Who is responsible, the incumbent or the former president? Motivated reasoning in responsibility attributions.
This recent election highlights how important it is for political scientists to consider how individuals ascribe responsibility to politicians. These responsibility judgments have a profound effect on candidate evaluation and voting behavior (see Feldman 1982; Lau and Sears 1981; Lowry, Alt, and Ferree 1998; Peffley 1984; Sniderman and Brody 1977), and deserve greater attention in the literature. Since people often vote based on whether or not they see a politician responsible for national conditions, it is vital for political scientists to develop a deeper understanding of how citizens make responsibility attributions in various contexts. Importantly, a voter's responsibility attribution does not occur in a vacuum; this article argues that the decision of whether to hold a current officeholder responsible will depend on voters' perceptions of the issue, their partisan attachments, and the responsibility they assign to the former officeholder. Following a transition of power from one administration to the next, the political significance of responsibility attributions can be dramatic; whether voters ascribe responsibility for conditions to the current or former president can greatly affect the incumbent's job approval, the "political capital" held by the incumbent when implementing his agenda, and the willingness of voters to support or oppose his party in upcoming elections.
Research has shown that responsibility attributions are often driven by partisan motives (e.g., Atkeson, and Maestas 2012; Brown 2010; Maestas et al. 2008; Malhotra and Kuo 2010; Rudolph 2003b; Sirin and Villalobos 2011). This study builds upon that literature by examining how citizens determine responsibility across presidential administrations in light of the confusion caused by a presidential transition of power. This article extends the findings of these studies to show that the partisan biases that often accompany the attribution process are not limited to the assessment of current political leaders. Instead, consistent with theories of motivated reasoning, citizens credit and blame both current and former officeholders according to their partisan leanings in order to achieve harmony between their partisanship and their evaluations of issue conditions.
Following a review of the attribution and motivated reasoning literatures, I first establish the relevance of presidential transitions in the attribution process by showing that uncertainty exists regarding who is responsible for conditions early in a presidential administration. Unfortunately, existing data sources cannot speak to how citizens handle this ambiguity. To answer the question of how people assign responsibility across presidential administrations, I present original survey data that shows citizens often confront this uncertainty by assigning responsibility in accordance with their partisanship and issue evaluations. In essence, the data shows that it is quite common for average citizens to shift their perceived responsibility for negative conditions toward the president of the opposite party while viewing a co-partisan president responsible for positive conditions. The results presented here suggest that scholars need to understand how severely partisanship can bias citizens' issue evaluations and responsibility attributions. Furthermore, by broadening the scope of the analysis to include attributions for the economy and the Iraq War, this study joins Sirin and Villalobos (2011) as the only research, to my knowledge, that directly examines how individuals form responsibility attributions for a foreign affairs issue.
For most issues confronting the nation, citizens can reasonably assign responsibility to a variety of leaders, institutions, and factors. Despite this, previous attribution studies have lacked a comprehensive examination of the responsibility shared by attribution targets. Instead, the vast majority of attribution studies have looked at the economic attribution process by couching the response options in general terms like "the President," "Congress," or "businesspeople" (e.g., Rudolph 2003b's analysis of the 1998 American National Election Studies [ANF.S]) or through the use of hypothetical situations in experiments (e.g., Sirin and Villalobos 2011). While such measurements are useful for looking at the role of institutions or evidencing causal relationships, they do not fully account for the political context in which the attribution process resides and may underestimate the role of partisanship when party labels and actual names are not provided (as is the case with the 1998 ANES). In rare cases, names are given, such as "President Reagan" rather than "the President" (e.g., Peffley and Williams 1985), which improves upon this issue, but still lacks a complete inclusion of the political context because it does not acknowledge history's role in the attribution process. Specifically, only including the current president as a response option fails to account for the possibility that individuals will blame or credit former elected officials for current conditions. Since there is regular turnover in American governments, and since national conditions often change slowly, studying the attribution process in this way provides an incomplete picture at best and an inaccurate one at worst.
Unfortunately, little work has seriously approached how individuals assign responsibility for national conditions after a governmental transition. Aside from studies that suggested Republican losses in the 1982 midterms were minimized because of the electorate's failure to blame Ronald Reagan for the recession (Peffley 1984; Peffley and Williams 1985; Petrocik and Steeper 1986), the role of previous officeholders in the assignment of credit and blame is not handled in the literature. Clearly, responsibility assignment after a transition of power is an area that is underexplored and ripe for further elaboration. This article uses two issues to highlight the problems of the standard incumbent-only approach: the national economy and the Iraq War. With both issues, it would be difficult to ascribe responsibility without being able to consider President Bush alongside President Obama. Yet despite this, it is surprising that previous work does not include former officeholders in the response option set when asking about attributions.
This study assesses how the public holds the former officeholder responsible for conditions in relation to the incumbent and vice versa. This is an important addition to the literature because governmental transitions of power are quite common in the United States, occurring at regular intervals due to electoral defeats, presidential term limits, and when other unfortunate situations cause one president to take over the term of another. When a new president takes over, he is not given a clean slate. Using the economy as an example, things like budgetary issues, government spending, trade agreements, and tax policies all carry over and affect conditions well after a president leaves office. Furthermore, we can easily apply the logic of presidential transitions to other instances in which one officeholder takes over after another (e.g., governors, mayors, and partisan shifts or leadership changes in Congress).
Instead of transitions, the closest examples of research examining how partisan biases motivate responsibility attributions deal with other situations where responsibility may be ambiguous, such as across institutions (Rudolph 2003b), in competing policy domains (Sirin and Villalobos 2011), in cases of "divided federalism" (Atkeson and Maestas 2012; Brown 2010), or in disaster response (Atkeson and Maestas 2012; Maestas et al. 2008; Malhotra and Kuo 2010). The results of these studies suggest that individuals tend to assign credit and blame in a partisan manner, which is consistent with earlier attribution work that claimed attributions to be rationalizations of previously held beliefs (Iyengar 1989) or the result of partisanship and previous voting behavior (Sigelman and Knight 1985).
The political science literature surrounding presidential approval ratings may provide additional context for this study. Like attributions, partisan predispositions often color approval ratings (Edwards and Gallup 1990). Attributions of responsibility and presidential approval attitudes may be closely related as well; if approval varies with changes in economic indicators and perceptions (McAvoy 2006, 2008) and other issues like war and foreign policy depending on issue salience (Edwards, Mitchell, and Welch 1995; McAvoy 2006), citizens must be connecting the president's performance to these issue conditions and ascribing him responsibility.
It is therefore not entirely surprising to see the partisan patterns found in responsibility attributions similarly present when considering variation in presidential approval. As Lebo and Cassino (2007) note, "Self-identified Democrats and Republicans look upon the president from very different viewpoints. Naturally, each group begins with either a basic empathy or aversion towards a president depending upon the president's party" (720). Their analysis of 50 years of monthly presidential approval data disaggregated by party shows a striking partisan pattern to approval--partisans as a group only reward/ punish opposite-party presidents based on economic performance; for members of the president's party, presidential approval is rarely affected by unemployment and inflation.
Clarity of responsibility for issue conditions is another factor that we should consider. Fiorina notes that "if responsibility was problematic in American politics even when government was unified, the problem is compounded when government is divided" (1992, 110). The presence of divided government increases presidential approval ratings because presidents essentially lose less in credit claiming than they gain in the ability to avoid blame because they share responsibility with opposite-party majorities in Congress (Nicholson, Segura, and Woods 2002). This finding meshes nicely with other studies that show an increase in accountability and blame attribution toward the president during periods of unified government (Lowry, Alt, and Ferree 1998; Nicholson and Segura 1999).
This article applies the reasoning of these studies to the period following a transition from one president to another. Survey results from the ANES and a brief content analysis of economic articles in the New York Times will establish that responsibility uncertainty exists early in a president's term and declines over time. This data also suggests that partisan opinions vary because of this ambiguity, but the fact that attributions are not explicitly measured limit our ability to make inferences about responsibility; thus, an original survey is necessary to clarify how people assign responsibility following a transition.
Motivated Reasoning and Partisanship in the Attribution Process
Attribution theory posits that individuals ascribe different levels of responsibility to actors based on the outcome of the actors' actions and whether or not the individuals believe that the actors should have done something else (Heider 1958). However, many researchers have recognized that situational factors often affect attributions, resulting in bias (Jones and Harris 1967; Jones and Nisbett 1971; Malle 2006; Malle, Knobe, and Nelson 2007; Regan, Straus, and Fazio 1974; Ross 1977). The most relevant of these biases to this article is group-serving attribution bias (Miller and Ross 1975), which suggests that individuals are more likely to ascribe responsibility in a manner that is supportive of their predispositions by crediting an in-group and blaming an out-group. When it comes to making responsibility attributions for national conditions, group-serving partisan biases may assert themselves through motivated reasoning (Kunda 1990). Motivated reasoning argues that two types of motives lead individuals when making decisions: accuracy and direction. Accuracy motives are incentives to arrive at an accurate conclusion, whatever that may be, while directional motives encourage a particular outcome. Motivated reasoning, thus, occurs when directional goals bias reasoning. Directional goals can be thought to arise from a variety of sources such as possible rewards/punishments (Balcetis and Dunning 2006), personal behaviors (Kassarjian and Cohen 1965; Kunda 1987), prior beliefs (Lord, Ross, and Leper 1979), and partisanship (Rudolph 2003b, 2006; Taber and Lodge 2006). Prior work has elaborated on the role of partisanship in shaping subjective political attitudes (e.g., Bartels 2002) and I, like others who have done so previously, contend that party identification acts as a source of directional motivations (e.g., Atkeson and Maestas 2012, Kopko et al. 2011, and Rudolph 2003b say so explicitly, while Brown 2010, Malhotra and Kuo 2010, and Sirin and Villalobos 2011 implicitly suggest as much).
It is reasonable to expect an individual's party identification to affect responsibility attributions through motivated reasoning. (1) Party identification is a stable personal trait that affects all sorts of political behaviors (see Campbell et al. 1960; Lewis- Beck et al. 2008). Prior research has shown that party identification can have an effect on attitudes similar to responsibility attributions for national conditions such as the increased effectiveness of frames that reinforce partisan predispositions in affecting attributions for politicized issues like gun control (Heider-Markel and Joslyn 2001). Party identification can cause individuals to attach value and emotional significance to their party membership and associate strongly with fellow partisans, even viewing the party's successes and failures as their own personal successes and failures (Green, Palmquist, and Schickler 2002; Greene 1999). This sense of a partisan social identity can have consequences for individuals' attitudes; as Maestas et al. (2008) articulate, partisanship can create in-group and out-group dynamics in the attribution process whereby group members attribute positive outcomes to in-group members and negative outcomes to out-group members in order to reconcile negative outcomes with their own prior beliefs about their own party. This group-serving attribution bias (Miller and Ross 1975) can lead to partisanship acting as a clear source of bias in responsibility attributions.
How the attribution process takes place is unknown, though motivated correction theory proposes that attributions are often formed in a two-stage process, in which individuals alter initial attributions in order to obtain consistency between the attribution and one's core values or ideology (Morgan, Mullen, and Skitka 2010; Skitka et al. 2002). A similar process may occur when asking respondents about responsibility for national conditions after a governmental transition, except motivated by party identification (Maestas et al. 2008). By crediting members of their own party when they have positive evaluations and blaming members of the opposite party when their evaluations are negative, individuals are avoiding the internal conflict that may exist among their issue evaluations, party identification, and responsibility attributions (Morgan et al. 2010; Skitka et al. 2002). Hence, individuals are motivated by their political leanings to assign responsibility in a manner that is consistent with their group identities and beliefs about the world.
While one can imagine many influences on individuals' responsibility attributions, elites and the media are perhaps the most apparent. By covering some stories but not others, the media increases the salience of certain issues (Iyengar and Kinder 1987) and primes the public to consider those issues when evaluating leaders and ascribing responsibility (Edwards, Mitchell, and Welch 1995; Iyengar 1990; Kelleher and Wolak 2006; Krosnick and Kinder 1990). Studies show that presidents themselves can affect their own approval by priming the criteria that citizens use for evaluation (Druckman and Holmes 2004; Jacobs and Shapiro 1994), and those citizens look to elite rhetoric as a barometer for issue judgments after major political events (Woessner 2005). These studies of presidential approval mirror studies that have shown the manipulation of voter perceptions by politicians to influence attributions of responsibility in a similar way (McGraw 1991; McGraw, Best, and Timpone 1995). As summarized by Lebo and Cassino (2007), elite and media effects are themselves often biased through selective exposure (seeking out messages that support existing opinions), selective judgments (counter-arguing dissonant messages), and selective perception (only acknowledging messages that reinforce existing beliefs and ignoring dissonant ones). This article documents the uncertainty of responsibility caused by a presidential transition of power and the biased response of partisans; theory suggests that such biases would be the result of the exposure to elite and media messages.
Presidential Transitions as a Source of Responsibility Uncertainty
All else equal, people prefer to see co-partisans (whether incumbent or former officeholders) as responsible for positive conditions and members of the opposite party as responsible for negative conditions (Maestas et al. 2008; Miller and Ross 1975). However, the likelihood of making a motivated attribution across presidential administrations is going to depend on whether or not such an attribution is plausible in the individual's mind; accountability mechanisms prevent individuals from engaging in motivated reasoning (Taber, Lodge, and Glather 2001), and I argue that a clear consensus regarding who is responsible could do so as well. Thus, the cognitive hurdles that one must jump in order to rationalize an attribution will depend on whether there is clarity or uncertainty regarding responsibility.
Before addressing the larger question of responsibility attributions following a transition of power, we must first establish that responsibility uncertainty exists after a presidential transition because such confusion would make it reasonable to expect citizens to make motivated attributions. If it is true that transitions create responsibility uncertainty, the ambiguity over whether the incumbent or the former president is responsible for economic conditions should be greatest in the early stages of a new administration and decline over time. Here, I use two sources of data to highlight this changing certainty of responsibility, the comparative mentions of the current and former presidents in economic newspaper articles and the differences in economic evaluations of partisans as a president's term progresses.
Responsibility Ambiguity in the Media: Evidence from Coverage
Unfortunately, an objective measure of the clarity of responsibility does not yet exist. However, to estimate responsibility uncertainty, I tracked the mentions of President Bush and President Obama in economic articles in the New York Times to get a rough estimate of how credible it would be to assign responsibility to the presidents. (2) Since the number of economic articles varies at different times, I use a ratio of Obama mentions to Bush mentions to provide a relative sense of how much the media is connecting each president to the economy. While this is certainly an inexact measure of responsibility certainty, it is a reasonable assumption that the more the media associates a president with the economy, the more credible it would be to attribute responsibility to that president. If both presidents are frequently mentioned, then responsibility is ambiguous; the more lopsided the mentions, the more responsibility clarity that exists and the more likely it is that individuals will see that president as responsible for economic conditions. Thus, an increase over time in the relative mentions of Obama to mentions of Bush would imply an increase in the clarity of responsibility as Obama's administration progresses.
Figure 1 shows that at the beginning of his term, for every economic article that mentioned Bush, there were less than three that mentioned Obama. However, by the end of Obama's first year in office, that ratio increased to 4:1. If the above-cited studies are any indication, this suggests a decrease in the attribution of responsibility to Bush and an increase in the perceived responsibility of Obama. Importantly, however, this was not a steady increase over time, which implies that the believability of ascribing responsibility to Bush was just diminished and not completely extinguished. The volatility can be largely explained by the 2010 and 2012 elections, which are indicated by the vertical dotted lines in Figure 1. As each of these elections neared, there were increasing mentions of President Bush in economic articles, which accounts for the sharp decreases in the ratio. This suggests that reporters, editors, and interview subjects were still very willing to inject the former president into economic discussions. Clearly, while Obama's share of responsibility increased over time, Bush's role was not forgotten either.
Responsibility Ambiguity in the Electorate: Indirect Evidence from Issue Evaluations
We find further evidence of transitions of power causing responsibility confusion in the issue opinions of partisans. If presidential transitions of power cause responsibility confusion, partisans will find it easy to credit or blame the president of their choosing, and it will be very easy for citizens to make motivated attributions. There would be no need for motivated reasoning to influence individuals' issue evaluations if motivated attributions are easy to make early in a new president's term, and there will be little difference in the issue evaluations of partisans. As time goes by, however, the responsibility confusion decreases and incumbent's perceived responsibility presumably grows while the former president's diminishes. If it becomes harder to ascribe responsibility to the former president as time progresses, then we would expect the issue evaluations of partisan groups to diverge as a president's term goes on. (3)
While one would normally expect citizens to view economic conditions through a partisan filter (see Bartels 2002), the motivation to have a positive economic evaluation when your party is in power (or a negative one when they are out of power) must be accompanied by the ability to connect responsibility for those conditions back to your co-partisans (or members of the opposite party in the case of negative conditions). Early in a president's term, the ability of citizens to ascribe responsibility to either administration largely removes this motivation, and partisans are not motivated to make biased economic assessments because it is still very easy to blame the opposite party's president if they see conditions as negative. As time progresses, the credibility of assigning responsibility to the former administration declines, and, as a result, partisans will be increasingly motivated to make biased issue evaluations.
The ANES's 2008-10 panel study illustrates this divergence in the issue evaluations of partisan groups. The surveys asked whether the national economy improved since the time of President Obama's inauguration. (4) The question was included in the May and July 2009 waves, and again in June/July 2010. Table 1 presents the percentage of respondents having positive retrospective evaluations for the two partisan groups as well as the difference in that percentage between Democrats and Republicans.
In May and July of 2009, only a few months into Obama's term, there was only a slight difference in the economic evaluations of partisans: 10.7% and 9-0%, respectively. We observe a sharp divergence of issue evaluations among Democrats and Republicans in the June/July 2010 wave, however, 18 months into Obama's term, where the partisan difference in economic perceptions triples to 27 points. The cause of this was a large rise in the percentage of Democrats seeing economic improvement under Obama; nearly 40% of Democrats saw improvement compared to just 13% of Republicans. This is consistent with what we would expect to see if responsibility was becoming more certain; Republicans were now unlikely to perceive economic improvement under Obama because they did not want to credit him, while Democrats perceived improvement because they did not want to blame him. This was not the case early on in his term because it was relatively easier for Democrats to blame President Bush.
The newspaper mentions and the ANES panel study provide support for the assertion that responsibility uncertainty exists early in an administration and declines over time. However, in order to assess if voters assign credit and blame across presidential administrations based on their partisanship, we need to ask them whom they see as responsible. The remainder of this article attempts this using an original survey. The survey asked respondents to assign responsibility for current conditions to either President Bush or President Obama. This eliminates much of the uncertainty surrounding the attributions inferred from the existing data and leads to a clearer picture of how citizens make responsibility attributions.
Motivated Reasoning in Responsibility Attributions after a Transition
Theory and prior work suggest that individuals' characteristics and predispositions will affect their economic evaluations and responsibility attributions. From motivated reasoning's perspective, the most interesting of these characteristics is party identification. Party identification is an accessible and key piece of information about politicians and is a strong predictor of the vote (Campbell et al. 1960; Lewis-Beck et al. 2008). As described earlier, there is reason to suspect party identification to affect motivated reasoning as well.
With this in mind, it is reasonable to expect members of the incumbent president's party to blame a former president of the other party when evaluating conditions negatively (and credit a co-partisan incumbent president when they have positive evaluations). Those who are not members of the incumbent's party, on the other hand, will blame the incumbent for negative conditions while crediting the former officeholder for positive ones. By forming responsibility attributions in this way, the individual is protecting himself from internal conflict caused by supporting a party whose officials are responsible for the poor conditions. (5) The following hypothesis summarizes this reasoning:
Individuals will ascribe responsibility in a partisan manner by attempting to credit a co-partisan president for positive conditions and blame a president of the opposite party for negative conditions.
It is not necessary to consider attributions as a dichotomous choice between the incumbent and former president such as in the above example. Instead, it is possible, perhaps even likely, that individuals will assign a great deal of responsibility (or very little) to both individuals. Thus, it is appropriate to also measure how much responsibility people assign because individuals may still do so according to their partisan predispositions. This article therefore utilizes a new measure of responsibility assignment that captures
the amount of responsibility assigned to the current and former presidents. It follows from the hypothesis, then, that as issue evaluations increase, individuals will assign more responsibility to co-partisan presidents and less responsibility to presidents of the opposite party.
A survey funded by The Ohio State University's Graduate School was mailed to 2,500 randomly sampled registered voters in Franklin County, Ohio, in August 2010. (6) August was chosen for the survey because it is long enough after the 2009 inauguration of President Obama for some voters to contend that his policies might have had some effect on economic or Iraq War conditions while also occurring before the midterm election. (7) The survey measured the dependent variable, responsibility attributions, in two different ways. First, the survey asked respondents if President Bush or President Obama was more responsible for current economic and Iraq War conditions. In addition to these dichotomous attributions, respondents were also asked to assign responsibility amounts to each president on a seven-point scale ranging from "0-No Responsibility" to "6-Full Responsibility." This approach gives the respondent flexibility to assign equal or disparate amounts of responsibility to both targets.
By itself, knowing whom an individual views as more responsible cannot provide evidence of motivated reasoning because all individuals will not view issues in the same way, particularly partisans (Bartels 2002). Therefore, it is necessary to measure individuals' evaluations of issue conditions and pair them with their responsibility attributions. (8) The survey asked respondents if the national economy (and Iraq War conditions) was getting better, staying the same, or getting worse, again measured on a zero--six scale. Measuring issue evaluations together with responsibility attributions is vital because we must know whether the respondent views the issue in a positive or negative light. (9) For example, the hypothesis expects Democrats to see Obama responsible if they view conditions positively but to see Bush as responsible if conditions are negative. By interacting party identification with issue evaluations, we can test whether individuals are more likely to blame opposite-party members and to credit those who share their party identification.
The remaining variables are intended to control for their potential relationships with attributions and issue evaluations. Since polls consistently show women, minorities, and young people have a preference for Democrats, while wealthy, married, and religious persons tend to favor Republicans, age (measured in years), religiosity (frequency of attendance at religious services measured in six categories), income (eight categories), and dummies for gender, minority self-identification, and marital status are included as controls. Since evaluations of issue conditions might be dependent on information, education (six categories), political interest (four categories), and knowledge are included. The knowledge variable is an additive scale of correct responses to the five- question Delli Carpini and Keeter (1996) political knowledge scale, supplemented with an additional question about the number of votes needed to stop a filibuster. Finally, governmental trust is an additive scale of two five-point Likert scale questions and is included because trust may affect the amount of responsibility people assign to government officials.
Because this study intends to show that ordinary citizens make responsibility attributions in a partisan manner after a presidential transition, its results do not hinge on the sample's generalizability to the larger population. Nonetheless, there are reasons to suggest that the results of this sample are not unusual. In addition to Franklin County's reputation as a representative test market (Bandapudi 2009), there was no statistically significant difference from national census estimates regarding the respondents' sex or income. Respondents were more likely than the general population to be Caucasian, homeowners, and have a bachelor's degree. However, these characteristics are all highly correlated with political participation, making it more likely that they would be over-represented on voter registration lists in the first place (Verba, Schlozman, and Brady 1995). Finally, there is little evidence of a partisan bias in the returned surveys. While Ohio only records the partisan affiliation of primary voters, the random sample of those sent questionnaires had a two-party split of 51% Democratic, 49% Republican; among the returned surveys, 52% identified as Democratic and 48% as Republican (excluding all Independents). When Independent-leaners are included and group with the partisans, 44% of those in the sample were Republican, 42% were Democrats, and 14% were pure Independents. The sample was split evenly between Obama and John McCain voters. This localized sample limits strict generalization to the national population, but at the same time provides an important benefit. In order to demonstrate the occurrence of motivated reasoning, we can rule out that the variation in responsibility attributions is not the result of different political messages across geographic regions; this localized sample holds the messaging environment constant. (10)
Dichotomous Attributions of Responsibility
It is immediately evident that not including former officeholders in the response set can be a grave error, simply given the fact that 39% of the sample viewed President Bush as more responsible than President Obama for current economic conditions a year and a half into Obama's term. A similar amount (34%) saw Bush as more responsible for Iraq War conditions. Concerning economic evaluations, most people though the economy was relatively stagnant; 74% of respondents chose one of the middle three points on the scale. Among those who thought the economy was changing, people were much more likely to say it was getting worse than better (21% to 5%).
Table 2 shows the results of logistic regressions predicting an individual seeing Obama as more responsible than Bush. Looking first as the economy, the significant and negative interaction between party identification and the positivity of economic evaluations indicates support for the hypothesis (p < 0.001). Republicans and those with positive economic evaluations are less likely to credit President Obama with responsibility. After holding party identification constant, the marginal effect of individuals' evaluations of current economic conditions on the assignment of responsibility is statistically significant from zero for all Democrats, Republicans, and Republican-leaning Independents (Brambor, Clark, and Golder 2006).
The partisan differences in economic attributions are stark when shown in terms of predicted probabilities. In the first panel of Figure 2, the predicted probability of ascribing responsibility to President Obama is practically indistinguishable from 1.0 for all but the most optimistic strong Republicans, while that probability is statistically indistinguishable from 0.0 for strong Democrats with the lowest three evaluation levels. If Republicans view current economic conditions as improving, the likelihood of crediting Obama drops by 0.63 from the "Much Worse" to "Much Better" evaluation level. The difference in predicted probability is even steeper for Democrats, increasing 71 points across the economic evaluation scale. The difference in predicted probability across the strong partisan groups is statistically significant at p < 0.05 at the five lowest evaluations, which covers 95% of the respondents because so few people were willing to say the economy had greatly improved during the previous year.
The results for Iraq War responsibility, shown in the second column of Table 2, similarly support the hypothesis. The interaction between party identification and the perceived positivity of the war's conditions is statistically significant and negative (p < 0.001), indicating that Democrats will be less likely to ascribe responsibility to President Bush as their evaluations of conditions improve. The marginal effect of seeing Iraq War conditions as improving is statistically significant for pure Independents, and all Republicans, and approaches the p < 0.05 level for strong Democrats (Brambor, Clark, and Golder 2006). The second panel of Figure 2 shows the probability of seeing Obama as more responsible than Bush for Iraq War conditions. The likelihood of an attribution to Obama increases 52 points across the evaluation scale for strong Democrats and decreases 73 points for strong Republicans. There is a statistically significant difference across these partisan groups for the 70% of respondents who viewed Iraq War conditions as deteriorating or staying the same.
These results all indicate that when making responsibility attributions, individuals appear to assign credit and blame in a biased manner, aligning their attributions with their party identification and economic evaluations. When conditions are seen as good, partisan bias leads individuals to ascribe responsibility to credit the president who shares their party affiliation, whether he is the incumbent or the former officeholder. When conditions are poor, people blame the president of the opposite party. Furthermore, while not hypothesized, the interactions are larger for the economy than the Iraq War. This finding contradicts Sirin and Villalobos (2011), whose experiment showed that the relationship between partisanship and issue conditions in predicting attributions was greater for foreign issues than domestic ones. Though they did not analyze attributions across administrations, Sirin and Villalobos's theory may provide an explanation for the difference. They argue that people assign greater responsibility to the president for foreign issues than domestic ones because he has greater unilateral foreign policy power than in the domestic sphere, where the president shares more power with Congress. However, in 2010, there was unified government and President Obama faced relatively little opposition from his own party over his economic agenda. Furthermore, both Bush and Obama handled the economic crisis in a very personal and active manner through stimulus spending, which likely further connected economic conditions to the two presidents in the voters' minds.
Assigning Relative Amounts of Responsibility
In some situations, it may be too crude to measure responsibility dichotomously; a more accurate measure would be to determine the amount of responsibility assigned to each president, especially if there is reason to believe people will hold both individuals similarly responsible. This survey also measured the assignment of responsibility amounts to the incumbent and former presidents on a 0-6 scale. In general, most respondents assigned a great deal of economic responsibility to both presidents, though it was typical to assign about a half-point more to Obama; the mean amount assigned to Obama was 3.9 (standard deviation [std. dev.] 1.5) and the mean assigned to Bush was 3.4 (std. dev. 1.8). President Obama was seen as having more responsibility for the current conditions of the Iraq War as well, getting about 1.3 points more responsibility assigned to him than Bush, on average. In fact, about three-quarters of respondents assigned Obama a four, five, or six (mean 4.3, std. dev. 1.4). The assignment of Iraq War responsibility to Bush, on the other hand, was rather uniform (mean 3.0, std. dev. 2.0).
Table 3 presents seemingly unrelated regressions for each issue to model the assignment of responsibility to each president. (11) The first two columns show the economic models for each president. As predicted, the significant Party ID * Positivity of Issue Evaluation coefficients are positive in Bush's model and negative in Obama's. In other words, as Republicans' evaluations of conditions improve, they are more likely to assign greater responsibility to President Bush and less responsibility to President Obama. In Column 1, the marginal effect of just a one-point increase in economic evaluations among strong Republicans leads to a predicted 0.32 increase in the economic responsibility assigned to Bush; among strong Democrats, a one-point increase in perceptions drops responsibility assigned to Bush by 0.27. Since the survey measures perceptions on a seven-point scale, this is a substantively important relationship.
Motivated attributions occur when people assign economic responsibility to President Obama as well (column 2), though the marginal effect of issue evaluations is only significantly negative among Republicans. A strong Republican who thinks the economy has gotten much worse in the past year is predicted to assign Obama 2.9 fewer responsibility points than a strong Republican who thinks the economy has gotten much better (p < 0.001).
The Iraq War models also support the hypothesis. In Column 3, the marginal effect of Iraq War perceptions on the responsibility assigned to Bush is almost identical for strong Democrats and strong Republicans, though in opposite directions. A one-point increase in Iraq War evaluations lowers the predicted amount of responsibility assigned to Bush -0.33 points among strong Democrats but raises it by 0.30 points for strong Republicans ip < 0.05 for both).
Finally, responsibility assignment to President Obama for the war also follows the same substantive pattern. Strong Republicans are predicted to decrease the responsibility they assign to Obama by two points, if they say conditions have improved greatly rather than have gotten much worse (p < 0.01). The model predicts strong Democrats to increase the responsibility they give to Obama when evaluations improve, though this marginal effect is not statistically significant.
In conclusion, the assignment of responsibility amounts to both presidents for both issues supports the hypothesis. The worse that Republicans viewed either the economy or the Iraq War, the more responsibility they gave to President Obama and the less responsibility they assigned to President Bush, while the opposite held true for Democrats.
Discussion and Conclusion
This is the first article to examine how individuals make responsibility attributions across presidential administrations following a transition of power. The newspaper and ANES data illustrated that uncertainty over economic responsibility often exists at the beginning of a president's term and suggests that partisans confront such ambiguity by assigning responsibility according to partisan biases. The survey confirms this interpretation by directly measuring responsibility attributions across presidential administrations. By extending theories of motivated reasoning to the attribution process, this work extends the results of prior studies that find evidence of partisan biases in situations of responsibility ambiguity (Atkeson and Maestas 2012; Brown 2010; Maestas et al. 2012; Malhotra and Kuo 2010; Rudolph 2003b). In doing so, it concludes that partisan-motivated reasoning occurs in the attribution process after a new president takes office. Since responsibility is ambiguous, the desire for consistency between an individual's issue evaluation, party identification, and responsibility attribution acts as a directional motive that causes individuals to assign responsibility in a partisan manner, crediting co-partisans for perceived successes and blaming members of the opposite party.
This study also improved the measurement of responsibility attributions in survey research. First, the inclusion of the former officeholder alongside the incumbent in the response set is a simple addition but yields rich data, allowing for a more accurate study of responsibility assignment in the long periods following a transition of power.
Second, the measurement of scaled responsibility amounts provides for a deeper analysis than just asking which actor is more responsible. It is advantageous to measure responsibility in both ways; the dichotomous question eliminates ties and forces respondents to decide who has greater responsibility while the ordered responsibility scale introduces a great deal of nuance because respondents can assign similar or disparate responsibility amounts to multiple targets. When researchers cannot make both measurements, the decision will depend on the aims of the survey: researchers should use dichotomous question for clear-cut results and ordered scales for nuanced comparisons of accountability. Third, as Malhotra and Kuo (2010) note, practically every other study of political responsibility attributions focuses on the economy (though Sirin and Villalobos 2011 is now an important exception). After demonstrating the applicability of responsibility attributions to military issues, political scientists should continue to apply attribution theory to other international issues. In addition, since there were nearly identical results with both issues, we can be confident that the poor 2010 economy did not cause the observed patterns.
As with all cross-sectional designs, the results can only present a snapshot in time and must remain silent about causality and the effects of changing national conditions. Fortunately, we can place this study in the context of other attribution studies who have addressed causality. Sirin and Villalobos (2011) showed that subjects attributed more blame for deteriorating conditions and less credit for improving conditions to hypothetical opposite-party presidents. The panel studies analyzed by Gaines et al. (2007) showed that while partisans held accurate beliefs about changing facts regarding the Iraq War, they were also able to alter their interpretations of those facts according to their partisan views. Certainly, these causal conclusions are consistent with the correlative findings presented in this article regarding responsibility attributions to the current and former presidents.
This article also established that responsibility uncertainty exists early in a presidential term and appears to decline over time. This raises the question of whether the motivated attributions found in the survey would occur at the end of a president's term. One could speculate that partisans could perhaps alter their issue evaluations to bring them in line with preexisting attributions, as the ANES data suggested. Alternatively, citizens not wishing to see the incumbent responsible may shift responsibility to Congress or outside of government as it becomes harder to credit or blame the former president. Indeed, 48% of respondents in the 1998 ANES thought "business" and "working" people were more responsible for the economy than the president or Congress (Rudolph 2003b). Since existing data cannot address this question, it is certainly an avenue for future work.
Another question that remains is whether the differences in the policies pursued by successive administrations affect the willingness of citizens to hold one or both administrations responsible for conditions. If citizens associate a change in a policy with a positive or negative change in the issue, those citizens will likely attribute responsibility to the political actor who caused that policy to change. Such clear changes in policy are not always present, however. For instance, while there are significant differences between Obama and Bush, the two presidents acted similarly with regard to some rather high-profile economic and military policies; both pursued a "stimulus-based" strategy to halt the economic downturn, and when it came to dealing with Iraq, Obama did not implement anything resembling a 180-degree turnaround from Bush's strategy. In such situations, political scientists can draw on Brickman, Ryan, and Wortman's (1975) work on causal chains, which suggests that the existence of a prior cause often cancels out the liability ascribed to the immediate cause. For example, respondents who perceived little difference in Bush and Obama's economic policies, and did not like the economic situation that resulted, might hold Bush more liable for economic conditions since he is the antecedent "prior cause." This is why measuring the amount of responsibility assigned to each president can be helpful; citizens were able to articulate in this survey that they absolved neither president of responsibility for the sluggish economy, yet still had the option to place greater blame on either Bush or Obama. It remains unclear, however, how a greater or lesser contrast across administrations would affect attributions of credit and blame. If one assumes a similarity or dissimilarity of polices based solely on party identification, perhaps political scientists might be able to answer such questions in 2017, depending on which party wins the presidency.
The reliability and strength of these findings stress the importance of partisanship and issue evaluations on the attribution process. Political scientists cannot ignore these variables when looking at credit and blame. Because of the directional motivation that partisanship provides, researchers must consider it when looking at matters of responsibility and never presume that attributions are measures of objectively correct policies or political actions; individuals do appear to pass their attributions through a partisan filter before assigning responsibility. Importantly, it is not just party officials, politicians, and pundits who cast issue conditions in the most positive light possible; in the electorate, regular citizens are quite effective at rationalizing their perceptions of responsibility by matching their attributions with their sense of issue conditions. With this in mind, we can now understand the effect of economic events or other issues on elections more deeply by recognizing that people are able to interpret the sources of both positive and negative events in a manner consistent with their personal predispositions.
AUTHOR'S NOTE: This article includes the analysis of a survey that was partially funded by an Alumni Grant for Graduate Research from the Graduate School at The Ohio State University. I would like to thank Herb Weisberg, Paul Allen Beck, Kathleen McGraw, Jeffrey Budziak, Christopher Devine, and anonymous reviewers for their helpful comments in shaping this article.
Atkeson, Lonna Rae, and Cherie D. Maestas. 2012. Catastrophic Politics: How Extraordinary Events Redefine Perceptions of Government. New York: Cambridge University Press.
Bandapudi, Neeli. 2009. "Columbus, OH: Test City, USA." Interview with Neal Conan. Talk of the Nation. NPR, October 14.
Balcetis, Emily, and David Dunning. 2006. "See What You Want to See: Motivational Influences on Visual Perception." Journal of Personality and Social Psychology 91 (4): 612- 25.
Bartels, Larry M. 2002. "Beyond the Running Tally: Partisan Bias in Political Participation." Political Behavior 24 (2): 117-50.
Brambor, Thomas, William Roberts Clark, and Matt Golder. 2006. "Understanding Interaction Models: Improving Empirical Analyses." Political Analysis 14 (1): 63-82.
Brickman, Phillip, Kathleen Ryan, and Camille B. Wortman. 1975. "Causal Chains: Attribution of Responsibility as a Function of Immediate and Prior Causes ."Journal of Personality and Social Psychology 32 (6): 1060-67.
Brown, Adam. 2010. "Are Governors Responsible for the State Economy? Partisanship, Blame, and Divided Federalism." The Journal of Politics 72 (3): 605-15.
Campbell, Angus, Phillip Converse, Warren Miller, and Donald Stokes. 1960. The American Voter. New York: Wiley.
Cass, Connie. 2012. "Many Voters Still Blame Bush for Bad Economy." Associated Press, November 6.
Delli Carpini, Michael X., and Scott Keeter. 1993. "Measuring Political Knowledge: Putting First Things First." American Journal of Political Science 37 (4): 1179-206.
Druckman, James N., and Justin W. Holmes. 2004. "Does Presidential Rhetoric Matter? Priming and Presidential Approval." Presidential Studies Quarterly 34 (4): 755-78.
Duch, Raymond M. 2007. "The Economy and the Vote." In The Oxford Handbook of Comparative Politics, eds. Charles Boix and Susan C. Stokes. New York: Oxford University Press, 805-44.
Duch, Raymond M., Harvey D. Palmer, and Christopher J. Anderson. 2000. "Heterogeneity in Perceptions of National Economic Conditions." American Journal of Political Science 44 (4): 635-99.
Edwards, George C., Ill, and Alec Gallup. 1990. Presidential Approval: A Sourcebook. 2nd ed. Baltimore: Johns Hopkins University Press.
Edwards, George C., III., William Mitchell, and Reed Welch. 1995. "Explaining Presidential Approval: The Significance of Issue Salience." American Journal of Political Science 39 (1): 108-34.
Evans, Geoffrey, and Robert Andersen. 2006. "The Political Conditioning of Economic Perceptions." The Journal of Politics 68 (1): 194-207.
Feldman, Stanley. 1982. "Economic Self-Interest and Political Behavior." American Journal of Political Science 26 (3): 446-66.
Fiorina, Morris P. 1992. Divided Government. New York: McMillan.
Gaines, Brian J., James H. Kuklinski, Paul J. Quirk, Buddy Peyton, and Jay Verkuilen. 2007. "Same Facts, Different Interpretations: Partisan Motivations and Opinion on Iraq." The Journal of Politics 69 (4): 957-74.
Green, Donald, Bradley Palmquist, and Eric Schickler. 2002. Partisan Hearts and Minds: Political Parties and the Social Identities of Voters. New Haven, CT: Yale University Press.
Greene, Steven. 1999. "Understanding Party Identification: A Social Identity Approach." Political Psychology 20 (2): 393-403
Heider, Fritz. 1958. The Psychology of Interpersonal Relations. New York: Wiley.
Heider-Markel, Donald P., and Mark R. Joslyn. 2001. "Gun Policy, Opinion, Tragedy, and Blame Attribution: The Conditional Influence of Issue Frames." The Journal of Politics 2 (2): 520-43.
Holbrook, Thomas. 2012. "Forecasting US Presidential Elections." In The Oxford Handbook of Political Behavior, ed. Jan E. Leighley. New York: Oxford University Press, 346-73.
Iyengar, Shanto. 1989. "How Citizens Think about National Issues: A Matter of Responsibility." American Journal of Political Science 33 (4): 878-900.
--. 1990. "Framing Responsibility for Political Issues: The Case of Poverty." Political Behavior 12 (1): 19-40.
Iyengar, Shanto, and Donald R. Kinder. 1987. News That Matters. Chicago: University of Chicago Press.
Jacobs, Lawrence R., and Robert Y. Shapiro. 1994. "Issues, Candidate Image, and Priming: The Use of Private Polls in Kennedy's 1960 Presidential Campaign." American Political Science Review 88 (3): 527-40.
Jones, Edward E., and Richard E. Nisbett. 1971. The Actor and the Observer: Divergent Perceptions of the Causes of Behavior. New York: General Learning Press.
Jones, Edward E., and Victor A. Harris. 1967. "The Attribution of Attitudes." Journal of Experimental Social Psychology 3 (1): 1-24.
Kassarjian, Harold H., and Joel B. Cohen. 1965. "Cognitive Dissonance and Consumer Behavior." California Management Review 8 (1): 55-64.
Kelleher, Christine A., and Jennifer Wolak. 2006. "Priming Presidential Approval: The Conditionality of Issue Effects." Political Behavior 28 (3): 193-210.
Kopko, Kyle C., Sarah McKinnon Bryner, Jeffrey Budziak, Christopher J. Devine, and Steven P. Nawara. 2011. "In the Eye of the Beholder? Motivated Reasoning in Disputed Elections." Political Behavior 33 (1): 271-90.
Kramer, Gerald H. 1971. "Short-Term Fluctuations in U.S. Voting Behavior." American Political Science Review 65 (1): 131-43.
Krosnick, Jon A., and Donald R. Kinder. 1990. "Altering the Foundations of Support for the President through Priming." American Political Science Review 84 (2): 497-512.
Kunda, Ziva. 1987. "Motivation and Inference: Self-Serving Generation and Evaluation of Evidence." Journal of Personality and Social Psychology 53 (3): 636-47.
--. 1990. "The Case for Motivated Reasoning." Psychological Bulletin 108 (3): 480-98.
Lau, Richard R., and David O. Sears. 1981. "Cognitive Links between Grievances and Political Responses." Political Behavior 3 (4): 279-302.
Lebo, Matthew J., and Daniel Cassino. 2007. "The Aggregated Consequences of Motivated Reasoning and the Dynamics of Partisan Presidential Approval." Political Psychology 28 (6): 719-46.
Lewis-Beck, Michael S., William G. Jacoby, Helmut Norpoth, and Herbert F. Weisberg. 2008. The American Water Revisited. Ann Arbor: University of Michigan Press.
Lord, Charles G., Lee Ross, and Mark R. Leper. 1979. "Biased Assimilation and Attitude Polarization: The Effects of Prior Theories on Subsequently Considered Evidence. "Journal of Personality and Social Psychology 37 (11): 2098-109.
Lowry, Robert C., James E. Alt, and Karen E. Ferree. 1998. "Fiscal Policy Outcomes and Electoral Accountability in American States." American Political Science Review 92 (4): 759-74.
Maestas, Cherie D., Lonna Rae Atkeson, Thomas Croom, and Lisa A. Bryant. 2008. "Shifting the Blame: Federalism, Media and Public Assignment of Blame Following Hurricane Katrina." Publius: The Journal of Federalism 38 (4): 609-32.
Malhotra, Neil, and Alexander Kuo. 2008. "Attributing Blame: The Public's Response to Hurricane Katrina." The Journal of Politics 70 (1): 120-35.
Malle, Bertram F. 2006. "The Actor-Observer Asymmetry in Causal Attribution: A (Surprising) Meta-Analysis." Psychological Bulletin 132 (6): 895-919.
Malle, Bertram F., Joshua Knobe, and Sarah E. Nelson. 2007. "Actor-Observer Asymmetries in Explanations of Behavior." Journal of Personality and Social Psychology 93 (4): 491-514.
McAvoy, Gregory E. 2006. "Stability and Change: The Time Varying Impact of Foreign Policy Evaluations on Presidential Approval." Political Research Quarterly 59 (1): 71- 83
--. 2008. "Substance versus Style: Distinguishing Presidential Job Performance from Favor ability." Presidential Studies Quarterly 38 (2): 284-300.
McGraw, Kathleen M. 1991. "Managing Blame: An Experimental Test of the Effectiveness of Political Accounts." American Political Science Review 85 (4): 1133-57.
McGraw, Kathleen M., Samuel Best, and Richard Timpone. 1995. "What They Say or What They Do?" The Impact of Elite Explanation and Policy Outcomes on Public Opinion." American Journal of Political Science 39 (1): 53-74.
Miller, Dale T., and M. Ross. 1975. "Self-Serving Attribution Biases in the Attribution of Causality: Fact or Fiction?" Psychological Bulletin 82 (2): 213-25.
Morgan, G. Scott, Elizabeth Mullen, and Linda Skitka. 2010. "When Values and Attributions Collide: Liberals' and Conservatives' Values Motivate Attributions for Alleged Misdeeds." Personality and Psychology Bulletin 36 (9): 1241-54.
Nicholson, Stephen P., and Gary M. Segura. 1999. "Midterm Elections and Divided Government: An Information-Driven Theory of Electoral Volatility." Political Research Quarterly 52 (3): 609-29.
Nicholson, Stephen P., Gary M. Segura, and Nathan D. Woods. 2002. "Presidential Approval and the Mixed Blessing of Divided Government." The Journal of Politics 64 (3): 701- 20.
Peffley, Mark. 1984. "The Voter as Juror: Attributing Responsibility for Economic Conditions." Political Behavior 6 (3): 275-94.
Peffley, Mark, and John T. Williams. 1985. "Attributing Presidential Responsibility for National Economic Problems." American Politics Quarterly 13 (4): 393-425.
Petrocik, John R., and Frederick T. Steeper. 1986. "The Midterm Referendum: The Importance of Attributions of Responsibility." Political Behavior 8 (3): 206-29.
Regan, Dennis T., Ellen Straus, and Russell Fazio. 1974. "Liking and the Attribution Process." Journal of Experimental Social Psychology 10 (4): 385-97.
Ross, Lee D. 1977. "The Intuitive Psychologist and His Shortcomings: Distortions in the Attribution Process." Advances in Experimental Social Psychology 10: 173-220.
Rudolph, Thomas J. 2003. "Institutional Context and the Assignment of Political Responsibility." The Journal of Politics 65 (1): 190-215.
--. 2003b. "Who's Responsible for the Economy? The Formation and Consequences of Responsibility Attributions." American Journal of Political Science 47 (4): 698- 713.
--. 2006. "Triangulating Political Responsibility: The Motivated Formation of Responsibility Judgments." Political Psychology 27 (1): 99-122.
Sigelman, Lee, and Kathleen Knight. 1985. "Public Views of Presidential Responsibility for the Economy: Understanding Personalization." Political Behavior 7 (2): 167-91
Sirin, Cigdem V., and Jose D. Villalobos. 2011. "Where Does the Buck Stop? Applying Attribution Theory to Examine Public Appraisals of the President." Presidential Studies Quarterly 41 (2): 334-57.
Skitka, Linda J., Elizabeth Mullen, Thomas Griffin, Susan Hutchinson, and Brian Chamberlin. 2002. "Dispositions, Scripts, or Motivated Correction? Understanding Ideological Differences in Explanations for Social Problems." Journal of Personality and Social Psychology 83 (2): 470-87.
Sniderman, Paul M., and Richard A. Brody. 1977. "Coping: The Ethic of Self- Reliance." American Journal of Political Science 21 (3): 501-23.
Taber, Charles S., and Milton Lodge. 2006. "Motivated Skepticism in the Evaluation of Political Beliefs." American Journal of Political Science 50 (3): 755-69.
Taber, Charles, Milton Lodge, and Jill Glather. 2001. "The Motivated Construction of Political Judgments." In Citizens and Politics: Perspectives from Political Psychology, ed. James Kuklinski. New York: Cambridge University Press, 198-226.
Trumbull, Mark. 2012. "How Obama Won in a Weak Economy: Voters Didn't Blame Him." Christian Science Monitor, November 8.
Verba, Sidney, Kay Lehman Schlozman, and Henry E. Brady. 1995. Voice and Equality: Civic Volunteerism in American Politics. Cambridge, MA: Harvard University Press.
Woessner, Matthew C. 2005. "Scandals, Elites, and Presidential Popularity: Considering the Importance of Cues in Public Support of the President." Presidential Studies Quarterly 35 (1): 94-116.
STEVEN P. NAWARA
(1.) When examining the relationship between issue evaluations and responsibility attributions, the political science literature has typically treated issue evaluations as the independent variable and attributions as the dependent variable (e.g., Peffley 1984; Rudolph 2003). However, it should be acknowledged that the reverse causal direction is also possible; in some circumstances, it is possible for responsibility attributions to be the causes of how people view the economy, rather than the effect of economic evaluations. Further research is needed to confirm the direction of causality in this relationship.
(2.) When counting the number of "Bush" and "Obama" occurrences, a Lexis-Nexis Academic search of New York Times articles was restricted to those included in the subject area "Economy and Economic Indicators."
(3.) While the evidence showing increased differences in issue evaluations by partisan groups is certainly supportive of the theory that responsibility uncertainty is greatest at the beginning of a presidential term, I cannot rule out the possibility of these changes having other causes.
(4.) The question was phrased, "Now thinking about the economy in the country as a whole, would you say that as compared to January 2009, the nation's economy is now better, about the same, or worse?"
(5.) Atkeson and Maestas (2012) found that such partisan rationalizations are by no means an automatic response by individuals to attribution questions; they argue that the anxiety of respondents during a crisis, as well as paying greatly increased attention to the media, increases accuracy motives while making an attribution (Kunda 1990).
(6.) The survey was administered in a mixed-mode format, and respondents could either return the questionnaire in a prepaid envelope or take the survey online. Responses were heaviest in the beginning and middle of August, with 92% received by September 1; the survey closed on October 5. There were 439 full and partial responses to the survey; the response rate based on the total number of mailed questionnaires is 18%. However, 284 questionnaires were returned as undeliverable mail due to inaccuracies on the voter registration list; correcting for this increases the cooperation rate to 20%.
(7.) The analysis of the ANES data earlier in this article suggests that we would observe similar results if this survey taken place earlier in Obama's first term, perhaps with an even greater partisan bias since the data suggested that it becomes harder to ascribe responsibility to the former president as time goes on.
(8.) Party identification can also shape an individual's perceptions of national conditions (Bartels 2002; Duch, Palmer, and Anderson 2000; Evans and Anderson 2006). While some degree of endogeneity certainly exists, I believe it incorrect to treat issue perceptions as completely determined by partisanship; the two variables are indeed distinct. Duch (2007), for instance showed that objective information does indeed influence national economic evaluations independently of partisanship.
(9.) To account for potential question order bias, surveys randomly ordered the issue evaluations and responsibility attribution batteries.
(10.) I would like to thank an anonymous reviewer for this insight.
(11.) I also replicated Table 3 using ordered- and generalized ordered-logistic regressions. The results are statistically and substantively similar and are available from the author.
Steven P. Nawara is assistant professor of political science at Lewis University. His research interests include voting behavior, public opinion, and political psychology.
TABLE 1 Percentage with a Positive Retrospective Economic Evaluation Democrats Republicans Partisan Gap May 2009 20.4 9.7 10.7 July 2009 15.5 6.5 9.0 June/July 2010 39.4 12.5 26.9 Source: 2008-10 American National Elections Panel Study TABLE 2 Seeing Obama as More Responsible than Bush National Economy Coefficient S.E. Party ID * Positivity of -0.27 * 0.08 Issue Evaluation Party ID 1.43 * 0.27 Positivity of Issue Evaluation 0.96 ** 0.32 Female 0.16 0.30 Minority -0.39 0.43 Married/Widowed -0.06 0.35 Age -0.02 * 0.01 Education -0.16 0.13 Income -0.04 0.10 Religiosity 0.23 ** 0.09 Interest 0.02 0.19 Trust in Government 0.06 0.08 Knowledge 0.10 0.12 Intercept -4.73 * 1.36 n 331 Log-likelihood -163.33 Iraq War Coefficient S.E. Party ID * Positivity of -0.19 * 0.06 Issue Evaluation Party ID 0.81 * 0.21 Positivity of Issue Evaluation 0.49 *** 0.23 Female 0.46 * 0.27 Minority -0.77 ([dagger]) 0.40 Married/Widowed 0.02 0.33 Age 0.00 0.01 Education -0.02 0.12 Income 0.10 0.09 Religiosity 0.03 0.08 Interest -0.14 0.18 Trust in Government 0.11 0.08 Knowledge 0.04 0.12 Intercept -2 44 *** 1.14 n 330 Log-likelihood -187.29 Logistic Regressions. DV: 1 = Obama more responsible, 0 = Bush more responsible. Party ID is coded on a seven-point scale, from Strong Democrat to Strong Republican * indicates p < 0.001, ** p < 0.01, *** p < 0.05, ([dagger]) p < 0.1. TABLE 3 Assigning Amounts of Responsibility to Individual Presidents Economy Pres. Bush Pres. Obama Party ID * 0.10 ** (0.04) -0.09 ** (0.03) Positivity of Issue Evaluation Party ID -0.60 * (0.12) 0.42 * (0.10) Positivity of -0.37 *** (0.18) 0.15 (0.16) Issue Evaluation Female 0.16 (0.19) 0.21 (0.17) Minority -0.27 (0.31) -0.07 (0.28) Married/Widowed -0.421 ([dagger]) (0.24) -0.50 *** (0.21) Age 0.01 (0.08) 0.00 (0.01) Education -0.02 (0.06) 0.03 (0.08) Income 0.04 (0.06) 0.05 (0.05) Religiosity -0.05 (0.06) 0.06 (0.05) Interest 0.26 *** (0.13) 0.19 ([dagger]) (0.11) Trust in -0.01 (0.05) -0.03 (0.05) Government Knowledge -0.11 (0.08) -0.02 (0.07) Constant 5.18 * (0.79) 1.56 *** (0.71) n 323 323 Iraq War Pres. Bush Pres. Obama Party ID * 0.10 *** (0.04) -0.08 *** (0.03) Positivity of Issue Evaluation Party ID -0.51 * (0.14) 0.33 ** (0.11) Positivity of - 0.44 *** (0.19) 0.21 (0.14) Issue Evaluation Female -0.00 (0.22) 0.22 (0.16) Minority -0.14 (0.36) -0.41 ([dagger]) (0.26) Married/Widowed -0.47 ([dagger]) (0.27) -0.42 *** (0.20) Age 0.00 (0.01) 0.00 (0.00) Education 0.04 (0.10) 0.11 (0.07) Income -0.02 (0.07) 0.03 (0.05) Religiosity -0.08 (0.06) -0.00 (0.05) Interest 0.18 (0.15) 0.17 (0.11) Trust in 0.03 (0.06) -0.04 (0.05) Government Knowledge -0.01 (0.10) -0.07 (0.07) Constant 4.80 * (0.92) 2.71 * (0.67) n 321 321 Results are separate seemingly unrelated regression models for each issue with standard errors in parentheses. DV = Amount of responsibility assigned to the target on a 0-6 scale. Party ID is coded on a seven-point scale, from Strong Democrat to Strong Republican * indicates p < 0.001, ** p < 0.01, *** p < 0.05, ([dagger]) p < 0.1.
|Printer friendly Cite/link Email Feedback|
|Author:||Nawara, Steven P.|
|Publication:||Presidential Studies Quarterly|
|Date:||Mar 1, 2015|
|Previous Article:||Shifting currents: Dwight Eisenhower and the dynamic of presidential opportunity structure.|
|Next Article:||Leviathan's reach? The impact of political consultants on the outcomes of the 2012 republican presidential primaries and caucuses.|