In the winter of 1998 and early spring of 1999, Democratic and Republican identifiers in the electorate were sharply divided on whether President Clinton should be impeached and removed from office. By more than two to one, Republicans favored removal, while Democrats, by an even larger margin, were sharply opposed (Jacobson 2000). Given the intensely partisan context in which attitudes on this issue were surveyed, such results are not surprising.
Looking beyond the immediate issue of impeachment, we have found evidence of a more general tendency for partisanship to color assessments of the president. There is always a gap in assessments of presidential performance by Democrats and Republicans in the mass public, as those who identify with the president's party tend to rate his job performance more favorably than those who identify with the opposition. But the effect of partisanship has strengthened over time: for most years in the 1980s and 1990s, the partisan gap in evaluations of the president's job performance was appreciably wider than in previous decades (Fleisher and Bond 2001). Although finding partisan differences in how citizens evaluate the president is not new, the growing impact of party on such evaluations has gone largely unnoticed in the literature. Furthermore, much of the literature analyzes aggregate trends. Although studies of aggregate trends are appropriate and useful, they do not reveal the individual-level processes that lead to the aggregate results. In this article, we present additional evidence at the aggregate and individual levels highlighting the growing impact of partisanship on how the public evaluates the president.
Presidential Performance Evaluations in the Aggregate and at the Individual Level
Most analyses of the public's assessment of presidential job performance are based on the over-time study of presidential approval ratings collected monthly, quarterly, or annually. These aggregate time series studies indicate that the condition of the economy systematically affects the president's job approval ratings.(1) While these studies have contributed much to our understanding of the conditions that influence the public's evaluation of the president, they cannot analyze changes in the effect of partisanship over time.(2) Studies analyzing the relationship between national economic conditions and presidential approval at the aggregate level tell us little about how this connection is made or the factors that increase or decrease the effect of the economic news on presidential performance. Brody (1991) noted that aggregate studies assume that individuals (1) receive evidence about the performance of the economy, (2) judge this evidence against some benchmark, and (3) blame or credit the president for the condition of the economy. This model, as Brody observed, is not without problems. Information about the state of the world, including the condition of the economy, must be made available to voters in a way that allows it to be easily received and processed. Few voters are economists with access to economic indicators that they regularly monitor. Instead, the typical citizen is dependent on others to make information available. And even if information is readily available, citizens are not equally receptive to new information. The condition of the economy can only influence the opinions of those who are sufficiently attentive to receive the information. Furthermore, if the effects of partisanship have changed, then the effects of the economy on individuals' assessments of the president may have changed as well.
Information available to citizens consists of more than objective indicators about the state of the world. The news media and political elites filter the flood of information produced and offer guidance about how to interpret it (Ansolabehere, Behr, and Iyengar 1993; Iyengar 1991; Iyengar and Kinder 1987; Mutz 1993). By deciding what gets reported and what is ignored, the media determine what information is on the agenda. …