Functions and Topics of 2012 American Presidential Advertising: An Increase in Attacks?

The President of the United States is the most powerful elected official in the world. The occupant of the Oval Office implements both foreign and domestic policy. It is no surprise that the quadrennial race for president attracts a great deal of attention from voters, media, and political scholars. In 2012 President Barack Obama (the Democratic incumbent) ran for re-election against Governor Mitt Romney (the Republican challenger). Television spots have been an important element of presidential campaigns since 1952 and both candidates employed them extensively in 2012. Wilson reported that over a billion dollars was spent in the general election campaign by Obama, Romney, and outside groups [1]. This is almost a half a billion dollars more than was spent on this race just four years earlier. Of course, political TV spots vary in persuasiveness and voters may react differently to any given commercial. Benoit, Leshner, and Chattopadhyay [2] applied meta-analysis to research on the effects of political advertising: Watching political spots enhances knowledge of the issues in a campaign, alters attitudes toward the candidates’ character, changed attitudes toward candidates, altered vote choice, and changed the probability that viewers would actually cast votes in the election. Political television commercials have been shown to influence voters; accordingly the presidential TV spots employed by the 2012 presidential candidates clearly merit scholarly attention. Only research on presidential advertising in each campaign can reveal the existence of possible trends in these important messages over time: This essay will report a content analysis of 2012 general presidential TV spots and compare those findings with analyses of advertising from earlier campaigns; Benoit supplies the data on presidential TV spots from 1952 through 2012 employed in this essay. This paper argues that presidential advertising has become sharply more negative in the last two campaigns because of changes in campaign fund-raising. This study focuses on ads sponsored by candidates; spots from the political parties and PACs use significantly more attacks than advertisements from candidates [3].


Introduction
The President of the United States is the most powerful elected official in the world. The occupant of the Oval Office implements both foreign and domestic policy. It is no surprise that the quadrennial race for president attracts a great deal of attention from voters, media, and political scholars. In 2012 President Barack Obama (the Democratic incumbent) ran for re-election against Governor Mitt Romney (the Republican challenger). Television spots have been an important element of presidential campaigns since 1952 and both candidates employed them extensively in 2012. Wilson reported that over a billion dollars was spent in the general election campaign by Obama, Romney, and outside groups [1]. This is almost a half a billion dollars more than was spent on this race just four years earlier.
Of course, political TV spots vary in persuasiveness and voters may react differently to any given commercial. Benoit, Leshner, and Chattopadhyay [2] applied meta-analysis to research on the effects of political advertising: Watching political spots enhances knowledge of the issues in a campaign, alters attitudes toward the candidates' character, changed attitudes toward candidates, altered vote choice, and changed the probability that viewers would actually cast votes in the election. Political television commercials have been shown to influence voters; accordingly the presidential TV spots employed by the 2012 presidential candidates clearly merit scholarly attention. Only research on presidential advertising in each campaign can reveal the existence of possible trends in these important messages over time: This essay will report a content analysis of 2012 general presidential TV spots and compare those findings with analyses of advertising from earlier campaigns; Benoit supplies the data on presidential TV spots from 1952 through 2012 employed in this essay. This paper argues that presidential advertising has become sharply more negative in the last two campaigns because of changes in campaign fund-raising. This study focuses on ads sponsored by candidates; spots from the political parties and PACs use significantly more attacks than advertisements from candidates [3].
Benoit reviews research on the nature and effects of political television advertising. Between 1952 and 2004 [11,26], general election presidential TV spots are mostly positive: 57% acclaims, 40% attacks, and 1% defenses. Incumbents are prone to use acclaims more (64% to 55%), and attacks less (35% to 44%) than challengers [26]. Historically, such commercials address policy (problems amenable to governmental action, past and future governmental action) more frequently than character (personality): 62% to 38% [26]. The advertising campaign in 2008 exhibited a sharp up-tick in attacks [27]: 34% acclaims, 65% attacks, and 0.4% defenses. Obama and McCain continued to emphasize policy (58% policy, 42% character) in their political commercials. Neither candidate in 2008 or 2016 was a true incumbent (President George W. Bush was term-limited and Vice President Dick Cheney

Functional Theory
Functional Theory [3,4,26,[28][29][30] posits a citizen will vote for the candidate who seems preferable to opponents on the dimension(s) most important to that voter [26]. Candidates can attempt to persuade voters of their desirability in three ways. First, a candidate can stress his or her positive features, engaging in acclaiming. The larger or more important the perceived advantages of a candidate, the more likely that politician will seem preferable to voters. Second, a candidate can offer criticisms of or attacks on opponents; the larger perceived the costs of a candidate, the less likely that candidate will appear less desirable to voters (candidates who attack must be aware of the possibility of backlash from voters who do like mudslinging). Finally, candidates who have sustained attacks can attempt to defend against or refute those criticisms. The fewer and/or smaller disadvantages, the more likely that a given candidate will seem preferable to voters. The three functions work as an informal variant of cost-benefit analysis. Note that Functional Theory does not assume that voters quantify acclaims, attacks, or defenses or engage in mathematical calculations, adding or averaging costs and benefits, in order to determine preferability: Acclaims can increase one's apparent benefits, attacks are meant to increase an opponent's perceived costs, and defenses could reduce one's own apparent costs. Audience response research on German chancellor debates has shown that voters react differently to acclaims and attacks [31].
Political candidates can discuss two potential topics, policy (issues) or character (image). Policy statements concern past or proposed governmental action or problems that call for action by the government. Character statements, in contrast, focus on candidates' personality. We must realize that the relative importance of the two topics of campaign discourse vary across voters. Furthermore, Functional Theory further analyzes candidate statements into three variants of policy and three aspects of character (the Appendix provides illustrations of acclaims and attacks on the forms of policy and character).
Functional Theory [26] predicts that candidates tend to acclaim more often than they attack: Acclaims do not have inherent drawbacks, but candidates may wish to moderate their attacks given that many voters say that they do not mudslinging [31,32] some incentive exists to moderate attacks. Each candidate chooses how much to attack; however, the fact that most voters do not like mudslinging is a reason that inclines most candidates to acclaim more than they attack. Defenses are predicted to be infrequent in political TV spots for three reasons: Most attacks concerns a candidate's potential weaknesses, which means that defending against an attack usually takes the candidate offmessage; second, a candidate has to identify an attack in order to refute it and doing so could inform or remind voters of a possible weakness; and third, attacks could foster the undesirable impression with voters that the candidate who is defending is reactive instead of proactive.
However, the underlying situation recently changed in an important way: Barack Obama declined federal financing for the general election in 2008 and both Obama and Romney eschewed federal financing in their clash in 2012. They believed, correctly as it turned out, that they could raise more money than the federal government would provide -and accepting federal funds meant they had to accept campaign spending limits. In 2008, arguably because of the decision to decline federal financing for the general campaign, the presidential candidates attacked more than they acclaimed, a clear break with tradition [26,27]. This raises the question of whether presidential advertising in 2012 would revert to more traditional (lower) levels of attack or begin to establish a trend toward greater negativity in these spots. Because the key situation (declining federal funds and presumably then attacking more to encourage the base to donate), we predict that: H1. Presidential television advertising in 2012 used acclaims more than attacks and attacks more than defenses.
Functional theory does not predict that candidates must mainly use attacks (or acclaims, or policy, or character). Each candidate faces different situations (e.g., the current state of the economy, presence or absence of international conflict, greater or less support from the news media) and each candidate chooses the content of his or her campaign messages.
It is important to recognize that a number of variables influence the use of attacks. For example, Benoit [3] divided presidential candidates into three groups: those who led the polls for the entire campaign, their opponents who trailed throughout the campaign, and candidates in close races in which the lead changed back and forth. Candidates who led the race attacked least often (38%), those who trailed attacked the most (51%), and candidates in close races fell in between (45%) [33]. Those who trailed throughout the campaign apparently felt that they needed to do something to change the polls and so attacked more; candidates who led felt less need to attack; candidates in close races moderated their attacks between these two extremes. Damore [23] also reported that candidates were more likely to attack when behind in polls; he also found that attacks were more likely as election day approached and when a candidate was the recipient of attacks. Elmelund-Praestekaer [34], who examined attacks in Danish political advertisements, reporting that challengers attack more than incumbents and those who were behind in the polls attacked more than leaders. Research conducted by Sullivan and Sapir [35] analyzed Taiwanese political advertisements. Challengers attacked more than incumbents, attacks were positively related to the race's competitiveness, campaigns became more negative as the campaign progressed, and spots sponsored by political parties were more negative than those from candidates. Airne and Benoit [36] contrasted non-presidential ads sponsored by candidates with those from political parties. Candidate sponsored ads in their sample acclaimed more than they attacked (74% to 26%) whereas spots sponsored by political parties attacked more than they acclaimed (59% to 41%). A study by Benoit and Airne [37] examined non-presidential ads from the 2008 election: Candidate ads acclaimed a bit more than they attacked (52% to 48%) whereas non-candidate ads (e.g., ads from Political Action Committees) attacked four times as often as they acclaimed (80% to 20%). Finally, research has confirmed that general election TV spots attack more, and acclaim less, than primary campaign ads [3]. Presidential primary ads acclaimed in 72% of utterances and attacked in 28%; general TV spots acclaimed in 54% of themes and attacked in 45%. This study focuses on one of these factors, incumbency, acknowledging that other variables also influence use of attacks in political advertising.
Research shows that incumbents are inclined to use acclaims more, and attacks less, than challengers [26]. This difference could arise from the sentiment that one should not "change horses in the middle of the stream" or the ideal that it is "better the devil we know than the devil we don't know," inclining challengers to attack more than incumbents to overcome the presumption that many incumbents enjoy.
H2. The incumbent in the 2012 American general election campaign spots acclaimed more, and attacked less, than the challenger. Arguably, the best evidence of how one will perform in an office is how that person performed in the office in the past. In recent American history, no challenger has ever served as president, although they often have experience in other elected offices, such as Senator or Governor. The fact that the incumbent's record is the most probative evidence means that both incumbents and challengers are inclined to focus on the incumbent's record. Of course, when incumbents discuss their own record, they acclaim; when challengers discuss the challengers' record, they attack. So, when discussion their record in office (past deeds), incumbents will acclaim more and attack less than challengers.
H3. The incumbent in the 2012 American general election campaign spots acclaimed more, and attacked less, than the challenger when discussing past deeds.
No incumbent can claim that everything is perfect, that no improvements are possible. However, specific proposals for change (future plans) implicitly attack the status quo and, accordingly, suggest a shortcoming of the incumbent. Therefore, challengers are more likely to acclaim (propose) future plans than incumbents, whereas incumbents are more likely than challengers to attack future plans.
H4. The incumbent in the 2012 American general election campaign spots attacked more, and acclaimed less, than the challenger when discussing future plans.
Research on previous presidential TV spots has found that policy is more common than character. Research on public opinion polls reveals that most voters say that policy is a more important influence of their vote for president than character. Candidates have an incentive to adapt to voter preferences so Functional Theory predicts that political advertising will address policy more often than character: H5. American presidential TV spots in the 2012 general election campaign discussed policy more than character.
This study also investigates the distribution of the three forms of policy and the three forms of character in these TV spots.

RQ1
. What was the relative emphasis on the three forms of policy in 2012 American presidential general election TV spots? RQ2. What was the relative emphasis on the three forms of character in 2012 American presidential general election TV spots?
Functional Theory predicts that candidates have a tendency to acclaim more, and attack less, when they discuss general goals and ideals [26]. Many ends or principles do not really lend themselves to attacks: How can a candidate a criticize a goal such as increasing employment or keeping America safe? How can a candidate attack ideals such as freedom or justice? The final two hypotheses predict that: H6. General goals were employed more frequently as the basis for acclaims than for attacks in 2012 American presidential general election TV spots.
H7. Ideals were employed more frequently as the basis for acclaims than for attacks in 2012 American presidential general election TV spots.
The method developed for Functional Theory will guide this analysis of television spots from the 2012 American presidential election. It is important to note that election campaign messages are deployed in a complex political environment. Some campaigns feature incumbents, others do not (e.g., neither John McCain nor Barack Obama were incumbents in 2008). Some races are more competitive than others. Benoit [3] found that candidates who trailed in the polls throughout the campaign attack more than their opponents (who led throughout the campaign) or candidates in races where the lead passed back and forth. He also reports that spots sponsored by candidates use fewer attacks than outside groups (PACs or 527 groups) or ads from political parties. Media coverage varies from year to year and office to office, although newspaper coverage of elections tends to emphasize horse race most, followed by character and then policy [38].

Method Sample
Texts of political TV ads from Obama and Romney used in the general (spots broadcast after these candidates secured their party's nomination) were obtained from two sources the two candidates' campaign webpages and the Stanford Political Communication Lab webpage [39]. This sample was limited to ads broadcast on television and to ads sponsored by the two candidates (not by political parties or 527 groups). 87 spots from Obama and 72 from McCain constituted the sample used in this study. The study employed content analysis following previous research using Functional Theory. First, the texts of each spot were unitized into themes, or utterances that address a coherent idea. Benoit [40] described the theme as "the smallest unit of discourse that is capable of expressing a complete idea" (p. 280). Similarly, Berelson [41] indicated that a theme is "an assertion about a subject" (p. 18). Holsti [42] states that a theme as "a single assertion about some subject" (p. 116). Themes vary in length from a short phrase to several sentences: The textual excerpt must focus on a single idea to qualify as a theme.
Second, each theme's function was classified using the following rules: ▸Acclaims portray the candidate who is sponsoring the ad favorably.
▸Attacks portray the opponent unfavorably.
▸Defenses respond to a prior attack on the sponsor of the ad.
Virtually all of the texts in this sample concerned one of the functions; other utterances were not analyzed.
Next, coders classified the topic of each theme: ▸Policy themes address past or future governmental action and problems that could be addressed by the government.
▸Character themes concern the attributes or personality of the candidates.
Defenses are infrequent in political advertising [26] so they were not classified by topic. The last step was to classify policy themes according to the three forms of policy and character themes according to the three forms of character. The Appendix offers illustrations of acclaims and attacks on the forms of policy and character, selected from the spots in this sample. An excerpt from an ad from Romney ("It Worked") illustrates how these texts were coded: "45 million Americans now on food stamps.... 23 million Americans struggling for work. Over 46 million Americans now in poverty (highest ever)." It includes three themes that are attacks (criticisms of Obama), all three concern policy generally (food stamps, unemployment, and poverty), and all address past deeds or record in office in particular (unfavorable outcomes of Obama's first term). Cohen's [43] κ was employed to calculate inter-coder reliability; this statistic controls for agreement by chance. Ten percent of texts were used to check reliability. κs were acceptable (0.98 for functions, 0.95 for topics, 0.93 for forms of policy, and 0.91 for forms of character); Landis and Koch [44] explain that κs of 0.81 and above reflect "almost perfect" agreement (p. 165). This means that the data can be considered sufficiently reliable for analysis. These procedures generate frequency data; accordingly, χ 2 will be used for statistical analyses. When appropriate (i.e., for analyses with two variables, such as functions and incumbency) φ will be reported to measure effect size.

Results
The first hypothesis predicted that these candidates would acclaim more than they attack and attack more than they defend. This prediction was not confirmed: Both candidates attacked more than they acclaimed. The least common function in these spots was defense. Combined, 66% of themes were attacks, 32% acclaims, and 2% defenses. For example, Obama's ad "Determination" illustrates an acclaim Obama: There's just no quit in America and you're seeing that right now. Over five million new jobs ,exports up forty one percent. Home values rising. Our auto industry back. And our heroes are coming home.... Here's my plan for the next four years: Making education and training a national priority; building on our manufacturing boom; boosting Americanmade energy; reducing the deficits responsibly by cutting where we can, and asking the wealthy to pay a little more and ending the war in Afghanistan, so we can do some nation-building here at home. That's the right path. Such accomplishments or goals as creating jobs, more exports, increasing home values, more domestic energy production, reducing deficits, and ending the war in Afghanistan would be seen as desirable by most voters. In contrast, an attack is exemplified in Romney's ad "Can't Afford Another Term": Announcer: Gutted the work requirement for welfare, doubled the number of able bodied adults without children on food stamps, record unemployment, more women in poverty than ever before, borrowed from China, and increased the debt to over $16 trillion, passing the burden on to the next generation.
Weakening the welfare to work program, increased reliance on food stamps, greater unemployment, more women in poverty, borrowing from China, and increasing the national debt would appear undesirable to many voters. Finally, Obama's advertisement "Dubious" is an example of a defense: Announcer: Seen this ad? Mitt Romney claiming the President would end welfare's work requirements. NBC calls this ad dubious and they're right. It's false. The Washington Post says, the Obama administration is not removing the bills work requirements at all. This spot rejects the attack that Obama weakened the welfare to work requirements (Table 1 displays these data).
H2 predicted that Obama, the incumbent, would acclaim more and attack less, than the challenger. This expectation was not confirmed in these data: Obama actually attacked even more than Romney, 70% to 61%, whereas Romney acclaimed more than Obama, 38% to 27%. See Table 1 for these data. The third prediction expected that Obama, the incumbent, would acclaim more, and attack less, than the challenger, Romney, when they discussed past deeds (record in office). This prediction was confirmed: Obama acclaimed more (43% to 5%) and attacked less (57% to 95%) than the challenger. Obama's ad "The Question" is an example of an acclaim on past deeds "30 months of private-sector job growth, creating 4.6 million new jobs." Clearly job creation is a positive accomplishment. Romney illustrated attacks on past deeds in his advertisement "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." In contrast to the excerpt from Obama, Romney's utterance identifies failures in the incumbent's first term in office. A chi-square test of cross classification confirms that this difference is statistically significant (χ 2 [df = 1] 43.5, p < 0.0001, φ = 0.46).The final contrast between the incumbent and the challenger (H4) anticipated that Obama would attack more and acclaim less than Romney when they discussed future plans. In these data Obama attacked more (74% to 47%) and acclaimed less (26% to 53%) than Romney when the candidates addressed specific future plans. An example of an acclaim on future plans occurs in Romney's ad "12 Million Jobs": Let me tell you how I will create 12 million jobs... First, my energy independence policy means more than three million new jobs. Many of them in manufacturing. My tax reform plan to lower rates for the middle class and for small business creates seven million more. And expanding trade, cracking down on China, and improving job training takes us to over 12 million new jobs.
The challenger outlined specific proposals (energy policy, tax reform, trade reform, job training) that would increase employment if Romney became president. In contrast, Obmana spot "Romney Misled the Middle Class About His Tax Plan" attacked a proposal from his opponent: "Romney's proposal would provide large tax cuts to high income households and increase the burden on middle and or lower income tax payers." Many voters would consider tax cuts for the rich and tax increases for others as an undesirable proposal. A chi-square test of cross classification indicates that these differences are significant (χ 2 [df = 1] 8.45, p < 0.005, φ =0 .23).
H5 expected that these candidates would address policy more frequently than character and the data confirm this hypothesis. The candidates discussed policy (66%) more than character (34%). To illustrate these results, an ad from Obama ("Remember") serves as an example of a policy discussion.
Mitt Romney's plan rolls back regulations on the banks that crashed our economy. Medicare voucherized. Catastrophic cuts to education. Millionaires will get one of the largest tax cuts ever. While middle class families pay more. These topics -bank regulations, Medicare, education, tax proposals -all concern policy. In contrast, Romney's spot "American Needs a Leader" addresses character: Romney: "We need American leadership. Where is American leadership? Ryan: "This is the moment where the moment and the man are meeting. We need someone who will be honest with us about our problems, who will not blame other people for the next 4 years, who will take responsibility, who will not duck the tough issues.... That's this man right here [Romney]." Leadership, honesty, and taking responsibility all concern

Acclaims
Attacks Defenses  Table 2 for these data.
The first research question addressed the use of the three forms of policy. Past deeds were discussed most (42%), followed by future plans (32%) and general goals (26%). These data are reported in Table 3. Past research reveals that ads from the 2012 campaign employed a greater emphasis on future plans and general goals than they did in the past: ads from earlier campaigns discussed past deeds (51%), future plans (19%), and general goals (31%).
RQ2 concerned the relative proportions of the three forms of character. Personal qualities were discussed most often (69%), with ideals (17%) and leadership ability (14%) at similar levels. Table 4 displays these data. Past research (Benoit, 2007;Benoit & Glantz, 2012) indicates a greater emphasis on personal qualities and leadership ability in 2012 than in the past: personal qualities 53%, leadership ability 29%, and ideals 18%).
The sixth hypothesis concerned the functions of general goals. As expected, general goals were used more to acclaim (64%) than to attack (36%). A chi-square goodness-of-fit test revealed that this difference was significant (χ 2 [df = 1] 10.29, p < 0.005). H7 predicted that ideals would more often be the basis of acclaims than attacks. This hypothesis was confirmed in these data, with 86% acclaims and 14% attacks. A chi-square goodness-of-fit test confirmed that this difference was significant (χ 2 [df = 1] 23.27, p < 0.0001).

Discussion
These data show both similarities to and differences from research on previous campaigns. The most obvious difference is the fact than both candidates combined attacked more than they acclaimed -and the incumbent attacked even more than the challenger. Figure 1 displays the functions of presidential TV spots over time. Following a negative campaign in 1952 (the first year to deploy presidential TV spots), ads became mostly positive, but gradually over time attacks became more prominent in these message forms. Both 2008 and 2012 were particularly negative. This finding underlines the importance of studying every presidential campaign (and studying multiple message forms, such as TV spots and debates). Not only is the race for the Oval Office important enough to study every year, it is clear that past trends are not invariably repeated.
These findings appear inconsistent with the predictions of Functional Theory. However, it is important to realize that Functional Theory identifies reasons rather than causes. Functional Theory does not assert that political candidates must acclaim more than the attack. Rather, it explains that candidates have a reason to acclaim more than they attack -avoiding a potential backlash from voters who dislike mudslinging -but the frequency of attacks is a choice made by candidates. They may choose to attack more than they acclaim for a variety of reasons: a strong personal dislike for the opponent, for example, or an ideological disagreement. Usually, American presidential candidates choose to acclaim more than they attack but exceptions do occur and general election TV spots in both 2008 and 2012 were exceptions. In 2008 and 2012 another factor arose: Obama (in 2008 and 2012) and Romney (in 2012) declined federal campaign funds for the general election. Accepting federal funds means a candidate must accept spending limits, and these candidates correctly believed they could raise more than they would get in federal funds     1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996   (remember that over a billion dollars was spent by and on behalf of Obama and Romney). However, raising such huge amounts of money mean that these candidates had to appeal heavily to strong partisans, and these candidates might have hoped that their attacks would appeal not just to voters but also to potential donors. In other words, these candidates had a reason (a change in how American presidential candidates financed their campaigns) to attack more in 2008 and 2012 than in previous campaigns. We speculate that this trend is likely to continue unless major changes occur in campaign financing occur. This new reason outweighed their reluctance to provoke backlash from too much mudslinging.
Comparing the level of attacks in TV spots and debates provides a reason to believe that the increased use of attacks in television spots in 2012 was a deliberate choice by candidates, rather than a general increase in negativity. Prior to 2008, American presidential debates used an average of 35% attacks. In 2008 the debates used 35% attacks [45] and 2012 saw similar results: Attacks comprised 33% of themes in debates [46]. American presidential TV spots from 1952-2008 averaged 40% attacks but attacks in ads increased to 65% in 2008 [27] and this study found 66% attacks in 2012. Attacks in debates remained about the same in 2008 and 2012 as in earlier debates; attacks in TV spots increased in 2008 and 2012 compared with earlier years. Presidential candidates leave as little to chance as possible; it is extremely unlikely that the candidates in 2008 and 2012 would accidentally or inadvertently use far more attacks in their TV spots than in their debates.
Success in an election campaign is a positive reinforcement. Obama won the presidency in 2008 with very negative ads [27] perhaps he continued that course in 2012 (because attacking McCain in ads worked for Obama in the 2008 general election, his most recent campaign) and Romney followed his lead. On the other hand, Romney's primary ads in his unsuccessful 2008 campaign had 24% attacks whereas in 2012 his primary ads had 72% attacks [47]. So, Romney might have modeled his 2012 general ads on his most recent successful campaign in the 2012 Republican primary, which was mainly negative. So, both candidates could have had reasons to "come out swinging" in their 2012 general TV spots.
Evidence indicates that candidates monitor and adapt to opponents' campaigns [23]. Once one candidate attacks heavily, the opponent may feel forced follow this lead. Monitoring their opponent's largely negative ads would not give them reason to moderate the level of attacks in their own spots. Once Obama started attacking McCain heavily in 2008, McCain may have felt it prudent to step up his attacks on the Democrat. Similarly, when Obama and/or Romney started attacking his opponent, the recipient may have felt it necessary to respond in kind. It is too early to conclude that presidential TV spots in the future will continue to be negative: Two campaigns is too small a sample to support a definite claim here. However, it is possible that we are witnessing a sea change in American presidential advertising as they attack markedly more than in earlier campaigns: Clearly we should watch future campaigns closely. This study found that the candidates for the White House in 2012 discussed policy more than character. As Figure 2 shows, after 1976 presidential candidates consistently stressed policy more than character in their television advertising. Benoit [48] reports public opinion poll data indicating that more voters report that policy is a more important determinant of their presidential vote than character; candidates' emphasis on policy is consistent with that poll data.
Although the candidates' overall functions were inconsistent with past results contrasting incumbents and challengers, some facets of their messages conform to past research. The incumbent acclaimed more and attacked less than the challenger when discussing record in office. Only the incumbent candidate has a record in the office sought and that record, arguably, is the best evidence of how the candidates would perform if elected. Both candidates discussed Obama's past deeds more frequently than Romney's past deeds. Of course, when discussing the incumbent's record (past deeds), incumbents acclaim whereas challengers attack. Both candidates attacked more than they acclaimed on past deeds (record in office) but a specific proposal for change (future plans) in the federal government can be considered to be an implicit attack on the incumbent. Of course, no system is perfect and incumbents should offer some proposals for improvement in the federal government. However, because a proposal for change is an implicit criticism of (or an admission of weakness by) the incumbent, we can expect challengers to acclaim their future plans more often than incumbents and for incumbents to attack the challengers' future plans more frequently than challengers. This effect (more attacks from the incumbent than the challenger when discussing future plans) tended to counterbalance the tendency for the previous tendency (more acclaims 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996   from the incumbent on past deeds than from the challenger), which may have played a role in the failure to confirm H2 on overall acclaims versus attacks for incumbents and challengers. When they talked about policy, past deeds was a more common topic than future plans or general goals. On character, personal qualities dominated, with leadership ability and ideals less common. The final two predictions were confirmed: It is often easier to acclaim a goal or an ideal than to attack either kind of utterance.

Conclusion
This study continued previous research investigating the content of presidential TV spots. The race for the president is so important that the 2012 campaign deserves scholarly attention on its own merits; furthermore, programmatic research also allows trends to be identified -or can alert us to possible changes in past trends. This study found similarities (policy was more common than character; themes on general goals and ideals used acclaims more than attacks; more acclaims and fewer attacks occurred on past deeds by the incumbent than the challenger; more attacks and fewer acclaims on future plans by the incumbent than the challenger) as well as differences (more attacks than acclaims; the incumbent did not acclaim more and attack less than the challenger) compared with research on earlier campaigns. The discussion helps illuminate the workings of election campaigns and Functional Theory's approach to understanding campaigns. For example, candidates generally have a reason to moderate their attacks but in some cases candidates may decide it is in their best interests to attack heavily. Presidential candidates usually do not choose to attack as much as Obama and Romney did in 2012, but the content of their messages -and particularly their television spots -is completely under their control (i.e., candidates' statements in debates are shaped in part by the questions asked). The change in how American presidential candidates financed their campaigns (declining federal funds and thereby avoiding spending limits), their most recent campaign successes [1] may have encouraged them to emphasize attacks in their ad campaigns, and the tendency for campaigns to monitor and respond to opponent's ads all are factors that may account for the increased negativity in 2012. The fact that the level of attacks in debates did not follow suit is evidence that the level of attacks in their ads was a deliberate choice (and also demonstrates the utility of using the same research method on multiple message forms). The two figures graphically display trends over time in the content of presidential TV spots.