approval ratings from polls taken in the third quarter prior to presidential elections and off-year congressional elections. Third-quarter data were used because they were gathered close to the election, but not so close that they would be available too late to be of much use. We used data from Pew Research Center and Newsweek polls where available, and Gallup poll data in other cases. If Gallup data had to be used for any particular year, they were used for both measures because of concern about house effects. We were able to find data for the last six presidential elections, as shown in Table 1.

These elections included four presidential races and two nonincumbent races. The incumbent party won four races and lost two. The average approval rating for the incumbent president leading into the election was 50 percent; the average percentage satisfied with the country was 40 percent.