12 Chapter 1 passing? More information would be gleaned from counting the number of policy (issue) and character (image) utterances that were contained in each spot. Fourth, although positive and negative ads may well predominate, some political television spots use defenses, or explicit responses to prior attacks from opponents (Trent and Friedenberg, 1995, acknowledge that such spots exist, but do not study them). Past research (e.g., Benoit, 2007a) found defenses in political TV spots. Although defenses are not nearly as common in television spots as acclaims and attacks (they accounted for 1 percent of the utterances in those spots), they are in option that is used in campaign discourse that should not be ignored by critics and analysts. Fifth, the sample of spots gathered for this study includes multiple commercials from both major party candidates in every presidential campaign that employed television spots, from 1952 to 2012 (as well as ads from the primary campaign, from 'I third parry candidates, nonpresidential candidates, and candidates for prime minister, chancellor, or president in other countries). This will permit an unparalleled description of political television spots. In this study I will address several ropics. Chapters 3-6 will present the results of analysis of general election campaign spots, considered four campaigns at a time (I discuss four campaigns in each chapter as a compromise: I did not want to devote a chapter to each of the sixteen presidential campaigns to use TV spots, but I also did not want to lump together all general spots in a single chapter). My compromise was 1 to discuss four general election presidential campaigns in each of four chapters. In t these chapters I begin with a brief background about the situation, the candidates, and their spots. I reproduce the Transcripts of several spots from these campaigns to try to give a flavor for each contest. In each of these chapters I take up four topics. First, I describe the functions of presidential television spots (acclaims, attacks, i defenses). Second, I consider these spots' treatment of policy and character topics. Third, I will discuss the subdivisions of policy (past deeds, future plans, general goals) and of character (personal qualities, leadership ability, ideals) as delineated in chapter 2. Then I offer chapters on third parry presidential ads, primary ads, non- j presidential ads, non-U.S. ads discussing the same basic ideas. :§ This analysis is followed by a comparative chapter. Chapter 11 discusses trends J in general television spots, compares primary with general campaigns, incumbenrs i| versus challengers, the advertisements of winners and losers, and a discussion the j source of utterances in spots. Chapter 12 discusses the development of several recur- j ring themes in presidential television advertising. Chapter 13 offers a discussion of | implications and conclusions derived from this study. -j .1 I 2 Method The Functional Approach to Political Advertising This chapter describes the Functional Theory of Political Campaign Discourse (Benoit, 2007a), which was used to produce much of the data discussed in this book (rwo other theories—Issue Ownership and Functional Federalism—will be discussed in chapter 9). Then I will describe the advantages inherent in this approach to analyzing political television spots. Finally, I will discuss the content analytic method used to generate these data. A functional analysis is especially appropriare for investigating political campaign advertisements because candidate srarements in these messages are intended as a means of accomplishing a goal: winning the election. Political campaign discourse is therefore inherently instrumental, or functional, in nature. Of course, some candidates campaign in order to espouse a particular poinr of view. This is presumably the case for some third parry candidates; it may also be the case for some of the candidates in the primary who have no realistic chance of winning. However, for those who do have a reasonable chance of winning—which at the presidential level in contemporary campaigns means the Republican or Democratic nominees—campaign messages function as the means to gaining votes and thus winning public office. ASSUMPTIONS OF FUNCTIONAL THEORY Functional Theory is based on six key Axioms. Each of these assumptions will be explicated here. Al. Voting is a comparative act. When voting, citizens face a relatively straightforward decision: For whom should I cast my vote? A vote is a choice between two (or more) compering candidates and 13 14 Chapter 2 Method 15 it clearly involves a comparison. No candidate For elective office should be expected to be completely without drawback; on the other hand, surely no candidate is utterly without redeeming qualities. A voter chooses between two or more candidates, and die candidate who appears most suitable (an the criteria are most important S to a given voter) will receive that person's vote (see Downs, 1957; Himmelweit, Humphreys, and Jaeger, 1985; Nie, Verba, and Petrocik, 1979; Pamper, 1975). A candidate does not need to win all votes to win the election; nor must a candidate f appear perfect to receive a citizens voter. All that is required is for a candidare to appear preferable to other candidates for the office for a majority (or plurality) of voters. This means that in any contested election, a citizens vote choice is funda- f mentally a comparative decision that one candidare appears preferable to the other candidate(s) on whatever criterion is most important to that voter. Use of the word "appears" acknowledges that a citizen's evaluation of a candidate is a perception; it is not possible to have objective knowledge of what a candidate would do if elected (policy) or what kind of person (character) he or she "really" is. Some voters, of course, may be sq certain that the candidate they prefer is better thar they consider this superiority to be a fact rather than a perception, but nevertheless their candidate preference is still a perception. Voters' candidate choices are best understood t as perceptions that they form on the basis of their own attitudes and values and the information they possess that appears relevant to them when they make their vote choice. This means that the ultimate goal sought by candidates, winning elections, is achieved by persuading enough voters to believe that he or she is the t better candidate in the race. As Popkin explains, "each campaign tries hard to make its side look better and the other side worse" (1994, p. 232). Therefore, political television spots have three basic functions: (1) acclaims, or utterances that enhance their own credentials as a desirable office-holder (positive utterances), (2) attacks, or comments that degrade rheir opponents credentials as a potential office-holder (negative utterances), and, if their opponent attacks them, (3) defenses, or remarks ( that respond to those attacks (rebuttals). J This idea that voting is a choice between competing candidates is becoming increas- S ingly important as political parties decline in influence. Popkin observed that "in an I environment of diminishing party loyalty, campaigns and candidates exert a greater I influence on voters than chey did in the elections of 1940 and 1948" (1994, p. 12; j see also Menefee-Libey, 2000; Wattenberg, 1991, 1998). In earlier contests, the party f nominee was selected at the convention. Patterson (2003) noted that in 1952 Estes | Kefauver won ! all but one of the twelve primaries he entered and was the clear favorite of rank-and- I file Democrats in the final Gallup Poll before the national nominating convention. it Nevertheless, the party's leaders chose Adlai Stevenson as the Democratic presidential f nominee, (pp. 145—146) :| We cannot know whether Kefauver could have defeated Eisenhower if he had 1 been the Democratic nominee in 1952 (that seems unlikely), but we know Ste- 1 venson lost. In 1968 only sixteen Republican and seventeen Democratic primaries were held (Crotty and Jackson, 1985). By 2012, primaries were scheduled across the country, although some occurred after the nominee had been determined through the current delegate count. The candidates do not officially become their party's nominee until after their party's national nominating conventions, but in recent campaigns we have known who would win the nomination well in advance of the conventions. The increasing prominence of primary contests has changed the nature of politics. One important consequence is an increased importance of individual candidates and their campaign advisors. Although many voters still cast their votes in the general election for whoever represents their polirical parry, the individual candidates, and their apparent preferability to voters, play increasingly important roles in election outcomes. Party loyalty is still important but has less influence on voting decisions today (see Benoit, 2007a; Menefee-Libey, 2000; Wattenberg, 1991, 1998); the individual candidates and their campaign messages are filling the void left by the diminishing role of party identification in vote choice. So, voters choose between the competing candidates, and an increasing number do not do so exclusively by party loyalty, but according to their perceptions of the candidates' preferability, impressions fostered by their TV spots and other messages. Furthermore, political party affiliation cannot help voters decide among the candidates from their own parry contesting the nomination: TV spots are an important source of information to support this decision. A2. Candidates must distinguish themselves fi-om opponents. The-idea that voting is a comparative act, in which the relative preferabiliry of the contendets determines vote choice, leads to the second assumption of Functional Theory: Candidates must appear different from one another. Voters cannor make a choice; they have no reason to prefer one candidate over another, if the candidates look exacdy the same an ever)' comparison. Candidates need not differ on every possible point of comparison; everyone wants to reduce crime, decrease inflation, and improve the economy. However, if the candidates agreed on every issue (and projected all of rhe same character traits) there would be no basis for preferring one candidate over another. This means that it is essential for candidates in contested races to offer some distinctions between themselves and their opponents. Candidates may attempt to differentiate themselves from opponenrs by discussing either policy (what they have done and/or will do if elected) or character (who they are). For example, in 2012, Obama's "Jobs" spot noted that "Our businesses have created almost 4.3 million new jobs over the last twenty-seven months," an example of a policy theme intended to improve perceptions of the president. In 2012, Romneys ad "Shame on You" said that Obama "attacked Hillary Clinton with vicious lies" during the 2008 Democratic primary. This statement illustrates a character theme designed to reduce his Democratic opponents desirability. Both of these statements 16 Chapter 2 Method 17 implicitly or explicitly draw a contrast between the candidates, providing voters a basis for choosing one as preferable to the other. Theories of candidate behavior developed in political science have made similar observations. As indicated above, candidates will usually adopt some of the same issue positions. For example. Page (1978) explained that Downss (1957) economic dieory of democracy calls for a candidate's policy stands to echo the policy preferences of the public, and many spatial models—especially those of the public opinion variety—predict that die midpoint of public opinion on issues has an important influence upon die stands that a candidate takes, (p. 29; see, e.g., Enelow and Hinich, 1984) Page offered evidence from the 1968 campaign that "Across a wide variety of issues, then, both Humphrey and Nixon rook positions which corresponded fairly closely with what the average American favored" (p. 47). However, he also found that both Humphrey and Nixon disagreedwith the mid-point of public opinion on 15 percent of the seventy-two issues he examined. Specifically, Democrat Humphrey took more liberal positions on some issues whereas Republican Nixon adopted more conservative stands on some issues. This result is, generally, what one would expect. Both candidates took similar issue positions on some issues, close to the majority of the public, but each candidate distinguished himself from the majority opinion on other issues, Humphrey (the Democrat) by moving to the left and Nixon (the Republican) to the right of the ideological spectrum. Page also suggested that in 1964 Goldwater may have been more of an ideologue who did not adapt to public opinion; this is nor a winning strategy. A3. Political campaign messages allow candidates to distinguish themselves. Once a candidate decides which distinctions between him- or herself and opponents to stress to voters, those points of difference must be conveyed to voters. Citizens must be aware of these differences before such distinctions can influence their candidate preferences. Citizens cannot and should nor depend solely on the news media to provide voters with information about the candidates' policy positions. News may inform voters, but it may not; research shows that the news concentrates most on the horse race (see, e.g., Benoit, Hemmer, and Stein, 2010; Benoit, Stein, and Hansen, 2005). Patterson and McClure (1976) reported that learning occurs from candidates' campaign messages: During die 1972 presidential campaign, people who were heavily exposed to political spots became more informed about die candidates' issue positions. . . . On every single issue emphasized in presidential commercials, persons widi high exposure to television advertising showed a greater increase in knowledge than persons with low exposure, (pp. 116-117) It is clear that election campaign messages, such as political TV spots, are an important source of political information. Rather than discuss policy in great detail (impossible in any event in a thirty-second spot) candidares might take a simplet and possibly less risky approach when ■s discussing issues is to focus on ends rather than means: "I favor a balanced budget [but I won't tell you whether I will increase taxes and/or reduce spending ro achieve it]." A certain amount of strategic ambiguity may be useful to political candidates; however, Alverez (1998) found that too much ambiguity is undesirable. Still campaign messages help candidates establish the distinctiveness among contenders that gives voters a basis for choosing one candidate over another. J A4. Candidates establishpreferability through acclaiming, attacking, and defending. i I Of course, it is not sufficient for candidares to be distinctive in their messages, > even on the issues that matter most to voters in that election year; a candidate must I appear different from his or her opponents in ways that most voters favor. For ex- ample, a candidate who declared that "I am the only candidate who will raise taxes 60 percent for everyone" would surely stand apart ftom opponents, but not in a way that is likely to attract many votes. So, candidates must appear different and better than opponents; conversely, one can portray opponents as different and worse. Pop-kin (1994) observed that "Somehow, candidates manage ro get a large proporrion of the citizenry sorted into opposing camps, each of which is convinced that the positions and interests of the other side add up to a less desirable package of benefits" (p. 8). Three kinds of srarements or funcrions of discourse are capable of helping a candidate appear preferable to opponents. Acclaims ' First, candidates may offer acclaims (Benoit, 1997), statements chat stress a can- didate's advantages or benefits. Such self-praise can address the candidates character or policy record and/or stands. In 2012, for example, Obama declared that Romney "would be so out of touch with the average person in this country" (Obama, "The f Question"), questioning the Republican nominees character. Candidates can also ac- I claim their policy accomplishments. In 2012, Romney told voters that "We cut our I spending. Our legislature was 85 percent Democrar and every one of rhe four years I was governor, we balanced the budget" ("Believe in Our Future"). It is clear that i. most voters would view this statement as acclaims, as a desirable accomplishment. I So, one way to increase the likelihood that voters will see a candidate as preferable I is for that candidate to produce campaign messages char acclaim, emphasizing the ■I candidate's desirable qualities. ,| Attacks : j Another way to increase one candidate's (net) favorability is to attack or criticize [ the opponent(s). Stressing an opponent's undesirable attributes or policy missreps 18 Chapter 2 Method 19 should reduce that opponent's desirability, particularly for voters who value the attribute or policy discussed in the attack. Because voters make a comparative judgment about which candidate is preferable (Axiom 1), a successful attack increases the attacker's net favorabiliry by reducing the desirability of an opponent. Obama's 2012 spot "Number One" criticized Romney: : When Micr Romney was governor, Massachusetts was number one. Number one in stare | debt. $18 billion dollars in debt. . . more debt per person than any other state in the \ country. At the same time, Massachusetts fell to forty-seventh in job creation.... One of the worst economic records in the country. First in debt. Forty-seventh in job creation. This advertisement illustrates several attacks, focused on policy generally and past deeds or record in office in particular. Of course, some candidates may be reluctant to attack opponents. Voters con- J sistendy report that they do not like mud-slinging (Merritt, 1984; Stewart, 1975) \ so some politicians may wish to avoid engaging in excessive character assassination. Candidates may refrain from attacking, attack less often, or even promise to eschew arracks because voters say they dislike mudslinging. However, attacks have the potential to reduce the preferability of an opponent, so candidates use this function in campaign their TV spots. Clearly, arracks are an option used strategically by political candidates with the potential to reduce the apparent preferably of opponents. ' Complaints about the level of negativity in political campaigns are fairly com- j mon (see, e.g., Ansolabehere and Iyengar, 1995; Jamieson, 1992a; Pfau and Kenski, " \ 1990). Kamber (1997), for example, notes that "previous eras saw severe personal • attack on political candidates, but they also saw detailed and sometimes inspiring I deliberation over the issues. Our present political discourse is nothing but spleen" I (p. 4). Of course, vicious attacks are uncalled for and false attacks are detrimental J to voters (Benoit, 2013d). Still, attacks can provide voters with useful information. | Kamber (1997) explained that ;j There is an argument to be made in defense of responsible negative advertisements. The ;| voters need to know die whole story, and solely posirive arguments do not provide ir. A 1 campaign is not going to willingly offer negative information about its own candidate, if and yet that is essential information for the voters to make an informed decision, (p. 7) : 1 So, accurate criticism of an opponent can be useful for voters who need to consider ;| both the pros and the cons of the candidates when making a vote choice. False at- | tacks, or attacks that are malicious in tone, are not justifiable (of course, false acclaims J are also wrong). But legitimate criticism is a form of attack that can help voters make an informed choice. Geer (2006) argues that negativity in political campaigns "creates a competitive dynamic that should yield a richer information environment than if candidates just talked about their own plans for government" (p. 13). We must realize that just because voters express distaste for attacks does not necessarily mean that attacking messages are never persuasive. Candidates use focus groups and public opinion polls to design messages—including attacking messages—and they obviously believe attacks can be persuasive. It seems clear that attacks are capable of reducing the desirability of the target of those attacks. However, some attacks may have a bacldash effect and thus hurt both the sponsor (because voters dislike mud-slinging) as well as the target. This means the most important question when deciding whether to attack may be who is likely to suffer the most from an attack: the target of the attack or the attacks sponsor? Meta-analyses have established that both positive and negative ads can be'persuasive (Allen and Burrell, 2002; Lau, Sigelman, and Rovner, 2007). The topic of the attack may be one important factor in audience response. Johnson-Carree and Copeland (1989) provide evidence that voters tend to consider policy attacks more acceprable than character attacks. Other studies (Pfau and Burgeon, 1989; Roddy and Garramone, 1988) indicate that policy attacks can be more persuasive than character artacks. Benoit (2003), analyzing multiple message forms (primary television spots, debates, and brochures; acceprance addresses; general television spots, debates, and brochures) over the last fifty years, found that winners are signiBcandy more likely to attack more on poliq', and less on character, than candidates who lose elections. Of course, this finding does not mean that policy attacks guarantee a win, or even that attacks on character can never be persuasive. It does suggest that it may be prudent to attack more on policy (and less on character) than one's opponent. Defense; The third function of campaign messages that is capable of affecting a candidates apparent preferability is defense. If a candidate is attacked by an opponent—or perhaps it would be more realistic ro say when one candidate is attacked by another—the recipient of the attack can choose to defend against (refute) that attack in a campaign message (see Bryant, 2004). Obama's spot "Blatant" defends against an attack from Romney: "Seen this? Mitt Romney claiming the President would end welfare reform's work requirements? The New York Times calls it 'blatandy false.' The Washington Past says, 'the Obama administration is not removing the bill's work requirements at all.'" This ad identifies a criticism and refutes it. Research has investigated the circumstances under which political advertising is likely to attack (see, e.g., Elmelund-Praestekaer, 2010; Sullivan and Sapir, 2012). Several potential factors have been identified including incumbency (challengers tend to attack more), standing in public opinion polls (those behind usually attack more than leaders), being attacked by opponents tends to provoke attacks in response, competitiveness of race (artacldng is positively related to competitiveness), and sponsor of advertisement (parties ate usually more negative than candidates). Defense can be important because a timely and appropriate defense may be able to prevent further damage from an attack and restore some or all of a candidate's damaged preferability. Defense, then, is the third potential function of campaign 20 Chapter 2 Method 21 discourse. It attempts to restore, or prevent additional damage to, a candidates perceived preferability. An interview with former presidential candidate Michael Duka- : Ids indicated that he believed defense could be important, explaining that "he was [. glad President Clinton was responding quickly to attacks, something Mr. Dukakis | said he failed to do in his 1988 campaign" (Clines, 1996, p. A12). Smith (1990) dis- j cussed two of these three functions when he explained that in politics "people pursue { and defend jobs by publicly boasting and attacking others" (p. 107). | At times candidates may decide to forgo defenses when they are attacked. Some I candidates may not wish to "dignify" an opponents accusations with a response. This I reluctance may also be related to the fact that defenses have three potential drawbacks. First, it is possible that presenting a response to an attack could make that can- ; didate sound defensive, appearing reactive rather than proactive. Candidates want to j project the image that they are in charge of events, not merely reacring to opponents. j Second, it seems likely that a candidate is most likely to attack on topics that favor j the attacker rather than the target of attack, which means that defending against an f attack probably rakes a candidate "off-message," devoting precious message rime to | issues that are probably better for one's opponent. Third, the only way to respond to ! a particular attack is to identify that criticism. Mentioning the arrack, in preparation J for refuting it, could inform or remind voters of the very weakness that the candidate is trying to combat. Defenses therefore have three porential drawbacks. Research has shown that candidates who are attacked more frequently in debates tend to defend more often than other candidates: Being the target of an attack provides both the opportunity and the motivation to defend (Benoit, 2007b). Scholars have offered other lists of functions. For example, Gronbeck (1978) identified a number of instrumental and consummatory functions of presidential t campaigning. Some of these sound like uses and gratifications for the audience. Of I course, it is important to know how auditors make use of campaign discourse. However, those sorts of functions supplement, rather than compete with, this analysis of campaign functions because I explicidy privilege the viewpoint of die candidates purposes rather than the voters' uses in this analysis. Certainly it is useful to consider the voters' petceptions, but Functional Theory is focused more on candidates. Similarly, Devlin (1986; 1987a) discusses several functions of political ads. However, I believe that these three functions (acclaims, attacks, defenses) are more basic than his list. For example, one of the functions Devlin lists is raising money. Candidates tout their desirable features (acclaim) and/or criticize their opponents (arrack) in order to convince donors to contribute. Another function identified by Devlin is reinforcing a candidates supporters. Supporrers are reinforced by stressing the good j qualities of the candidate (acclaims; and, quite possibly, by attacking or stressing the 1 negative qualities of the opponent). Thus, these three activities—attacking, acclaim- I ing, and defending—are the fundamental functions of political advertising. Sabaro j (1981) made a similar point, albeit from rhe voters' poinr of view, when he observed I that there are a limited number of ways to vote: "for or against either of the party j nominees or not voting at all" (p. 324). Scholars who investigate televised political 1 advertising often distinguish between positive and negative spots (see, e.g., Kaid and Johnston, 2001), which correspond to the functions of acclaims and attacks. Trent and Friedenberg (2000) noted that televised political advertisements can accomplish three basic functions: extol the candidates' own virtues; condemn, attack, and question their opponents; and respond to attacks or innuendos. These three functions obviously correspond to acclaims, attacks, and defenses. Pfau and Kenski (1990) noted that television spors can be categorized in four types: positive, negative, comparative (positive and negative elements together), and response (defense). I explicidy privilege the candidate's purposes in this analysis, rather than voters' or reporters' purposes. So, several political scholars have recognized that political campaign messages acclaim and attack role of defensive or response advertisements. Only research from the Functional perspective investigates the frequency of defense in campaign messages. Political candidates and their campaign advisors also recognize che fundamental principle that campaign discourse performs multiple functions. For example, H. R. Haldeman gave advice to President Richard M. Nixon on the 1972 reelection campaign: "Getting one of those 20 [percent] who is an undecided type to vote for you on the basis of your positive paints is much less likely than getting them to vote against McGovern by scaring them to death about McGovern" (Popkin et al., 1976, p. 794n). Thus, Haldeman argued that the election hinged on the undecided voters and that Nixon could seek their votes by praising himself—acclaiming Nixon's "positive points"—or by attacking his opponent—"scaring them to death about McGovern." Similarly, Vincent Breglio, who was a part of Ronald Reagan's successful 1980 presidential campaign, acknowledged that "It has become vital in campaigns today that you not only present all the reasons why people ought to vote for you, but you also have an obligation to present the reasons why they should not vote for the opponent" (1987, p. 34). So, political campaign advisors, like political communication scholars, tecognize that candidates can praise themselves and attack their opponents. This is why the Functional approach analyzes political campaign discourse into utterances that acclaim the preferred candidate, attack the opponent, and defend. the candidate from opponent's attacks. Although these thtee functions may not be equally common in discourse, they are three options that ever}' candidate has available for use. These functions ate very important because they provide vatets a reason to vote for a candidate or against an opponent. A complete understanding of political campaign communication should consider all three functions. One useful way to think about these three functions is as an informal form of cost-benefit analysis. Acclaims stress a candidates benefits. Attacks reveal an opponents costs. Defenses attempt to refute or minimize potential costs. A vote decision requires an understanding of the pros (acclaims) as well as the cons (attacks, defenses) of the contending candidates. This means that attacks serve a useful purpose— identifying costs—as long as they are neither false nor misleading. Political candidates can inform voters of an opponents potential costs through attacks. Consistent with this analysis Kelley and Miter (1974), using survey data from the 1952-1968 22 Chapter 2 presidential elections, found that 82-87 percent of citizens voted for the candidate for whom they reported the largest number of reasons for lilting that candidate and the smallest number of reasons for disliking that candidate (in other words, benefits J and costs; this figure may be less than 100 percent because some pros or cons are j more important to a particular voter than others). > It is important to acknowledge that characterizing vote choice as similar to f cost-benefit analysis does not mean that I assume that every voter takes a rational [ approach to voting: gathering, weighing, and integrating as much information as | possible ro guarantee that they make the most rational decision possible. As Zaller | (1992) correcdy explained, "citizens vary in their habitual attention to politics and i hence in their exposure to political information and argumentation in the media" J (p. 1), Only political junkies avidly seek out huge amounts of information about the [ various candidates. As Popkin argued (1994; see also Downs, 1957), many voters use I information shortcuts. They do not seek out Information about the candidates or j they wait until just before the election to do so. They base their voting decisions on j the information they happen to encounter, including TV" spots. This is why political j candidates employ multiple media and repeat their basic campaign message: They i want their message out there for whatever voters might be attending to a particular j medium at a given point in time. TV spots are particularly important because they have the potential to reach all voters, even those who might not watch debates or i read campaign news. Voters do not quantify bits of information or place the infor- j mation they obtain about the candidates into mathematical formulas (i.e., benefits minus costs) to calculate their votes. Thus, although I believe that deciding how to vote is similar in principle to cost-benefit analysis, I do not claim that votets use numbers to quantify pros and cons or even that they systematically and consciously weigh the pros and cons of competing candidates. Acclaims tend to increase a can-didace's perceived preferability, attacks tend to reduce an opponent's preferability, and defenses may resrore lost preferability. All three functions work to make one candidate appear preferable to an opponent. : We must realize that the power of campaign messages has limitations. As noted above, many votets have little interest in political campaigns and are unlikely to watch debates or to read or watch political news. Some voters who do pay attention to candidate messages may not accept a candidates statements at face value. Candidates may not always address the most prominent concerns of voters, and when that happens it surely would reduce the impact of the message. Different voters may interpret I a message in different ways, and their reaction may not be what the candidate hoped (Reinemann and Maurer, 2005, reported that acclaims in German political leaders debates generated general support in the audience whereas attacks tended to polarize j the audience). Furthetmore, we should not assume that a single message is capable of j making a voter choose the candidate touted in that message. Nevertheless, the messages : to which are exposed during a campaign gradually shape their perceptions of the can- J didates' character and issue stands and, ultimately, a citizen enters a polling place and ■ Method 23 casts a vote based on those perceptions. Undecided and independent voters, as well as potential vote defectors, may be particularly susceptible to these messages. Functional Theory argues that these three functions are likely to occur with different frequencies. Acclaims, if persuasive (if accepted by the audience) can increase a candidates apparent preferability and have no drawbacks. This means that acclaims should be the most common campaign discourse function. In contrast, attacks, if persuasive, can increase a candidates apparent net favorability by decreasing an opponent's preferability. However, many voters dislike mudslinging as noted above so the risk of bacldash may encourage candidates ro moderate their attaclts. Accordingly, Functional Theory expects attaclts to be less common than acclaims. Finally, defenses, if they are accepted by a voter, can help restore a candidates lost preferability. As noted above, defenses have three drawbacks: They are likely to take a candidate off-message (because attaclts are likely to concern the target candidate's weaknesses), they risk informing or reminding voters of a potential weakness (a candidate must identify an attack to refute it), and they may cteate the impression that the candidate is reactive (defensive) rather than proactive. Thus, Functional Theory makes two predictions about the functions of political campaign discourse: Hi. Candidates will use acclaims more frequently than attacks and attacks more often than defenses. Studies have investigated this prediction with a variety of American presidential campaign messages. Research on American presidential primary and general debates (1952-2004) confirmed that the most common function was acclaims (62 percent); nominating convention Acceptance Addresses from 1952-2004 also emphasized acclaims (77 percent), as do primary and general election direct mail brochures from 1948-2004 (77 percent; Benoit, 2007a). As ptedicted, attaclts were the second most common function in U.S. TV spots (34 percent), Acceptance Addresses (23 percent), and direct mail (23 percent; Benoit, 2007a). Defenses were the least common function in debates (5 percent), Acceptance Addresses (1 percent), and direct mail (0.3 percent; Benoit, 2007a). Subsequent chapters will offer daca that test this prediction in different lands of political TV spots (American presidential general, presidential primary, gubernarorial, senate, house, and local ads as well as election ads from other countries). A5. Campaign discourse occurs on two topics:policy and character. The fifth axiom of Functional Theory posits that political discourse can occur on rwo broad topics: policy (issues) and character (image). In other words, candidares try to persuade voters of their preferability on policy—what they do—and character— who they are. Pamper (1975), in fact, observed that many voters "change their partisan choice from one election to the next, and these changes are most closely related 24 Chapter 2 Method 25 to their positions on the issues and their assessment of the abilities of the candidates" (p. 10). FunctionaJ theory defines policy and character in this way: Policy utterances concern governmental action (past, current, or future) and problems amenable ro governmental action. Character utterances address characteristics, traits, abilities, or attributes of the candldares. Thus, these are the two broad topics on which candidates contend over their prefer-ability (Functional Theory also subdivides policy and character utterances into finer categories, as discussed later). Rountree (1995), for example, distinguishes between actus (behavior, action) or what we do and status (nature) or who we are in political campaign discourse. Although Functional Theory dichotomizes the two potential topics of political campaign discourse, it aclcnowledges that policy and character have a complex and dynamic relationship (Benoit, Blaney, and Pier, 1998). First, it is possible that an utterance which focuses explicidy on policy could have some influence perceptions of the candidates character. For example, this passage from Bill Clinton's 1996 Acceptance Address discusses his first term successes with the economy: Four paint four million Americans now living in a home of their own for the first time; hundreds af thousands of women have started their awn new businesses; more minorities own businesses rhan ever before; record numbers of new small businesses and exports. . . . We have the lowest combined rates of unemployment, inflation, and home mortgages in twenty-eight years. . . . Ten million new jobs, over half of them high-wage jobs, ten million workers getting the raise rhey deserve wirh the minimum wage law. Surely this is a policy uttetance, for it discusses home ownership, business ownership, exports, unemployment, inflation, mortgages, jobs, and the minimum wage. Of course, these successes all work to implicidy reinforce Clinrons apparent leadership ability, a character attribute, because they implicidy demonstrate that he possesses the skills necessary to enact legislation (leadership ability is one aspect of character). Similarly, a message that touted programs ro help the poor or disadvantaged could serve to create or reinforce an impression of that candidates compassion (another element of character). On the other hand, this passage from one of Vice President George Bush's 1988 Republican primary television spots recounted his experience in the military, focusing on his experience and courage: "How does one man come so far? Maybe for George Bush, it began when he became the youngest pilot in the Navy. Or perhaps it began this day in 1944 when he earned the Distinguished Flying Cross for bravery under fire." This passage clearly concerns Bush's character, the personal quality of bravery, not what he will do if elecred president. Nevertheless, vorers might reasonably infer that this land of person, a person with this land of character, is likely to support a strong military. These two ldnds of comments have distincdy different content. One passage (from Clinton) explicitly addresses policy and die other (from Bush) explicidy discusses character. These messages tell us more about Clinton's policies than Bush's policies; we can learn more from them abour Bush's personal qualities than Clintons. However, we should not be surprised if voters form impressions from these passages that ate not explicidy addressed in the text (see Hacker, Zakahi, Giles, and McQuirry, 2000). Furthermore, it appears that candidates sometimes attempt to shift the grounds of discussion from one topic to the other. For example, in the first Clinton/Dole debate of 1996, Jim Lehrer posed this question about Clinton's character "Mr. President, what do you say to Senator Dole's point that this election is about keeping one's word?" Clintons honesty (his character) was challenged, and he offered this answer: Lets look at that. When I ran for president I said we'd cut die deficit in half in four years; we cur it by 60 percent. I said that our economic plan would produce eight million jobs, we have ten and a half million new jobs. We're number one in autos again, record numbers of new small businesses. I said we'd put, pass a crime bill that wouldf put 100,000 police on die street, ban assault weapons, and deal with the problems that ought to be dealt with with capital punishment, including capital punishment for drug kingpins, and we did that. I said we would change the way welfare works, and even before the bill passed we'd moved nearly two million people from welfare to work, working with stares and communities. 1 said we'd get rougher wirh child support and child support enforcements up 50 percent. I said that I would work for tax relief for middle class Americans. The deficir was bigger dian I thought it was going to be. I think they're better off, all of us are, that we got the interest rates down and the deficit down. Clintons response shifted the discussion away from the question of honesty or keeping one's word generally to keeping one's word on campaign promises, or policy accomplishments: jobs, autos, crime, welfare, middle-class tax curs, interest rates, the deficit. He tesponded to an attack on character by shifting grounds and acclaiming his past successes on policy. This process can also work in the other direction, moving from policy to character. For instance in the second debate of 2000, Vice President Gore attacked Governor Bush's record in Texas on the issue of healrh care. GORE: I'm sorry to tell you diat, you know, there is n record here, and Texas ranks forty-nindl our of die fifty scares in health care—in children with health care, forty-ninth for women with health care, and fiftieth for families with health care. LEHRER: Governor, did Vice President—arc the vice president's figures correct about Texas? BUSH: You can quote till the numbers you want, but I'm telling you, we care about our people in Texas, we spend a lot of money to make sure people get health care in the state of Texas, and we're doing a better job than they are at the national level for reducing uninsured. LEHRER: Is he right? Are those numbers correct? Are his charges correct? 26 Chapter 2 Method 27 BUSH: If he's trying to allege that I'm a hard-hearted person and I don't care about children, he's absolutely wrong. We spend S4.7 billion a year in the state of Texas for uninsured people, and chey get health care, (emphasis added) Bush repeatedly tried to turn this policy question into an issue of character. Bush did talk some about spending in Texas on health care (policy), but there is a clear effort to shift this attack from policy to character: "we care about our people." Bush even responds to character attacks (that Bush is hard-hearted, that he doesn't care about children) that Gore never articulated: "If he's trying to allege that I'm a hard-hearted person and I don't care about children, he's absolutely wrong." Again, this is a clear effort to shift the topic from policy to characcer. Functional Theory predicts that, particularly in presidential campaigns, policy will be a more frequent topic of campaign messages than character. We elect presidents to run our government, to implement policy. Although some voters believe diat they elect positive role models—and surely we all hope our elected leaders are positive role models—the primary duty of our elected officials is to administer policy. Hofstetter (1976) explains that "issue preferences are key elements in the preferences of most, if not all, voters" (p. 77; see also Patterson and McClure, 1976). Furthermore, public opinion poll data from 1976 to 2000 reveals that the majority of voters believe that policy is more important than character in their vote for president (Benoit, 2003). i Presidential candidates who discuss policy more, and character less, than their opponents are more likely to win elections (Benoit, 2003). Character does matter, of course. We must trust candidates to work to achieve their campaign promises, and we must trust them to implement suitable policies in unexpected situations on which they did not take policy stands during the campaign. However, King (2002) summarized the results of several studies of the role of character in 51 elections held in 6 countries between I960 and 2001: It is quite unusual For leaders' and candidates' personality and other personal traits to determine election outcomes. ... [T]he almost universal belief that leaders' and candidates' personalities are almosc invariably hugely important in determining the outcomes of elections is simply wrong, (p, 216) Because of voter preferences, Functional Theory considers policy to be more important, in genetal, than character. Specifically, Functional Theory holds that candidates are likely to respond to these preferences so that policy will be discussed more frequendy in presidential campaign messages than character. Of course character is discussed in campaign messages. These considerations lead to a second prediction: H2. Policy comments will be more frequent than character comments in presidential campaign discourse. Published research has investigated the topics of presidential campaign messages (Benoit, 2007a). In American presidential primary debates, policy was 78 percent of statements whereas character was 28 percent. Presidential primary brochures discussed policy more than character (62 percent to 38 percent). In nominating convention Acceptances, policy was discussed more often than chatactet (55 percent to 45 percent). In general election debates, policy was a more frequent topic than character (75 percent to 25 percent). In direct mail advertising from the general election campaign, policy (76 percent) was addtessed more frequently than character (24 percent). As with the first hypothesis, subsequent chaptets will provide evidence on this prediction in a variety of political campaign advertisements. A6. A candidate must win a majority (or a plurality) of the votes cast in an election. The last axiom might appear to be so trivial that it is not worth mentioning. However, this proposition implies several key tenets of campaigning. First, candidates do not need to try to win every vote. This is extremely important because some policy positions are inherently divisive and will simultaneously attract some voters and repel odiers. That is, many issues dichotomize the electorate. For instance, in 2012 Barack Obama and Mitt Romney disagreed on such issues as how health care should be provided or federal tax policy. It is unrealistic to expect either candidate to win the votes of ever)' citizen given the existence of divisive issues such as this one. Luckily, however, candidates need not receive all of the votes that are cast to win the election. Second, it is important to realize that only those citizens who actually cast votes in the election matter to the outcome. This means that a candidate does not even have to win the votes of most citizens, but only of most citizens who actually vote on election day. Some candidates have explicitly attempted to encourage turnout, which seems to be consistent with the ideals of democracy. For example, in 1964 at least seventeen of Johnson's television spots included the statement "The stakes are too high for you to stay home." Thus, it should be possible to enhance a candidate's chances of winning by increasing the turnout of voters who favor that candidate (or, although diis seems reprehensible, reducing the turn-out of voters who favor an opponent; see Ansolabehere and Iyengar, 1995). Third, American presidential elections are peculiar because of the Electoral College and its rules. In a presidential election, a candidate only needs to persuade enough of those who are voting in enough states to win 270 electoral votes. This encourages candidates to maximize their resources by campaigning more vigorously in some states than others. The 2000 presidential election underlined the importance of the Electoral College vote. As voting returns came in on Tuesday night Florida was "given" to Gore, taken back, given to Bush, and then taken back again. Then the recounts in Florida made the nation wait for the winner to be determined as the outcome of the election hinged on whether Florida's twenty-five electoral votes belonged to Bush or Gore. The U.S. Supreme Court (in a five to four vote) decided to halt recounts in Florida, giving the Electoral College majority to Bush. Al Gore won the popular balloting by a margin of half a million votes, but because Bush won Florida 28 Chapter 2 Method 29 by 537 votes, he won all of its Electoral College votes and the ptesidency (New York Times, 2001), Thus, a U.S. presidential candidate only needs to win a majority of votes in enough states to amass 270 electoral votes to win the presidency, and that influences the placement of campaign discourse. These principles suggest six specific strategies candidates can use in an attempt to maximize the probability of winning the election. First, a candidate can attempt to increase the election day turnout of voters who prefer that candidate. If a citizen fails to vote, it does not matter which candidate that person prefers. This means that if the same number of people prefer the two leading candidates, but more of one candidates supporters actually vote, that candidate will win the election (indeed, a candidate with less support than a tival coulc win if his or her supporters vote at a sufficiendy higher rate than the other candidates adherents). Second, a candidate can seek the support of undecided voters. The numbet of independent voters has increased over time as the importance of political parties has diminished. Although there are some vote defectors, most Republicans will vote for the Republican nominee and most Democrats will vote for the Democratic nominee. Thus, a wise candidate will focus much of the general election campaign on the undecided voters. In 1996, for example, we heard a great deal about the so-called "soccer moms," swing voters who allegedly held the keys to the 'White House. Independents are less likely to vote than partisans; still, the difference between the number of Republicans and Democrats is so small, and the number of Independents is so large, that Independents are important even if a smaller percentage of Independents than partisans actually vote. Third, a candidate can attempt to attract potential vote-defectors from the other political party. Candidates are unlikely to attract votes from those partisans who are strongly committed to the other political party, but some party members are willing to vote for the candidate of the other party (Nie, Verba, and Petrocik, 1999)—if they are given an adequate reason to do so in the candidates' campaign messages. This is a surprisingly large group, ranging from 14—27 percent (Nie, Verba, and Petrocik, 1999). Thus, political candidates can try to poach voters who have only soft support for their opponents. Fourth, a candidate can attempt to prevent members of his or her own partyfrom defecting. As just indicated, political candidates ate not likely to lose the votes of strong partisans, but some party members may be open to persuasion from opponents. So, candidates can try to keep partisan supporters from defecting to the opposing party's candidate. We do not know how many partisans considered defecting but ultimately decided not to do so. It could be as many as the number who do defect, 14-27 percent. Fifth, candidates may attempt to discourage voter turnout from those who support another candidate. This strategy runs counter to the ideals of democracy and I consider it to be reprehensible, so I would never recommend it to a candidate. However, it is a possible option, and Ansolabehete and Iyengar (1995) have argued that some negative political advertisements are intended to do so. So, candidates should adopc positions on some issues in an attempt to build a winning coalition of votets. Adopting a desirable position on a particular issue (e.g., private school vouchets, tax cuts) could help the candidate achieve three goals: (1) attracting the votes of independent or third parry voters, (2) discouraging one's own party members from defecting or voting for one's opponent, and (3) enticing some members of the opposing party to defect to you. Forms of Policy and Character Functional Theory offers more detail on the two topics of campaign messages, policy and character. Policy remarks can be divided into three subforms, past deeds, future plans, and general goals. Past deeds concern the record in office of an elected official (accomplishments or failures). Mention of, for example, jobs gained or lost concerns past deeds. The second form of policy utterance is fiitttre plans, or means to an end, specific proposals for policy action. A 15 percent across the board tax cut, such as proposed by Dole in 1996, exemplifies a future plan. The third form of policy utterance is general goals. Unlike future plans, goals refer to ends rather than means. Cutting taxes, without specifying which how much or which taxes to cut would illustrate a general goal. Acclaims and attacks can occur on each form of policy. Character is divided inro three subforms. Personal qualities are the personality traits of the candidate, such as honesry, compassion, strength, courage, friendliness. Leadership ability usually appears as experience in office, the ability to accomplish things as an elected official. Finally, ideals are similar to goals, but they ate values or principles embraced by the candidates rather than policy outcomes. These three forms of character can be used to acclaim and attack. Appendix 2.1 illustrates acclaims and attacks for each form of policy and character. Functional Theory offers predictions about the forms of policy and character (see also chaptet 6 on Incumbency). Broad goals (e.g., creating jobs, keeping American secure, reducing the federal deficit) are easier to acclaim than to attack. Ideals, such as justice or equality, are also easier to acclaim than to attack. For this reason, Functional Theory predicts: H3. Candidates will ttse general goals more to acclaim than to attack. H4. Candidates will use ideals more to acclaim than to attach. Past research on other campaign message forms (see Benoit, 2007a) confirms these predictions. In presidential primary debates, general goals are more often the basis of acclaims than attacks (91 percent to 9 percent) and ideals are more frequendy about acclaims than attacks (87 percent, 13 percent). In direct mail advertising from the primary campaign, general goals were used more often in acclaims than attacks (96 percent to 4 percent); this is true of ideals as well (91 percent, 9 percent). In Acceptances, general goals more frequently employed more for acclaims (92 percent) than 30 Chapter 2 attacks (8 percent); ideals are used more often for acclaims than attacks (85 percent to 15 percent). The same relationship occurs in general debates (general goals: 85 percent acclaims and 15 percent attacks; ideals: 82 percent to 18 petcent). Finally, general goals in direct mail advertising reveal the same relationship (88 percent acclaims to 12 percent attacks) as do ideals in brochures (81 percent to 19 percent). Future plans are more specific than goals; they are means to an end (the end being a goal). It is more difficult to attack a goal, such as reducing the deficit, than means to achieve that end, such as raising taxes or reducing Social Security benefits. Accordingly, Functional Theory anticipates that H5. Candidates will attack more and acclaim less on future plans than general goals. Research on the functions of these two forms of policy confirms this prediction. Presidential debates (primary, general, and vice presidential) attack in 31 percent of themes on future plans but only 11 percent of themes on general goals (Benoit, 2014). When discussing future plans in acceptances, candidates attack 11 percent of the time compared with genera) goals, which are the basis for attacks in 21 percent of utterances (Benoit, 2007a). In presidential primary and general campaign brochures, 16 percent of statements on future plans are attacks whereas 9 percent of statements on general goals are attacks (Benoit and Stein, 2005). Thus, Functional Theory views political campaign discourse as the means to an end—convincing voters to cast voters for a candidate—which is achieved through three functions: acclaiming, attacking, and defending to create die impression that you are the preferable candidate in the race. Functional Theory predicts that these functions are not equally likely to be used in campaign messages: Acclaims should be more common, defenses least common. These functions can address two topics, policy and character. Given the fact that more American voters consider policy more important than character, Functional Theory predicts that American presidential campaign discourse will address policy more often than character. Note that if more voters considered character more important than policy, Functional Theory would then predict that character utterances would outnumber policy comments. Functional Theory divides policy and character comments into more specific topics and predicts that acclaims will be more common than attacks when both general goals and ideals are discussed and attacks will be more common on future plans than general goals, These predictions are consistendy confirmed through content analysis of presidential campaign discourse. ADVANTAGES OF THE FUNCTIONAL APPROACH Functional Theory enjoys several clear advantages over other approaches to studying political campaign discourse. This approach is consistent with other approaches to analyzing televised political advertisemenrs, categorizing statements in spots as nega- Method 31 tive (attacking) or positive (acclaiming). However, it adds a third function, defense, which is overlooked in most approaches to understanding the nature of televised political spors. For example, a 1960 Nixon commercial started by acknowledging attacks on Nixon, who was running on the record of the Eisenhower/Nixon administration. The announcer then told viewers that "President Eisenhower answers the Kennedy-Johnson charges that America has accomplished nothing in the last eight years." Then viewers saw Eisenhower, who declared that "My friends, never have Americans achieved so much in so short a time," clearly denying rhe attack. Campaign discourse of this nature cannot be fully understood as negative—even though it rejects the opposition—or as positive—even though it refers to accomplishments. It begins by identifying an attack from an opponent ("the Kennedy-Johnson charges that America has accomplished nothing in the last eight years") and then explicitly rejects that attack: "never have Americans achieved so much in so short a time." Thus, one advantage of the Functional approach is that it extends analysis of campaign messages to include a third function, defenses. Defenses are not as common as acclaims or attacks in campaign messages, but they are distinctive utterancesand they are capable of reducing perceived drawbacks (costs). Defenses are more common and arguably even more important in debates. A second advantage of the Functional approach stems from its use of the theme (idea unit, argument, claim, assertion) as the coding unit instead of the entire spot. Most previous research on political spots classifies entite spots as positive or negative (a few studies add a third category, "comparative ad") or issue versus image. However, many television spots contain multiple utterances which may perform different functions, so each theme in an ad is categorized separately. Many political advertisements are mixed, containing both attacks and acclaims and/or policy and character, and that mix is not always 50/50. Benoit and Airne (2009), investigating nonpresidential TV spots from 2004, found that 42 percenr of ads contained at least one acclaim and at least one attack; 75 percent of the ads in this sample addressed both policy and character. More importantly for the current project, it is important to unitize the candidates' statements in debates into themes (one could hardly code the "entire debate" as scholars code the "entire spot"). This provides a more precise measurement of these messages content (functions, topics) than coding entire ads. Using the theme as the coding unit also facilitates comparisons of different campaign messages. For example, if those who content analyze television commeicials using the entire spar as the coding unit were to analyze other messages, what would they use as the coding unit? An entire acceptance address? An entite debate? An entire candidate webpage? Using the theme as the coding unit facilitates comparison of different kinds of campaign messages by content analyzing all messages with the same coding unit. This book relies on data produced by content analysis of political campaign messages. This method produces nominal or frequency data, which count the number of times cerrain lands of content (e.g., acclaims, attacks, or defenses; polity' or characrer), so many predictions will be tested with chi-square. This is a nonpaxametric statistic 32 Method 33 appropriate far investigating differences using frequency data. As the chapters will make dear, a large amount of data has been generated using Functional Theory. That is obviously desirable because it means that the conclusions drawn here are supported from many campaigns, many candidates, multiple message forms, multiple elective offices, and multiple countries. It also permits comparisons of the nature of various message forms. The chi-square statistic is sensitive to sample size (AO; that is, this statistic is more likely to find significance with larger Ns. Ir is important to understand the difference between significance and effect size. The significance test tells us how likely a given result would occur by chance. The statement "p < .05" means that these results would occur just by chance fewer than 5 times out of 100. Similarly, "p < .0001" means that these results should occur by chance only once out of ten-thousand times. The effect size, in contrast, indicates the magnitude of the relationship between the independent and dependent variable. For example, these are two different questions: Do challengers arrack more dian incumbents than would be expecred by chance? How much more do challengers arrack than incumbents? The former question is answered with a test of significant differences; the latter is answered with a measure of effect size. A possibility exists that a sample with a large TVa result could be statistically significant (say, "significant" even atp < .0001) and yet not make much of a difference. Research using parametric statistics increasingly reports both the significance level and the effect size, like r, ff, or eta2. I will report comparable statistics for non-parametric data: Cramers Kand (p. This statistic, like Pearsons r, can vary from 0 (no relationship whatsoever) to 1 (a perfect relationship between two variables). Unlike r, however, V\s always positive (and so it does not indicate direction of relationship; a positive r indicates a positive or direct relationship between variables; a negative r indicates an inverse or indirect relationship). This approach—reporting significance tests, consistency of effect, and effect size whenever possible—provides the best insight into the relationships investigated here. Two lands of chi-squnres are used. A chi-square goodness of fit test is used for predictions involving only one variable, such as "policy is mote common than character." Because there is only one variable (topic) no effect size can be calculated for this test: only when two variables are tested, such as policy versus character (topic) for incumbents versus challengers (incumbency), can one estimate the size of die effect of one variable on another variable. A chi-square test of cross-classification tests predictions with two variables, such as "incumbents acclaim more, and attack less, than challengers," with two variables (incumbency, function). CONTENT ANALYTIC APPROACH There are four basic steps in the coding method used to generate the data discussed in this book. First, the messages must be unitized into themes, the coding unit in this method. A theme is the smallest unit of discourse capable of expressing a coherent idea (in this case, not just any idea, but acclaims, attacks, and defenses). Themes can vary in length from a phrase to a paragraph (several sentences). Second, themes are classified into function: acclaim, attack, or defense. Because defenses are rare in TV spots, they are not coded further. Third, acclaims and attacks are classified by topic: policy or character. Next, the proper sub-form of policy (past deeds, future plans, general goal) or character (personal quality, leadership ability, ideal) is identified. First, the candidates' utterances must be unitized into themes. For example, a statement which said "I will reduce taxes, create new jobs, and keep our country safe from terrorism" would be unitized into three themes, one for each topic (taxes, jobs, terrorism), even though these are al! contained in a single sentence. On the other hand, a statement which said "Jobs are the backbone of a sttong economy. We cannot have economic recovery without jobs. That's why I will increase jobs" would be coded as one theme, jobs (the first two senrences explain why jobs are important, but do not actually establish the existence of a problem or offer a solution for jobs). Finally, a message which said "The present administration,has lost over a million jobs. If elected, I will create new jobs" would be coded as two themes: the problem of lost jobs under the current administration; the candidate's solution: a goal of creating more jobs if elected. Second, each themes function is classified as an acclaim, an attack, or a defense. A few themes do not function as acclaims, attacks, or defenses (themes which do not enact these functions are not coded). Codets must decide whether a theme performs one of these functions, and, if so, identify which one. Acclaims are themes that portray the candidate (or the candidate's political parry) in a favorable light. Obama's ad "Wonderful" in 2012 declared that "He believes smaller class sizes and great teachers are a key to a stronger economy and a stronger middle class." A stronger economy and middle class would be seen as desirable by voters. Attacks are themes that portray the opposing candidate (or diat candidate's political party) in an unfavorable light. In the spot from 2012 "Who Will Do More," Romney criticized Obama, claiming that the President "Obama took GM and Chrysler into bankruptcy, and sold Chrysler to Italians who ate going to build Jeeps in China." American firms going bankrupt, particularly those as large and as important as GM and Chrysler, would be perceived as undesirable. Acclaims can be identified in two ways. First, acclaims are positive: Virtually all statements made by a candidate about himself or herself are positive; almost all statements about an opponent are negative. Occasionally, a candidate may say something like, "My honorable opponent," but those statements are throw-away lines designed to show the speaker is a reasonable person, not to genuinely praise 34 Chapter 2 an opponent. Second, acclaims are about the candidate supported by the spot, whereas attacks are about the opponent. These two statements jointly differentiate acclaims from attacks. Defenses are themes that explicitly respond to a prior attack on the candidate (or a political party). Obama ran an ad that defended against Romneys attack ("Collapse"): "After Romney s false claim of Jeep outsourcing to China, Chrysler itself has refuted Romneys lie." To count as a defense, a statement must acknowledge or allude to a criticism and then attempt to refute it. In this example, Obama points to "Romneys false claim of Jeep outsourcing to China" before refuting that criticism. Third, themes which were classified as acclaims or arracks are then coded by topic (because defenses are so rare, topic of defense is not identified). Policy. Utterances that concern governmenral action (pasc, current, or future) and problems amenable to governmental action. In 2012, Romney aired a spot called "Secretary of Business": "Under Obama, millions of people can't find work. And more families on welfare and a record number of Americans on food stamps." Unemployment and poverty are examples of policy topics. Character. Utterances that address characteristics, traits, abilities, or attributes of the candidates (or their parries). In the ad "Seen," Obama criticized Romneys character, saying that he was "Trying to mislead us." Honesty is a clear example of a character trait. The next step in the content analytic method is to classify each policy or character utterance according to the forms of policy and character. Policy utterances can address past deeds, future plans, or general goals. Past deeds are, of course, actions . taken in the past (a candidate's record in office), whereas future plans are proposed actions (means) and general goals are ends. Character utterances can address personal qualities (e.g., courage, compassion, honesty), leadership ability (e.g., experience, vision), or ideals (i.e., values, principles). As noted earlier, appendix 2.1 provides examples of acclaims and artacks on each form of policy and character. As campaign messages are coded, other relevant information is also recorded, such as the candidates' political party, incumbency status, campaign phase (primary or general), office sought, and country. The data generated using the Functional Approach have strong reliability on each variable (functions, topics, forms of policy, forms of character); reliability is reported in each study. I compared data on presidential TV" spots reporred here with data from similar studies by Kaid and Johnston (1991, 2001), West (1997), and Geer (2006). Kaid and Johnson's data on negative ads correlated highly with the data reported here (r [n = 10] = .95, p < .0001) as do Geer's data (r [n = 12] = .87, p < Method 35 0001)- Wests data, based on a flawed sample (limited to "prominent" ads), do not correlate with the data reported here (r [n = 10] = .24, ns). The sample employed in this book includes 7256 political advertisements: 1465 presidential primary ads, 1313 presidential general spots, 66 presidential third party advertisements, 3467 nonpresidential commercials, and 945 ads from other countries. I also present data on Issue Ownership Theory and Functional Federalism Theory in chapter 9. These data are generated using compurer content analysis. Benoit and McHale (2003) used grounded dieory—the method of constant comparison (Glaser and Strauss, 1967)—to develop a typology of traits discussed by political candidares. Texts discussing personal qualities from American presidential television spots were used to develop this typology. These procedures produced four clusters or dimensions of personal qualities: Morality (decency, integrity, responsibility, fairness), Empathy (fights for the people, compassionate, understanding), Sincerity (honesty, trust, promises, openness, consistency), and Drive (strength, hard work, derermination, courage). Each term had multiple synonyms in the search list. Examples of each dimension can be found in Benoit and McHale (2003). In each of the four clusters these qualities were used to praise the candidate who sponsored the commercial (e.g., honest) and their opposires were used to attack the opponent (e.g., dishonest). After developing this typology, Benoit and McHale (2003) used the categories (and associated word lists) in computer content analysis to investigate the use of personal qualities in American primary and general election TV spots. Benoir and McHale (2004) extended this work to American congressional TV spots. The program Concordance used the word lists to count the number of rimes each word was used in each group of TV spots. ANALYTICAL CHAPTERS The next section of the book (parr II) presents my analysis of presidential advertising campaigns, grouped four campaigns per chapter (chapters 3-6). Presidential primary (chapter 7) ads and general election ads from third party candidares (chapter 8) are also addressed. The third section (part III) offers of political advertising and two theories from political science, Issue Ownership and Functional Federalism (chapter 9) and U.S. nonpresidential, and non-U.S. political advertising (chapter 10). When giving examples, I "cite" excerprs from ads by giving the candidate, year, and title of the spot (not all spots have clear titles, so I created some titles). Material placed inside square brackets ("[," "]") are descriptions of visual images (including words displayed on the screen) or sounds in the commercial. In these chapters, I discuss the functions (attacks, acclaims, and defense), the topic (policy versus character), forms of policy utterances (past deeds, furure plans, general goals) and forms of character comments (personal qualities, leadership ability, ideals). To illustrate the resuirs of my analysis, 1 also provide excerprs from spots, without specific citations. In part IV, chapter 11 discusses several contrasts in political advertising, chapter 12 discusses the development of American presidential election advertising, and chapter 13 addresses conclusions and implications of the study. 3G Chapter J APPENDIX 2.1: EXAMPLES OF ACCLAIMING AND ATTACKING ON EACH FORM OF POLICY AND CHARACTER Policy Past Deeds: Acclaim Reagan (1984): Today, inflation is down, intetest tares are down. "We've creared six and a half million new jobs. Americans are working again, and so is America. Past Deeds: Attack Eisenhower (1952): Man: General, the Democrats are telling me I never had it so good. Eisenhower: Can that be true when America is billions in debt, when prices have doubled, when taxes break our backs, and we are still fighting in Korea? Future Flam: Acclaim Ford (1976): Under my proposal ro increase the personal exemption, you would get an additional exemption of $1,250. Now that would make a sizeable increase in your weeldy take-home pay. Future Plans: Attack Johnson (1964): The other candidate wants to go on testing more atomic bombs. If he's elected, they might start testing all over again. General Goals: Acclaim Reagan (1980): We must act to put Americans back to work. We must balance the budget. We must slow the growth of governmenr. We must cut tax rates. General Goals: Attack Forbes (1996 Primary): The politicians can keep on raising your taxes and wasting your money. Method 37 Leadership Ability: Acclaim Nixon (1960): Above everything else, the American people want leaders who will keep the peace without surrender for America and the world. Henry Cabot Lodge and I have had the opportunity of serving with President Eisenhower in this cause for the last seven and a half years. We both know Mr. Krushchev. We have sat opposite the conference table with him. Leadership Ability: Attack Kennedy (I960): [Republicans want you to believe Mr. Nixon has experience in the White House.] A reporter recently asked President Eisenhower for an example of a major idea of Nixon's that Eisenhower had adopted. [Eisenhower]: If you give me a week, I might think of one. I don't remember. Ideals: Acclaim Humphrey (1968): [Woman] Humphrey is a man who has a very strong liberal background. Ideals: Attack Johnson (1964): Senator Goldwater said on Oct. 12, 1960, in Jacksonville, FL, the child has no right to an education. In most cases the child can get along just as well without it. Character Personal Qualities: Acclaim Humphrey (1968): Humphrey is without question a man that I feel everyone in this country can trusr. Personal Qualities: Attack Dukakis (1988): The other side has pursued a campaign of distortion and distraction, of fear and of smear.