5 Getting the Word Out and Going for Votes John McCain's 2008 presidential run was an uphill climb. He clawed his way back to victory in the Republican primary only to find himself pitted against Barack Obama in the general election, a candidate whose message of hope and change had propelled him to celebrity-like status around the world. Although initial polling put Obama and McCain in a dead heat, it soon became evident that the Republican senator would have to close the enthusiasm gap by enhancing his personal image and convincing voters that his experience and policies were right for the country moving forward.1 McCain set out to make his case with all of the usual campaign strategies and a heavy dose of entertainment talk show interviews. He made countless speeches across the country, his team aired millions in television ads and developed a sophisticated web presence, and he maintained a hectic schedule of appearances across an array of shows. He was a frequent guest on news programs like Meet the Press, Larry King Live, and Good Morning America while also making stops at the most prominent entertainment venues from The Tonight Show, The Late Show, and The Daily Show to Live with Regis and Kelly, The View, and Ellen. In fact, McCain, who had developed an affinity for these shows over the years, engaged in the most prolific entertainment talk show strategy to date. Some of his interviews highlighted his playful spirit, others delved into his policy proposals, and still others combined personal and political topics. The press kept tabs on McCain's interviews along the way. Although they modestly praised most of his performances, including his "good natured sparring" with David Letterman, his willingness to laugh along with Jimmy Kimmel, and his ability to communicate policy ideas on The Daily Show, they also made note of his less successful appearances.2 There were stories, for example, describing the awkwardness with which McCain returned to The Late Show after ducking out on Letterman four weeks earlier, how the audience booed when McCain became defensive as he was "clobbered over the head with questions" during "a pretty good grilling" on The View, and how he "clashed" with Ellen DeGeneres over the fact that they "don't exactly see eye to eye when it comes to gay marriage."5 These descriptions, while telling, 130 Talk Show Campaigns paint little more than an anecdotal picture of the effect that McCain's entertainment talk show strategy had on his chances in the race. They give strong impressions but leave important questions unanswered about the extent to which these interviews actually helped (or hurt) McCain. Did talk shows make any real difference in what viewers knew about him and how they felt towards him, or were they merely entertainment with little strategic value? This kind of media commentary and assessment is common for most entertainment talk show interviews. Articles and reports generally start by highlighting candidates' objectives with claims that they are hitting the late night stage to "polish their regular guy aura" or "to show that they can laugh at themselves," while daytime talk shows are characterized as the perfect place for candidates to "soften their image" and "show their more personal side."4 Reporters might then assess the interview with casual comments on the extent to which the candidate "appeared relaxed," "joked with the host," or "engaged the audience."5 There is almost never any serious discussion of how the interview might have actually affected the millions of viewers watching from home. Even researchers have only just begun to explore the impact that these interviews have on those who watch them. A small number of studies have produced some useful findings, but most of this work has either focused on a limited number of interviews and/or used imprecise exposure measures to estimate effects." There is still a great deal that we do not know about what, if anything, viewers learn from these interviews and how it affects their feelings about the race. The result is that candidates have spent more than 20 years going on these shows with clear and widely publicized objectives, and yet there is relatively little precise information on how viewers are likely to react. In this chapter, I use results from six online experiments conducted during the 2008 campaign to explore the impact that entertainment talk show interviews have on viewers' political decisions. I am specifically interested in the extent to which these interviews inform, prime, and ultimately persuade viewers. What, if anything, do viewers learn from watching? When are these interviews most likely to affect the criteria viewers use to evaluate candidates? And, most importantly, when are candidates most likely to generate support by going on these shows? I start in the next section with a theoretical discussion about the potential impact that these encounters could have on those who watch them. I then describe the experimental process used to test these predictions before presenting the results. The chapter concludes with a brief discussion of what these findings say about candidate and voter behavior in an age of entertainment politics. THE POTENTIAL IMPACT OF ENTERTAINMENT TALK SHOW INTERVIEWS Entertainment talk show interviews have the potential to influence viewers in three critical ways. They are likely, under certain conditions, to affect Getting the Word Out and Going for Votes 131 what viewers know about the candidate, what they think about when evaluating the candidate, and ultimately the extent to which they are inclined to support the candidate. As such, they have the potential to significantly help those making a bid for the White House. Informing voters is a critical first step in the persuasion process. Long-shot hopefuls and even well-known contenders recognize that voters must first receive and comprehend their message if they are to support it, and while some voters need little more than the candidates' party affiliation to make their decision, most seek out additional information about candidates and their positions on certain issues. There is considerable debate about what viewers learn from "soft news" programs that include political material (e.g., Entertainment Tonight, Oprah). While some studies show higher political knowledge among viewers of certain soft news shows (e.g., The Daily Show) and many Americans claim to regularly learn about politics from entertainment-based sources, critics contend that they actually absorb little, if anything, from these shows. They argue that although amusement-seeking viewers may encounter bits of political information, most of what they come across focuses on the sensational or arguably less important aspects of political life—"roughly sex, drugs and celebrities"—and even then, they show only a weak tendency to process and recall this information.'- Questions remain about the precise educational value of watching these shows. Candidate interviews on entertainment talk shows, however, are different in that they routinely feature more of the substantive information that may be lacking from day-to-day episodes (see Chapter Four). Matthew Baum's research shows that when "substantive information is presented in an entertaining context, it can be piggy-backed (i.e., attached) to information intended primarily to entertain [which] renders political information cost-effective for even apolitical individuals."8 In other words, viewers can learn about the candidate as an "incidental byproduct" of being entertained. These interviews have the additional advantage of allowing candidates to speak in simpler terms, in an atmosphere—i.e., with music, energetic discussion, and a live studio audience—that is known to stimulate cognitive involvement and learning.9 There is even some limited empirical evidence to further support the idea that viewers pick up information about the candidates by watching their interviews. For example, Paul Brewer and Xiaoxia Cao show that exposure to a candidate appearance on a late night or comedy program during the 2004 Democratic primary was "positively related to knowledge about the campaign."10 Still, additional analysis is required to determine the extent to which these learning effects are widespread and the factors that make them most likely to occur. Of course, what viewers learn will be a reflection of what is discussed. Interviews focused on personal anecdotes will obviously help viewers understand the candidate better, while those focused on policy ought to inform them about where the candidate stands on the issues. It follows then that interviews combining personal and policy information would maximize the 132 Talk Show Campaigns Getting the Word Out and Going for Votes 133 breadth of what viewers learn about the candidate. It also stands to reason that viewers' political interest will play an inverse role in the learning process such that those with the lowest levels of political interest will, if they are sufficiently engaged by the interview, be in a position to learn the most from information that is likely to be new to them.11 They simply have more to learn than those who have been following the campaign closely because highly interested viewers cannot "learn" something they already know. This means that candidates ought to get their messages out to voters, particularly those with limited campaign interest, by appearing on these shows. These interviews also have the potential to shape the criteria viewers use when assessing the candidate. That is, they can "prime" viewers to alter the relative weight they give to certain considerations. The basic idea is that exposure to a message will activate specific considerations, making them easily accessible and therefore more likely to be emphasized in subsequent judgments. For example, hearing a candidate talk about the outsourcing of jobs could prime viewers to place more weight on their assessment of his or her economic policies when deciding for whom to vote. However, priming is not a simple or automatic process in that the primed criteria has to be made easily accessible and deemed relevant to the evaluation. In fact, it is "something like a two-stage process" in which "the priming stimuli should influence the accessibility of some knowledge constructs more than others, but whether people use those primed constructs as evaluative criteria depends on the degree to which they are perceived as applicable to the judgment task."12 Priming is therefore predicated on the enhanced accessibility and saliency of certain criteria. Entertainment talk show interviews provide candidates with perhaps their best opportunity to promote and prime their personal empathy—i.e., their compassion and ability to connect with ordinary people—by routinely featuring a good deal of self-deprecating humor and personal stories.13 This ought to make perceptions of the candidates' "regular guy" image more accessible and perhaps even more relevant to viewers' evaluations. Although some studies show that late night television viewing in general has little impact on how viewers measure up candidates, work by Patricia Moy, Michael Xenos, and Verena Hess suggests that late night interviews can sometimes prime a candidate's empathy.14 They found that late night viewers were more likely than non-viewers to emphasize George W. Bush's "caring" disposition after his second 2000 interview with David Letter-man, although there was no evidence of similar priming effects for three other candidate interviews, highlighting the conditional nature of this type of priming. These interviews also have the potential to prime policy considerations. As shown in Chapter Four, hosts routinely ask candidates about their positions on the key issues of the day. Of course, some hosts will challenge them from time to time, but in most cases candidates are given the opportunity to lobby viewers without much of the harsh confrontation found on other programs. This amiable discussion ought to increase the accessibility of certain issues so that viewers who find them relevant can emphasize them in their evaluations. Evidence of issue priming has been found in numerous studies, including the Moy, Xenos, and Hess study mentioned above and an experimental study of John Kerry's 2004 appearance on The Late Show where viewers were primed to focus on economic and security matters at the expense of personal image considerations because the interview "made policy concerns . . . readily accessible to viewers who then elected to use them as the basis of their evaluation."15 Although research finds that candidate interviews on entertainment talk shows can prime personal and policy considerations, there is also a sense that these priming effects might be conditional. All else equal, the amount of discussion on particular topics ought to affect the type of priming that occurs. Interviews featuring abundant personal material should prime empathy, those featuring policy discussion should prime the issues most discussed, and those that mix content should prime both personal and policy considerations assuming that a "hydraulic effect" does not occur in which viewers feel compelled to focus on certain criteria over others.10 The general priming literature further suggests that viewers have to be convinced that the topics being discussed are pertinent to their evaluation and are, in fact, worthy of more weight than voters are currently giving them. The volume of discussion or explicit statements about the importance of certain considerations may help with this, although it can be difficult to increase the saliency of considerations that are already on voters' minds. For example, an interview that features an extended discussion of the Iraq War will only produce priming effects if viewers believe the issue is pertinent and deserving of more weight than they were planning to give to it. In the end, these interviews are likely to prime viewers to the extent that viewers find certain considerations easily accessible, salient, and deserving of extra attention. For these priming effects to be beneficial, candidates must also convince viewers that they are reasonably empathetic and earn their support on the issues. Interviews that prime negatively evaluated considerations will obviously have an adverse effect overall. Candidates who can win over viewers by recounting personal stories and elaborating on their proposals stand to gain the most from any priming effects that occur because viewers will be emphasizing positive criteria that will enhance their composite ratings.17 Viewers will be thinking about the candidate's favorable attributes when making their evaluation. This suggests that candidates, by appearing on entertainment television and talking about their personal lives and political objectives, have the opportunity to potentially focus viewer attention on things that could help later in the evaluation process. Entertainment talk show interviews should also help to increase the candidate's overall support. That is, viewers should feel more warmly towards the candidate and be more likely to vote for the candidate.18 It stands to reason that a friendly conversation in which the candidate is given an 134 Talk Show Campaigns uninterrupted opportunity to promote his or her candidacy ought to have some positive impact on viewers at home. These effects could also be conditional, however, in that interviews that mix personal and political information ought to be the most persuasive because those with little more than personal anecdotes, although they may increase warmth ratings, give viewers littie substantive reason for strengthening their vote intentions. Moreover, viewers who are accustomed to the normal energy and excitement of entertainment television may ignore candidates and hosts who mire themselves in extended policy discussions. Some candidates might see these interviews as an opportunity to explain their policies to ordinary voters, but they should remember that persuasion is predicated on the ability to keep viewers engaged with more than stump speech boilerplate. Interviews that balance iighthearted stories with substantive policy information ought to draw viewers in while also giving them multiple reasons to increase their support. The implication is that candidates might benefit most from appearances on shows with a reputation for mixing content (e.g., The Late Show) and/or by controlling the conversation as much as possible so as to cover both personal and political topics. Viewers' political interest and partisanship could also play a role in the persuasion process. Those with little interest in politics rarely encounter political information on their own and thus have relatively weak partisan preferences. Watching a candidate interview on entertainment television should therefore provide them with new information that stands a reasonably good chance of altering their opinions because "politically inattentive persons will often be unaware of the implications of the persuasive communication they encounter, and so often end up 'mistakenly' accepting them."15 Highly interested individuals, on the other hand, have well-formed belief systems that allow them to better resist "persuasive communications that are inconsistent with [their] political predispositions."20 This means that candidates appearing on entertainment talk shows should be more likely to persuade politically disinterested viewers, including those from the opposing party, than their highly interested counterparts. These predictions suggest that, under certain conditions, candidates ought to benefit from appearing on entertainment television. Their messages should get across to viewers, and there is a chance that their conversations might prime and/or persuade the millions who are watching at home. For their part, viewers should also benefit as they learn about the candidate and receive information that could be useful in making political decisions. All of this suggests that entertainment talk show interviews have the potential to be much more than merely entertaining. EXPERIMENTAL TESTS Extant research highlights the challenges of studying viewer reaction to entertainment talk show interviews. While experimental studies have Getting the Word Out and Going for Votes 135 provided tight exposure measures and control over a host of variables, they have been constrained by low levels of mundane reality, limited and non-representative samples (e.g., college sophomores), and a small number of cases that hampers the generalizability of their results.21 Meanwhile, survey-based studies have often included a slightly larger number of interviews and better samples, but they have struggled to provide airtight exposure measures that ensure that viewers have actually seen the interview being studied. In fact, some studies have had to rely on national survey data that treat specific interview "viewers" as anyone who has watched a particular type of entertainment talk show at least once in the past week—e.g., at least one episode of either The Late Show or The Tonight Show.12 To address these challenges and investigate the predictions made above, I conducted a series of online experiments during the 2008 presidential campaign. I started by creating a standard online survey template that measured demographics, political attitudes, and media use habits. Then, as the campaign unfolded, I captured video of six late night candidate interviews that seemed to represent a cross section of typical appearances—three each by Barack Obama and John McCain (described below). I then embedded each video into its own survey and added a few political knowledge questions based on the specific interview. Each survey, along with a nearly identical control group survey that did not include the video, was then sent within 24 hours of the interview's initial airing to Market Tools, an online survey-research firm. Market Tools then sent email invitations to its national sample for three days, and those willing to participate were randomly assigned to either the control or test condition. The online survey was designed with an initial question to weed out anyone who had already seen the interview. Participants then answered a few demographic questions, and while those in the control group simply continued with the rest of the survey, those in the test condition watched one of the six candidate interviews before answering some questions about it and completing the rest of the survey. To ensure their exposure to the interview, test condition participants had to confirm that they had watched the entire video and correctly answer questions about which candidate was in the interview, which host conducted the interview, and whether the interview took place face-to-face or via satellite. The survey software was programmed to immediately drop any participants who did not verify their exposure or got at least one of the exposure confirmation questions wrong. The entire study included 1,503 participants, with 761 answering the control condition surveys while the other 742 participants served in the test conditions.23 Participants came from all over the country and represented the typical late night viewing audience fairly well (see Chapter Three) with 50% of the sample being female, 88% white, and 37% college educated. In addition, the median age was 41, median household income was between $50,000 and $60,000, and 33% of the sample was Republican compared to 39% Democratic and 28% independent or other party. Additional analysis confirms that there were no statistically significant differences between 136 Talk Show Campaigns conditions or between any of the six independent studies on any of these measures. This approach exceeds other experimental studies in that it includes multiple interviews and, while the samples for each experiment are not overly large (approximately 250 participants for each of the six experiments), they are nationally representative. I also took steps to ensure an accurate exposure measure—something that has hampered past studies—and to control for a number of key variables such as show type (late night), candidates (Obama or McCain), and campaign year (2008) in order to isolate the impact that interview content has on viewers' reactions. Of course, the tradeoff is that I can only speculate about the impact that interviews on other types of shows (i.e., daytime), with other candidates during other campaigns, have on those who see them. The results in Table 5.1 show that the six interviews covered a range of content and tone. John McCain's April 1 interview with David Letterman was fairly serious for a late night talk show. Letterman and the presumptive Republican nominee exchanged some humorous jabs during the opening monologue, but they got down to business once the actual interview started. In fact, the amount of policy discussion (58% of comments dealt with policy) far exceeded any jokes (9%) or references to personal (1$%) or campaign (24%) topics. McCain started by defending his conservative principles and warning that Republicans needed to do more to cut government spending if they were to regain the trust of the American people. He then launched into a lengthy explanation of how he would rebuild the middle class, including proposals for increased mortgage interest deductions and refinancing options, job training for high-tech industries, and controlling the excesses on Wall Street. Although he claimed, "the fundamentals of the American economy are very strong," McCain conceded that "we have to take action to help people in danger of losing their home, losing the American dream, who are sitting at the kitchen table saying I may have to get another job, may have to dip into our savings." McCain may have started with humor, but his interview with Letterman ultimately ended up being one of the most detailed and policy-laden entertainment talk show interviews of the 2008 campaign. Barack Obama's April 21 visit to The Daily Show also favored political topics over personality. The interview occurred via satellite just before the Pennsylvania primary, prompting Obama and Stewart to discuss the race in detail before talking about some of its major issues. Although Stewart brought up the Reverend Jeremiah Wright controversy and wondered if Obama represented "too much change," the candidate nevertheless pushed forward with an idealistic vision for America's foreign and domestic policy. "Folks are a lot more concerned now about the economy. They're not talking as much about Iraq, but they see a connection between us spending ten billion dollars a month there [and] the lack of investment here at home . . . What they're hoping is for significant change in Washington," Getting the Word Out and Going for Votes 137 Table 5.1 Content and Tone of Experiment Interview's. Policy Personal Campaign Jokes McCain on The Late Show 58 18 24 9 Obama on The Daily Show 34 16 50 26 McCain on The Daily Show 36 28 36 29 Obama on Jimmy Kitnmel Live 7 78 15 31 McCain on The Tonight Show 16 52 32 33 Obama on The Late Show 44 40 16 21 Note: Entries arc the percentage of all references in the interview. Obama said. Although Stewart and Obama shared some laughs (26% of comments included a joke), most of their conversation focused on policy (34%) and the campaign (50%) with very little mention of Obama's personal life or background (16%). John McCain's visit to The Daily Show on May 7 featured more of a balance between humorous personal references and serious policy discussion. Stewart promised to be "more respectful"—a reference to their testy encounter a year earlier—but the two still traded barbs over McCain's age and what his secret service code name should be—"trailblazer," "McSteamy," "Grumpelstiltskin," or "jerk." They also joked about the Democratic primary candidates "beating each other over the head." "It's terrible, 1 hate to watch," McCain mocked. However, sprinkled throughout the humor and personal references were some serious questions about McCain's strategy for the Iraq War. The senator reminded Stewart that he had long criticized the Bush administration's handling of the war and that he recognized that withdrawing would be a long process that might include a "security arrangement" like those with South Korea, Kuwait, Japan, and Germany. In the end, McCain's interview was almost equally balanced between good natured ribbing (29% jokes), campaign talk (36%), personal references (28%), and policy debate (36%). Then, on June 15, Barack Obama appeared on Jimmy Kimmel Live for an extremely lighthearted and casual chat. The interview, which came just a week after Hillary Clinton dropped from the race, featured an upbeat Obama looking to retool his appeal for a general voting audience. Kimmel introduced the presumptive nominee as "number 23, point guard, Senator Barack Obama." They then talked about Father's Day and Obama's tendency to lose things before spending the better part of nine minutes chatting about basketball: Obama's love of the game, his plan to replace the White House bowling alley with a basketball court, and the pick-up games he plays on the road. Kimmel ended the interview by offering to serve as 138 Talk Show Campaigns Getting the Word Out and Going for Votes 139 Obama's running mate, leading the candidate to diplomatically explain, "I think obviously you got some popularity and you'd help bring in some votes, but frankly, I don't think you'd vet." The vast majority of this interview mentioned personal references (78%) with a healthy share of jokes (31%) throughout, while issues (7%) and the campaign (15%) were really only mentioned in passing. John McCain's August 25 interview on The Tonight Show was also quite friendly, although he managed to sneak in a few political references from time to time. With the race in a dead heat, McCain opened with some self-deprecating cracks about his age—"my Social Security number is eight"— and although he offered Leno a place on his ticket, the host kindly declined, noting, "I can make more doing a week in Vegas, but thanks." The two then chatted like old friends for the rest of the interview, focusing most of their attention on the running mate selection process, McCain's friendship with Joe Biden, and his respect for Hillary Clinton. The only really earnest exchange came when the senator asked for a minute to explain his home ownership gaffe.24 McCain mentioned his time as a prisoner of war in Vietnam and explained that his wife's good fortune and generosity had allowed them to buy multiple homes. Other than this self-described "moment of seriousness," the vast majority of John McCain's Tonight Show interview featured jokes (33%) and relaxed conversation about personal topics (52%) with only the occasional reference to issues (16%) and a few more to the campaign (32%). Finally, there was more of a balance between personality and politics when Barack Obama hit the Late Show stage on September 10. Obama started with a series of humble references to Sarah Paiin's growing popularity and jokes about how it was causing him to field offers from magazines like Popular Mechanics rather than Newsweek or Time. They're offering me "a centerfold with a wrench," he told Letterman. Obama then promoted his '"hope" and "change" message by claiming to represent "something fundamentally different" that could help the average American. Although there were jokes (21 %) and other personal references to his youth and family (40%) throughout the rest of the interview, a good portion also covered the campaign (16%) and the central issues of the day (44%), most notably the economy, the Iraq War and especially the fight against terrorism in Afghanistan. "I think that if we had stayed focused on Afghanistan, if we, instead of spending a trillion dollars in Iraq, had focused our energy on problems here at home |and] implemented the 9/11 commission report," Obama said, "then we . . . would be further along in making sure that America was safe." Just as McCain had done months earlier on The Daily Show, Obama was sure to mix humorous passages with serious policy considerations during his interview with David Letterman. THE POLITICAL IMPACT OF TALK SHOW INTERVIEWS i present the results from these experiments in three sections. I start with a focus on how these interviews affected viewers' knowledge about the candidate and his policies. I then look for priming effects to see if and when they altered the criteria viewers use in their candidate evaluations. Finally, I test for persuasion effects to see if viewers became more likely than non-viewers to support the candidate. The results section concludes with a brief discussion of how these effects relate to each other as part of the candidate's overall entertainment talk show strategy. Informing Participants in each experiment were asked a small number of questions to assess their knowledge of the candidates. Those participating in the Obama-Kimmel experiment were asked only two questions due to a lack of policy specifics in the interview—one about the effect that Hillary Clinton's candidacy had on Obama and another on Obama's tendency to misplace things. Those participating in the other experiments, however, were asked three questions that typically covered both personal and political information found in the interview. For example, those in the McCain-Leno experiment answered questions about his economic plan, his thoughts about drilling for oil, and the story about him possibly running with John Kerry in 2004, while Obama-Stewart participants were asked about what Obama thought was keeping the country from progressing, the name of his controversial pastor who had made anti-American remarks, and the fact that he was often criticized for not wearing an American flag lapel pin.2-1 Figure 5.1 shows the difference in the proportion of control and test condition participants who gave the correct answer for each of the 17 questions. The first thing to note is the consistency with which viewers outpaced non-viewers in answering these questions correctly. These differences are statistically significant in all but one case: the question on whether or not John McCain supported offshore drilling {"Drilling").-** Moreover, the proportion of viewers who knew the correct answer almost always exceeded the 50% mark, which demonstrates that the majority of those who watched a particular interview were "in the know." This provides clear and compelling evidence that these interviews can convey information about the candidates and where they stand on some critical issues. For example, those who watched Barack Obama's Late Show interview learned about what he would have done differendy after 9/11 if he had been president: his plan to maintain his friendship with Bill Clinton without offering him a cabinet position; and his positive thoughts about George W. Bush's efforts to deal with the AIDS epidemic in Africa. These learning effects span both "easy" and "hard" questions. On easy questions—where most of the control group answered correctly—those who saw the interview were even more likely to know the right answer. For example, whereas 65% of non-viewers correctly answered that John McCain's plan for fixing the economy included spending less on foreign oil, 91 % of those who watched his interview with Jay Leno got it right ("Fix Economy").27 On harder questions—where most of the control group 140 Talk Show Campaigns McCain on The Uue Show and Obama on The Daily Show 64 63 78 61 59 r~ 48 i 47 36 35 MB . Illlii..... H Announce Spending Economy Progress Rev. Wright Lape! Pin Run McCain on The Daily Show and Obama on Jimmy Kimmel Lire 88 77 si 1-! »RB ■i lottuILJl Lk.-i. Diflurra\ Iraq Critic Better Candidate 56 16 Misplace McCain on The Tonight Show and Obama on The Late Show 60 l-> n than they are already receiving. This provides candidates with an opportu- i nity to steer the conversation towards favorable criteria so that viewers will be more supportive of the candidate overall. ; Most importantly, these interviews provide candidates with what they most covet: the opportunity to win support. The ability to do this, hew-i ever, appears to be conditional in that mixed-content interviews do more to i generate favorable assessments than interviews emphasizing either personal i qualities or policy positions. Viewers may enjoy the lighthearted chats and i they may learn from the policy-focused discussions, but they react most : positively to the combination of personal and political topics. This, again, ' gives candidates an incentive to highlight both aspects of their candidacies whenever possible. : Of course, different interviews with different candidates on different shows could produce different results. It could be, for example, that viewers are less likely to pick up on the candidate's message on daytime talk shows, or that priming and persuasion effects are weaker when the candidate is struggling to gain traction. Only additional studies will tell. However, these I results show that learning effects are found on a wide range of late night i interviews and that both McCain and Obama, on different shows, had simi- i lar results in terms of priming and generating support. This suggests that content—something that has rarely been emphasized in past studies—plays 154 Talk Show Campaigns an important role in determining the impact that these interviews have on viewers. Consider, for example, the consistency between McCain's interview with Jon Stewart on May 7 and Obama's interview with David Letterman on September 10. Differences between the candidates, the timing, and the hosts were seemingly irrelevant compared to the fact that both interviews mixed personal and political content. These results also imply that viewers can gain a great deal from watching candidate interviews on entertainment television. Chances are they will learn something useful, particularly if they do not follow politics too closely, and the interview could provide them with some guidance on what to consider when evaluating the candidate and how to think about him or her overall. In short, these interviews could significantly help viewers participate more effectively in the electoral process by giving them information and guidance that they might not otherwise encounter. Some viewers may even find it more useful to get this information from an entertainment-based source as it makes them more attentive than they would be with a more formal news source. In fact, while they are likely to engage with candidates appearing on entertainment talk shows, they could tune out the same candidate being interviewed in a more formal setting.61 This is not to say that these interviews are a panacea for the challenges facing American electoral democracy, but they do provide at least some viewers with an ideal venue for getting the political information they need to participate in the process. This chapter provides new information on the effect that entertainment talk show interviews have on viewers, and while there is still more to learn, the results can help candidates and commentators move the conversation from speculation and conjecture to more widely verified empirical realities. The findings provide a better understanding of when the entertainment talk show strategy might work and when it might fail, and show that hitting the talk show stage can have a real impact on viewers' political decisions. Appendix 5.1 Knowledge Regressions Coefficient (Standard Error) Exp(B) Announce Run 3.674*** 39.427 (0.797) Spending 2.047*** 7.742 (0.498) Economy 1.520*** 4.571 (0.419) Progress 0.758* 2.135 (0.381) Rev. Wright 0.407* 1.503 (0.199) Lapel Pin 0.928*** 2.530 (0.349) 'iown Hall 3.j56/' 35.023 (0.867) Bush Difference 1.787*** 5.971 (0.672) Iraq Critic 0.663* 1.940 (0.319) Better Candidate 2.234* 9.340 (1.143) Misplace 2.053*** 7.788 (0.589) Kerry '04 2.932*** 18.772 (0.507) Fix Economy 2.441*** 11.485 (0.586) {Continued) Appendix i.l (Continued) Coefficient (Standard Error Exp(B) Drilling 1.052* 2.863 (0.545) September 11 th 0.740* 2.096 (0.308! Clinton Cabinet 0.493* 1.638 (0.261) Bush Africa 1.214*** 3.369 (0.354) Note: Entries are "condition" coefficients, standard errors in brackets, and logged odds from independent logits controlling for party ID, age, gender, race, education, income, and political interest. *** p < .01, ** p < .05, * p < A in two-tailed tests. Appendix 5.2 Priming Regressions Empathy Economy Iraq Terrorism McCain on The Late Show 0.038 0.332** -0.171 -0.030 (0.193) (0.163) (0.132) (0.153) Obama on The Daily Show 0.043 0.196 -0.042 -0.299** (0.168) (0.145) (0.121) (0.144) McCain on The Daily Show 0.351* -0.053 0.349*** -0.175 (0.209) (0.165) (0.131) (0.132) Obama on jimmy Kimmel Live 0.236* 0.137 -0.001 -0.033 (0.144) (0.189) (0.163) (0.200) McCain on The Tonight Shotv 0.053 0.212 -0.177 0.042 (0.193) (0.179) (0.156) (0.149) Obama on The Late Show 0.567*** -0.279* -0.228 0.362** (0.190) (0.165) (0.154) (0.178) Note: Entries are interaction coefficients ("condition" ;< "empathy" or issue) with standard errors in brackets from independent OLS regressions controlling for party TD, age, gender, race, education, and income. *** p < .01, p < ,05, * p < A in two-tailed rests. Appendix 5.3 Feeling Thermometer and Vote Intention Regressions Feeling Vote Thermometer Intentions McCain on The Late Show 0.621* 0.111 (0.334! (0.146) Obama on The Daily Show 0.427 0.024 (0.314) (0.147) McCain on The Daily Show 0.792** 0.492*** (0.333) (0.172) Obama on jimmy Kimmel Live 0.292 0.197 (0.357) (0.171) McCain on The Tonight Show 0.261 0!033 (0.317) (0.152) Obama on The Late Show 0.814** 0.468 (0.323) (0.151) Note: Encries are "condition" coefficients with standard errors in brackets from independent OLS regressions controlling for parry ID, age, gender, race, education, and income. *** p < .01, ** p < .05, * p < .1 in two-tailed tests. Getting the Word Out and Going for Votes 159 NOTES 1. See, e.g., Roger Simon, "Obama's Change vs. McCain's Fear," Chicago Sun-Times, August 25. 2008, 25. 2. See. e.g.. "People," St. Louis Post-Dispatch, April 3, 2008, A2; Bill O'Reilly and Tonya Reiman, "Body Language: John McCain, Billy Ray Cyrus, and Naomi Campbell," The O'Reilly Factor, June 23, 2008; David Jackson, "McCain to Pitch Environmental Policy in Oregon; Speech to Outline Cap-and-Trade Plan," USA Today, May 12 2008, 9A; Andrew Malcolm, "McCain, 71, Nods Off on 'Late Night with Conan'," Los Angeles Times.com, July 18, 2008, accessed August 14, 2013, http://latimesblogs.latimes.com/washington/2008/07/mccain-conan.html. 3. Katherine Q. Seeyle, "The View Couch Not So Cozy for McCain," The Caucus, September 12, 2008, accessed August 15, 2013, http://thecaucus. blogs.nytimes.com/2008/09/12/the-view-coLich-not-so-cozy-for-mccain/; Rebecca Traister. "The McCain's Should Spend Everyday with Rachael P»ay," Saion.com, September 22, 2008, accessed August 11, 2013, wwvv.saion. com/2008/09/22/rachael_ray_mccains/; Jason Szep, "McCain Faces Tough Questions on Abortion," Reuters, September 12, 2008, accessed August 11, 2013, Yvww.reuters.com/article/2008/09/12/us-usa-politics-niccain-women-idUSN1231594920080912; Robert Barnes and Michael D. Sheat, "McCain Gets an Earful on The View'," Washington Post, September 12,2008, accessed August 14, 2013, http://voices.washingtonpost.com/44/2008/09/mccain-gets-an-earful-on-the-v.html; "Cindy McCain Slams Women of The View," ABC News.com, September 15, 2008, accessed August 14, 2013, http://abcnews. go.com/Politics/video?id = 5800473; Tirdad Derakhshani, "Sideshow," The Philadelphia Inquirer, May 24, 2008, D03. 4. Jennifer Harper, "Late Night Laughs Give Serious Gift; Politicos Polish Regular-Guy Aura," Washington Times, May 27, 2004, A06; David Bander, "Road to White House Runs Through TV Land," The Virginian-Pilot, December 31, 2008, E5; Sean Scully, "Bush Courts Female Vote on Talk TV," Washington Times. September 22, 2000, A12. 5. See, e.g., Lloyd Grove, "October 25'\ 2011, Burbank, CA," Newsweek, November 14, 2011, 15. 6. See, e.g., Matthew A. Baum, "Talking the Vote: Why Presidential Candidates Hit the Talk Show Circuit," American Journal of Political Science 49-2 (2005): 213-34; Patricia Moy, Michael A. Xenos, and Verena K. Hess, "Priming Effects of Late-Night Comedy," International Journal of Public Opinion Research 18-2 (2005): 111-31; Michael Parkin, "Taking Late Night Comedy Seriously: How Candidate Appearances on Late Night Television Can Engage Viewers," Political Research Quarterly 63-3 (2010): 3-15. 7. Markus Prior, "Any Good News in Soft News? The Impact of Soft News Preference on Political Knowledge," Political Communication 20 (2003): 149-71; also see Young Min Baek and Magdalena E. Wojcieszak, "Don't Expect Too Much! Learning from Late-Nigln Comedy and Knowledge Item Difficulty," Communication Research 36-6 (2009): 783-809; Marshall Sella, "The Stiff Guy vs. The Dumb Guy," New York Times Magazine, September 24, 2000, 72-80, 102; Julie Mason, "Kerry Makes Rounds to Aid Image; More Candidates Seeking Advantages of Popular Shows," Houston Chronicle, September 22, 2004, A14. For an overview of the debate, see Paul Brewer and Xiaoxia Cao, "Late Night Comedy Television Shows as News Sources: What the Polls Say," in Laughing Matters: Humor and American Politics in the Media Age, eds. jody C. Baumgartner and Jonathan S. Morris (New York: Routledge, 2008), 263-79. 160 Talk Show Campaigns Getting the Word Out and Going for Votes 161 8. Matthew A. Baum and Angela S. Jamison, "The Oprah Effect: How Soft News Helps Inattentive Citizens Vote Consistently," Journal of Politics 68-1 (2006): 948; also see Matthew Baum, Soft News Goes to War: Public Opinion and American Foreign Policy in the New Media Age, (Princeton, NJ: Princeton University Press, 2003); Matthew Baum, "Soft News and Political Knowledge: Evidence of Absence or Absence of Evidence?" Political Communication 20 (2003): 173-90; Dannagal Goldthwaite Young, "Late Night Comedy in Election 2000: Its Influence on Candidates Trait Ratings the Moderating Effects of Political Knowledge and Partisanship," Journal of Broadcasting and Electronic Media 48 (2004): 1-22; Amy B. Becker, "What About Those Interviews? The Impact of Exposure to Political Comedy and Cable News on Factual Recall and Anticipated Political Expression," International Journal of Public Opinion Research 25-3 (2013): 344-56. For an overview of the debate, see Patricia Moy, "The Political Effects of Late Night Comedy and Talk Shows," in Laughing Matters: Humor and American Politics in the Media Age, eds. jody C. Baumgartner and Jonathan S. Morris (New York: Routlcdge, 2008), 295-313. 9. See, e.g., Doris A. Graber, "Seeing is Remembering," Journal of Communication 40-3 (1990): 134-55; Doris A Graber, Processing Politics: Learning from Television in an Internet Age (Chicago: University of Chicago Press, 2001); James N. Druckman, "The Power of Television Images: The First Kennedy-Nixon Debate Revisited," Journal of Politics 65-2 (2003): 559-71. 10. Paul R. Brewer and Xiaoxia Cao, "Candidate Appearances on Soft News Shows and Public Knowledge about Primary Campaigns," Journal of Broadcasting and Electronic Media 50 (2006): 18-35. Also see Parkin, "Taking Late Night Comedy Seriously," 3-15. 11. This logic stems from Baum, "Talking the Vote," 213-34; and John R. Zaller, The Nature and Origins of Mass Opinion (New York: Cambridge University Press, 1992). Also see Xiaoxia Cao, "Political Comedy Shows and Knowledge about Primary Campaigns: The Moderating Effects of Age and Education," Mass Communication and Society, 11-1 (2008): 43-61; Barry A. Hollander, "Late-Night Learning: Do Entertainment Programs Increase Political Campaign Knowledge for Young Viewers?" Journal of Broadcasting and Electronic Media 49-4 (2005): 402-15; Michael A. Xenos and Amy B. Becker, "Moments of Zen: Effects of The Daily Show on Information Seeking and Political Learning," Political Communication 26-3 (2009): 317-32. 12. Scott L. Althaus and Young Mie Kim, "Priming Effects in Complex Information Environments; Reassessing the Impact of News Discourse on Presidential Approval," journal of Politics 68-4 (2006): 960-76; and Joanne M. Miller and Jon A. Krosnick, "News Media Impact on the Ingredients of Presidential Evaluations: Politically Knowledgeable Citizens are Guided by a Trusted Source," American Journal of Political Science 44 (2000): 295-309. 13. This conceptualization of "empathy" comes from Donald R. Kinder, "Presidential Character Revisited" in Political Cognition: The 19th Annual Carnegie Symposium on Cognition, eds. Richard R. Lau and David O. Sears (Hillsdale, NJ: Lawrence Erlbaum, 1986), 233-55 and Carolyn L. Funk, "Implications of Political Expertise in Candidate Trait Evaluations," Political Research Quarterly 50 (1997): 675-97. 14. General viewing studies include Young, "Late Night Comedy in Election 2000," 1-22; Dannagal Goldthwaite Young, "Late-Night Comedy and the Salience of the Candidates' Caricatured Traits in the 2000 Election," Mass Communication and Society, 9-3 (2006): 339-66. Moy, Xenos, and Hess, "Priming Fjfects of Late-Night Comedy," 111-31. 15. For an overview see James N. Druckman, Lawrence R. Jacobs, and Eric Ostermeier, "Candidate Strategies to Prime Issues and Image," Journal of Politics 66-4 (2004): 1180-202. Also see Parkin, "Taking Late Night Comedy Seriously," 9. 16. Joanne M. Miller and Jon A. Krosnick, "News Media Impacr on the Ingredients of Presidential Evaluations: A Program of Research on the Priming Hypothesis," in Political Persuasion and Attitude Change, eds. Diana C. Mutz, Paul M, Sniderman, and Richard A, Brody (Ann Arbor: University of Michigan Press, 1996), 79-100. 17. Druckman, Jacobs, and Ostermeier. "Candidate Strategies to Prime Issues and Image," 1180-202. 18. Early reports found connections between political attitudes and entertainment talk show viewing in general. See Michael Pfau and William P. Eveland, Jr., "Influence of Traditional and Non-Traditional News Media in the 1992 Election Campaign," Western Journal of Communications 60-3 (1996): 214-32; Michael Pfau, Jaeho Cho, and Kirsten Chong, "Communication Forms in U.S. Presidential Campaigns: Influences on Candidate Perceptions and the Democratic Process," The Harvard International Journal of Press/ Politics 6 (2001): 88-105. Also see Jonathan S. Morris, "The Daily Show with Jon Stewart and Audience Attitude Change during the 2004 Party Conventions," Political Behavior 31-1 (2009): 79-102. 19. Zaller, The Nature and Origins of Mass Opinion, 44-5. 20. Zaller, The Nature and Origins of Mass Opinion, 44. Also see Baum, "Talking the Vote," 228 and Heather L. LaMarre and Whitney Walther, "Ability Matters: Testing the Differential Effects of Political News and Late-Night Political Comedy on Cognitive Responses and the Role of Ability in Micro-Level Opinion Formation," international Journal of Public Opinion Research 25-3 (2013): 303-22. 21. For an example of these issues, see Parkin, "Taking Late Night Comedy Seriously," 3-15. 22. Examples include Baum, "Talking the Vote," 213-34 and Moy, Xenos, and Hess, "Priming Effects of Late-Night Comedy," 111-31. The problem is not so much with viewers remembering what they watched (see Susanna Dilli-plane, Seth K. Goldman, and Diana C. Mutz, "Televised Exposure to Politics: New Measures for a Fragmented Media Environment," American Journal of Political Science 57-1 (2013): 236-48) but with the probabilistic measure of exposure based on watching multiple shows over a lengthy period. 23. The sample breaks down in the following way: McCain-Letterman control group has 128 participants and the test condition has 126 (254); Obama-Stewart control is 126 and test is 125 (251); McCain-Stewart control is 130 and test is 124 (254); Obama-Kimmel control is 119 and test is 122 (241); McCain-Leno control is 128 and test is 118 (246); and Obama-Letterman control is 130 and test is 127 (257), 24. Jonathan Martin and Mike Allen, "McCain Unsure How Many Houses he Owns," Politico.com, August 21, 2008, accessed November 16, 2012 www. politico.com/news/stories/0808/12685.html. 25. The specific questions wete as follows with correct answers in italics. For McCain-Letterman: Announce Run: "On which of the following shows did John McCain announce his intention to run in the 2008 presidential election? The Daily Show with Jon Stewart, Good Morning America, The Late Show with David Letterman, Face the Nation, don't know"; Spending: "Accotding to John McCain, which of the following has hurt the Republican party most in recenr years? The Republican position on the environment, the Republican position on spending, the Republican position on the economy, don't know"; and Economy: "John McCain believes that the fundamentals of the American economy have_during the recession. Been damaged beyond repair, weakened slightly, remained strong, don't know." For 162 Talk Show Campaigns Getting the Word Out and Going for Votes 163 Obama-Stewart: Progress: "According to Barack Obama, which of the following is most responsible for keeping Americans from believing that progress can be made? A belief that the economy is in ruins, a belief that the country is divided between 'red' and 'blue' states, a belief that the country will never win in Iraq, don't know"; Rev. "Wright: "Who was Obama's former pastor in Chicago who stirred controversy with critical statements about race and the American government? The Reverend Al Sharpton, The Reverend Jeremiah Wright, The Reverend Jerry Falwell, don't know""; and Lapel Pin: "Which of the following was a criticism repeatedly leveled against Barack Obama during the Democratic primary season? He did not do well in law school, he did not attend church with his family, he did not wear an American flag lapel pin, don't know." For McCain-Stewart: Town Hall: "According to John McCain, which of the following helped him the most in winning the Repub lican nomination? His position on immigration, his history as a war veteran. his willingness to conduct town hall meetings, don't know": Bush Difference: "John McCain claims to be most different from George W. Bush on which of the following issues? Immigration, taxes, climate change, don't know"; and Iraq Critic: "John McCain says that he repeatedly _ the Bush administration's handling of the early part of the Iraq War. Praised, Defended, Criticized, don 'tknow." For Obama-Kimmel: Better Candidate: "Accordingto Barack Obama, running against Hillary Clinton in the primary h as made him... a better candidate, a weaker candidate, don't know"; Misplace: "Which of the following does Barack Obama apparently do quite often? Misplace things, sleep in, bowl, don't know." For McCain-Leno: Kerry '04: "Who said, in 2004, that John McCain should consider running as John Kerry's Vice Presidential candidate? Joe Lieberman, foe Biden, Ted Kennedy, Mitt Romney, don't know"; Fix Economy: "According to John McCain, the United States can solve some of its economic problems by doing what? Increasing taxes, spending less on foreign oil, raising the minimum wage, don't know"; and Drilling: "True or false, John McCain supports offshore drilling for oil? True, false, don't know." For Obama-Letterman: September 11th: "Barack Obama says that he would have done which of the following after 9/11 if he were president. Focused U.S. efforts on Afghanistan, focused U.S. efforts on Iraq instead of Afghanistan, focused U.S. efforts on Iraq and Afghanistan simultaneously, focused U.S. efforts on neither Iraq nor Afghanistan, don't know"; Clinton Cabinet: "True or false, Barack Obama has said that, if he becomes president, he will ask Bill Clinton to serve in his cabinet? True, False, don't know"; and Bush Africa: "Barack Obama has praised the way in which George W. Bush has handled which of the following? Hurricane Katrina relief efforts, the African AIDS crisis, the U.S. mortgage crisis, all of the above, none of the above, don't know." 26. All differences are statistically significant at the .1 level or better in two-tailed difference of means tests, except for the difference on "drilling" (79 to 86, t244 = 0.619, p .536). 27. Viewers are significantly more knowledgeable than non-viewers (0.91 to 0.65, t244 = 3,677, p .000! on the "fix economy" question. 28. Viewers are significantly more knowledgeable than non-viewers (0.63 to 0.17, t244 = 7.637, p .000) on the "spending" question. 29. Partisanship is measured on a five-point scale from "strong Republican" (1) to "strong Democrat" (5); age is in years; gender is female (1) or male (0); race is white (1) or other (0); education is the highest level of education completed from grade school (1) to graduate or professional degree (5); income is the participant's approximate annual household income in $10,000 categories from $0-10,000 (1) to $210,000,- (22); and political interest ranges from "not at all" interested (1) to "very" interested (4). 30. The percentage change in odds would be 94% given the formula: 100 " (Exp (b) - 1). 31. Prior, "Any Good News in Soft News?" 149-71. 32. Knowledge differences between low-interest viewers and low-interest non-viewers compared to high-interest viewers and high-interest non-viewers are statistically significant at the .1 level in two-tailed difference of means tests for all questions except: Better Candidate (8.0 difference between low-interest viewers and low-interest non-viewers compared to 8.6 difference between high-interest viewers and high-interest non-viewers, t239 = 0.424, p .671); Kerry '04 (52.5 to 53.3, t,44 = 0.057, p .944); Clinton Cabinet (4.0 to 3.0, t,5, = 0.707, p .479); Rev. Wright (8.0 to 7.1, t24s = 0.636, p .542); Lapel Pin (21.0 to 20.2, t,49 = 0.057, p .932): and Announce Run (37.0 to 36.1, tia = 0.636, p .525! ~ 33. The three questions on which low- and high-interest viewers are close are: Bettei Candidate (84.6 to 89.8. t = 0.368, p .713); Clinton Cabinet (54.1 to 55.7, t12. = 0.103, p .910); and Iraq Critic (53.6 to 56.8, t,„ = 0.226, p. 821). 34. See, e.g., Moy, Xenos, and Hess, "Priining Effects of Late-Night Comedy," 201. 35. See, e.g., Lydia Saad, "Economy Reigns Supreme for Voters: More than Half Rate it Extremely Important to the Vote for President," Callup.com, October 29, 2008, accessed April 2, 2012, www.gallup.com/poll/111586/ economy-reigns-supreme-voters.aspx. 36. See, e.g., Funk, "Implications of Political Expertise in Candidate Trait Evaluations," 675-97. The Cronbach's alpha for McCain's empathy score is 0.873 and 0.895 for Obama's empathy score across the six experiments. 37. Although viewers rated McCain's empathy highly, the difference between non-viewers' and viewers' empathy rating is not statistically significant (2.6 to 2.9, t,,2 = 0.405, p .686). 38. Viewersrated Obama's empathy highly, and the difference between non-viewers' and viewers' empathy rating is statistically significant (2.9 to 3.2, t,„ = 3.375,p.001). 39. Although viewers rated Obama's empathy highly, the difference between non-viewers' and viewers' empathy rating is not statistically significant (2.7 to 2.8, t25J = 0.303, p .762). 40. Frank Newport, "Obama Has Edge on Key Election Issues," Gallup.com, June 24, 2008, accessed November 12, 2012, www.gallup.com/poll/108331/ obama-has-edge-key-election-issues.aspx; Jeffrey M. Jones, "Obama Holds Lead over McCain on Top Issue of Economy," Gallup.com, August 25, 2008, accessed April 2, 2012, www.galhip.com/poll/109786/obama-holds-lead-over-mccain-top-issue-economy.aspx. 41. Data from the February 8—10 Gallup survey comes from Lydia Saad, "Obama Wins the Economy, McCain on Terrorism," Gallup.com, October 14. 2008, accessed November 12, 2012, www.gallup.com/poll/111130/ obama-wins-economy-mccain-terrorism.aspx. Also see "Clinton Backers Cooi to Obama—White Female Support in Question," Pew Research Center, May 29, 2008, accessed November 12, 2012, www.people-press.org/files/ legacy-pdf/425.pdf. 42. Race relations and affirmative action accounted for the other 50% of policy mentions. 43. Data from the February 8-10 Gallup survey comes from Saad, "Obama Wins the Economy," Gallup.com. Also see "Clinton Backers Cool to Obama." 64 Talk Show Campaigns Getting the Word Out and Going for Votes 165 44. Data from the August 21-23 Gallup survey conies from Saad, "Obama Wins the Economy." 45. Frank Newport, "On Economy, McCain Gains Ground on Obama," Gallup, com, September 10,2008, accessed November 12, 2012, www.gallup.com/ poll/110170/economy-mccain-gains-ground-obama.aspx. 46. The difference between viewers and non-viewers on the perceived ability to handle terrorism is statistically significant (3.1 to 2.7, t2J5 = 1.923, p .056). 47. Differences between viewers and non-viewers on thermometer scores are statistically significant (5.3 to 4.5, t, = 1.920, p .056) but not quite significant on vote intentions (2.7 to 2.4, t244 = 1.437, p .152). 48. Differences between viewers and non-viewers on rhermometer scores are not significant (5.8 to 5.5, t,49 = 0.841, p .401), nor are the differences on vote intentions (3.139 to 3.136, t2J4 = 0.014, p .988). 49. Differences between viewers on non-viewers on thermometer scores are significant (5.8 to 5.1, t2i, = 1.882, p .061), as are differences on vote intentions (3.4 to 3.0 t252 = 2:213, p .028). 50. Differences between viewers and non-Wewers on Obama thermometer scores are not quite statistically significant (6.4 to 5.9, t23s = 1.387, p .166), while differences between viewers and non-viewers on McCain thermometer scores are clearly insignificant (6.0 to 5.9, t,44 = 0.277, p .782). Differences between viewers and non-viewers on vote intentions are also insignificant for Obama (3.4 to 3.1, t2„ = 1.06, p .289) and for McCain (2.81 to 2.79, t244 = 0.323, p .747). 51. Differences between viewers and non-viewers on thermometer scores are significant (5.7 to 5.1, t,„ = 2.121, p .035), as are differences on vote intentions (3.3 to 2.9, t25J = i:i789, p .075). 52. Baum, "Talking the Vote," 227-8. 53. Differences between high and low-interest McCain-Letterman viewers on rhermomerer scores are statistically significant (5.0 to 4.1, t = 2.304, p .022). Differences between high-interest Democratic viewers and low-interest Democratic viewers on thermometer scores are statistically significant (9.1 to 8.2, t118 = 2.305, p .023), as are differences between high-interest Republican viewers and low-interest Republican viewers on thermometer scores (3.9 to 2.7, tm = 3.073, p .003). Independents are excluded from the party/interest analyses, 54. Differences between low-interest viewers and low-interest non-viewers on thermometer scores are statistically significant (5.4 to 4.4, t„, = 2.561, p .011), as are differences between low-interest viewers and low-interest non-viewers on vote intentions (2.5 to 3.2, t2., = 1.793, p .074), Differences between Democratic viewers and Democratic non-viewers on feeling ther mometer scores and vote intentions are statistically significant (3.2 to 4.0. tI2, = 2.048, p .042 and 1.7 to 1.2, t12, = 1.777, p .079), as are differences between Republican viewers and Republican non-viewers (8.3 to 6.9, t]2l _= 1.793, p .075 and 4.7 to 3.7, tm = 1.768, p .079). 55. Differences between low-interest Democratic viewers and low-interest Democratic non-viewers on thermometer scores are statistically" significant (4.4 to 3.2, t = 2.121, p .038), as are differences between high-interest Republican viewers and high-interest Republican non-viewers (7.8 to 6.7, t54 = 1.944, p .058). 56. Differences between low-interest Democratic viewers and low-interest Democratic non-viewers on vote intentions are statistically significant (2.6 to 1.8, toi = 2.828, p .003), as are differences between high-interest Republican viewers and high-interest Republican non-viewers (4.3 to 3.7, t = 2.121, p .039). 57. Differences between low-interest viewers and low-interest non-viewers on thermometer scores and vote intentions are statisticallv significant (5.9 ro 4.7, tJ2s = 4.242, p .000 and 3.5 ro 2.9, t,2| = 2.121, p .036). 58. Differences between Democratic viewers and Democratic non-viewers on thermometer scores and vote intentions are statistically significant (8.2 to 7.2, t126 = 1.768, p .079 and 4.5 to 3.9, t126 = 2.121, p .036), as are differences between Republican viewers and Republican non-viewers (4.1 to 3.2, t]20 = 3.182, p .002 and 2.2 to 1.6, t,,0 = 2.121, p .036). Differences between low-interest Republican viewers and low-interest Republican non-viewers on thermometer scores and vote intentions are statisticallv significant (4.8 to 3.3, t = 1.768, p .082 and 2.4 to 1.7, tM = 2.475, p .016). 59. Details are available from the author. 60. See Vladimer O. Key, Jr. The Responsible Electorate: The Rationality of Presidential Voting, 1936-1960 (Cambridge, MA: Harvard University Press, 1966). 61. See, e.g., Baum and Jamison, "The Oprah Effect," 946-959; Parkin "Taking Late Night Comedy Seriously," 3-15.