«A Dissertation Submitted to the Graduate Faculty of the Louisiana State University and Agricultural and Mechanical College in partial fulfillment of ...»
This study also measured political interest and knowledge. Political interest was measured using five items. Two of these items have been used in the past to assess interest as long-term interest in politics (Jin, An, & Simon, 2009). These items asked participants, “how involved in politics would you say you currently are?” and, “in the past, how much have you been interested in political campaign during election times?” Two other items asked, “how much do you enjoy keeping up with the news?”; “how much do you enjoy news about national politics?”; and, “how much do you enjoy talking about politics with family, friends, or peers?” A principal components factor analysis revealed a single dimension to these items. Subsequently, the items were combined into a composite scale (M = 1.99, sd =.67, range = 0 to 4, alpha =.84).
Political knowledge was assessed with the five-item index recommended by Delli Carpini and Keeter (1996). Close-ended items were used that asked participants if they could identify the vice president (Joe Biden), whose responsibility it is to determine if a law is constitutional (the Supreme Court), how much of a majority is required for Congress to override a presidential veto (2/3 majority), which party has the majority in the House of Representatives (Republican Party),
combined into an composite index (M =.82, sd =.23, range = 0 to 1, alpha =.65).
Independent Variables. The primary independent variable was random exposure to the experimental treatments (Appendix B). Participants were randomly exposed to a news story involving a Republican or Democratic official (party cue) accused of financial or sexual misconduct (scandal type cue). Each story uses similar language to describe the scandal scenario, and was informationally similar, except for details specific to the type of scandal described.
Treatment groups were approximately balanced, with n = 34 assigned to each treatment.
Dependent Variable. Evaluations of the accused official, Senator Sam Hall, were assessed using a 100-point feeling thermometer. This measure is identical to those currently used in the American National Election Study (ANES). After exposure to one of the news stimuli, all participants were asked, “to indicate how you feel about Senator Sam Hall” from 0 (very negative attitude) to 100 (very positive attitude) (M = 22.82, sd = 20.90, range = 0 to 83).
Since the dependent variable was interval-level, the appropriate statistical modeling procedure was ordinary least squares (OLS) regression (Cohen, Cohen, West & Aiken, 2002;
Kennedy, 2005). The main effects of receiving the Republican or Democratic party cue manipulation, and receiving the Financial or Sexual scandal cue manipulation were entered in Model 1 as independent variables, along with interaction terms in Model 2.
The analysis first examines the main effects of the party cue and the scandal type cue on evaluations of the accused official, Senator Sam Hall (Table 3). Two models are shown in Table 3: Model 1 is unconditional and Model 2 shows results with interactions entered.
This study did not make specific predictions about the main effect of party cue on participant evaluations, since it is assumed that participant evaluations of the accused official will be at least partially based on participant political ideology, which should mitigate the main
Effect of Scandal Type Cue on Evaluations of Accused Official Note: n = 148. The Y-axis shows predicted values on a 100-point feeling thermometer, generated from Table 3, model 1. Bars indicate scandal type.
related to more negative evaluations of the accused official. It also indicates that exposure to the sex scandal treatment led to slightly more positive evaluations. H5, which predicted learning about an official’s financial misconduct would result in more negative evaluations than learning
Interaction Between Party Cue and Participant Ideology Note: n = 148.
The Y-axis represents the predicted value of evaluations, by participant ideology and scandal type. Predicted values were generated from Table 3, model 2. Trend lines indicate party cue.
was not expected, as there should be an interactive effect between participant ideology and receiving information about a likeminded or opposition party official. Indeed, there is an
relationship is illustrated in Figure 7 by plotting predicted values by ideology and party cue.
As shown in the figure, conservative participants give more positive evaluations on a Republican official accused of scandal and more negative evaluations of a Democrat. The reverse is true of liberal participants, who made more negative evaluations of the Republican, and relatively more positive evaluations of the Democrat. H6A predicted that conservatives would be more negative toward a Democrat accused of scandal than a Republican. H6A receives support. H6B expected that liberals would be more negative toward a Republican, and relatively more positive evaluations of a Democrat accused of scandal. H6B is supported. Consistent with research showing humans naturally protect members of their group (Brewer, 1979; Tajfel 1982), including their political groups (Huddy, 2013), these results suggest liberals and conservatives consider the party affiliation of the accused official when making evaluations about them.
Based on the assumption that ideologically motivated moral values concerning Fairness and bodily Sanctity may influence scandal evaluations (Graham et al., 2009; Haidt, 2013), H7A and H7B predicted an interaction between participant ideology and scandal type. The results do not support H7A or H7B. There was no interaction between participant ideology and the type of scandal described in the treatments – financial or sexual – suggesting that moral values associated with political ideology may not exert independent effects on evaluations of the accused official. Nevertheless, it appears that officials accused of a sex scandal are generally evaluated more positively than officials accused of financial scandals.
political knowledge and scandal type. Here it was predicted that high knowledge levels of knowledge would place harsh evaluations on an official accused of financial misconduct than would individuals with low knowledge levels (H8A). Conversely, it was expected that individuals with high knowledge levels may rate officials accused of sex scandals more harshly
Interaction Between Scandal Type Cue and Political Knowledge Note: n = 148.
The Y-axis represents the predicted value of evaluations, by political knowledge and scandal type. Predicted values generated from Table 3, model 2. Trend lines indicate scandal type cue.
type (β = 0.23, p.05). As shown in Figure 8, going from the minimum to the maximum value of the political knowledge scale resulted in a 40-point decrease in mean evaluation scores for participants exposed to the sex scandal cue. Participants with high knowledge levels appeared to judge the sex scandal more negatively than individuals with low levels of knowledge. A similar pattern of results is found with regard to financial scandals, to a lesser degree. The more knowledgeable the participant, the lower the mean evaluation for an official accused of a financial scandal. However, even participants who scored zero on the knowledge scale judged an official accused financial misconduct more severely than an official accused of an affair.
Democratic citizens must perform a number of duties. To be considered “active” and “engaged” members of their polity, they must participate in the popular election of government representatives; they must serve on juries of their peers; they may elect to serve as civic leaders in their local communities, or at the state or national level; they must be willing to be drafted into the armed services in times of military conflict; and they must be willing, at times, to remove popular but corrupt representatives from office (Clawson & Oxley, 2012).
In this chapter, a survey-experiment demonstrated that citizens are willing and able to penalize representatives who cheat them financially. Citizens also punish representatives accused of sex scandals, but not to the same degree as representatives accused of financial misconduct.
Combined, these results suggest that citizens are “rational,” insofar as they appear to recognize typological differences in scandal severity when making evaluations about the accused, punish members of the opposition partisan group more harshly than members of their own group, and are not swayed by rather rigid moral values in assessing officials accused of sexual and financial
protect members of one’s political group (Huddy, 2013) and to preserve representative power in Washington, rather than a moral judgment about the official’s conduct.
These results can be interpreted optimistically. As hypothesized (H4), individuals deposit more negative evaluations on more severe offenses of public trust, in this case, financial scandals.
Financial scandals represent violations of a person’s duty as an elected representative, violating norms of democracy, as well as ethical and legal norms. Sex scandals can involve activities that violate each of these norms, but they are often merely private violations of social mores governing what is considered “correct” or “acceptable” sexual behavior. Citizens respond negatively to accusations of sexual scandal, but they appear to account for the egregiousness of the offense, adjusting evaluations to meet the severity of the alleged misconduct. Consistent with previous analyses (Doherty et al., 2011; Sigal et al. 1988), it seems that citizens are able to make distinctions in scandal severity. However, this assumes that citizens receive equitable information about various scandal accusations. This is not a safe assumption, as news media tend to focus on sexual scandals to a greater extent than they do less salacious financial scandals.
Citizens also appear to make group-related calculations about scandal. Human beings have a natural affinity for, and are predisposed to protect members of their group (Peters, 1979;
Tajfel, 1982). In the realm of politics, this includes protecting partisans who identify with your ideological identity (Huddy, 2013). This chapter found evidence to support the prediction that ideology sways evaluations of accused representatives (H5A and H5B). Liberals and conservatives evaluate members of their group negatively in wake of a scandal, but evaluate the opposition party members even more negatively. This is quite rational, considering that partisan officials theoretically represent one’s political interests in larger governing bodies. A scandal
from office, and will no longer be able to represent the citizen’s interests. Given this knowledge, citizens are less willing to punish members of a group who likely promote their own interests.
Moral judgments related to political ideology do not appear to play a significant role in individual assessments of scandal. Specifically, the moral value dimensions of Fairness / Cheating and Sanctity / Degradation do not appear alter assessments of scandal (Graham et al., 2009; Haidt, 2012). Counter to the prediction (H6A and H6B), there was not a significant interaction between participant ideology and scandal type, suggesting that moral values associated with ideology do not play a major role in scandal evaluations, all else being equal.
One explanation for this finding may be that the fictitious scandal scenario was not sensational or vivid enough to trigger feelings of moral outrage. To trigger disgust, the moral foundations approach uses scenarios describing sexual relations that are often considered deviant (e.g., homosexual relations), designed to activate judgments based on the Sanctity / Degradation value (Haidt, 2012). The stimuli used in this study may not have been effective at activating this value, since it was a news story providing a detached depiction of improper relations between a congressman and a mistress (Appendix B). It is possible that treatments highlighting more salacious sex scandals could prime the Sanctity / Degradation value dimension, subsequently altering scandal judgments. Likewise, stories that emphasize an official’s lack of fairness to his constituents may prompt evaluations based on the Fairness / Cheating value.
Political knowledge also plays a key role in scandal evaluations (H8A and H8B). The evidence suggests that judgments about officials accused of misconduct is shaped by the participant’s level of political sophistication. Participants with high levels of knowledge appear less capable of processing the magnitude of the misconduct, assigning more negative evaluations
compared to high knowledge individuals, gave a much more positive rating to officials accused of an extramarital affair, suggesting that low knowledge participants may not have been able to attach the information about scandal to their prior beliefs.
Since knowledge is related to making what could be called “partisan errors” in information processing (Chong & Druckman, 2007a; 2007b; Druckman et al., 2013), it appears that knowledge also motivates selective processing of scandal information. Rather than being able to gauge the severity of scandal, knowledge may motivate citizens to connect news about the “game” of politics – scandal, in this case – to their prior partisan affiliations and ideological orientations. This translates into more negative evaluations of an official accused of sexual dalliances. Individuals at every level of knowledge, meanwhile, judged officials accused financial misconduct more negatively than officials accused of sexual misbehavior.