Internet Political Ads in 2012

Abstract

A posttest-only experimental design is used to test the effects of humorous negative video ads from the 2012 presidential campaign on evaluations of Barack Obama and Mitt Romney among young people. Findings show that ads targeting Romney had a negative effect on attitudes toward him. Romney also had his evaluations lowered as the result of respondents viewing third-party ads attacking Obama. This may be consistent with some research that suggests that negative political ads have a “backlash,” or “boomerang” effect on their source. Obama, on the other hand, was largely insulated from both target and source effects. The study suggests first that candidates who decide to “go negative” may not be able to insulate themselves from unintended negative effects by framing their ads in a humorous fashion. Moreover, because the anti-Obama ads were sponsored by a third party, it suggests that the source really might matter in terms of the effects ads have on “their” candidates.

Keywords

Internet political ads negative ads political humor

During the 2012 presidential election, a record amount of money was spent on Internet advertising. The Internet campaign included both “traditional” banner ads and the posting of hundreds of video ads, many of which were produced exclusively by candidates, parties, and interest groups for the web. Do these Internet video ads have any effect? In this study, I address this question, examining the influence of humorous negative video ads from the Internet during the 2012 presidential campaign on evaluations of both President Barack Obama and his Republican challenger, Mitt Romney, among young people. While I test for the effects the ads have on their targets, I am especially interested in determining if humor can mitigate any negative effects that attack ads may have on the source of the ads.

Consistent with research on both televised negative ads and political humor, I find that ads targeting Romney had a negative effect on attitudes toward him. Romney evaluations were also lowered as the result of respondents’ viewing third-party ads attacking Obama. This may be consistent with some research that suggests that negative political ads have a “backlash,” or “boomerang” effect on their source. Obama was largely insulated from either target or source effects.

The study is important because while there is a fair amount of research on the effects of televised political advertising much less work has been done examining the impact of Internet advertising. As importantly the study seeks to reconcile research that focuses on negative political ads and that which focuses on humorous messages. The former suggests that there may be unintended negative effects on the sponsor of negative political advertisements, while the latter intimates that a humorous message may raise the likeability and/or credibility of the source of the message. From a substantive perspective the research is important because it suggests that candidates who decide to “go negative” may not be able to insulate themselves from unintended negative effects by framing their ads in a humorous fashion. Moreover, because the anti-Obama ads were sponsored by a third party, it suggests that the source really might matter in terms of the effects ads have on “their” candidates.

The 2012 Presidential Ad Campaign

One estimate suggests that in 2012, US$78 million was spent by the Obama and Romney campaigns (combined) on online advertising. This total almost quadruples the amount (US$22.2 million) spent in 2008 (Stampler, 2012). While these figures are relatively small compared with the amount of money spent on television advertising, it should be remembered that advertising on the Internet is far less expensive and a relatively smaller amount of money can yield numerous ads that “air” for longer periods of time. Both candidates produced numerous ads that appeared on the web. The Romney YouTube channel featured 290 videos, while Obama’s had a staggering 2,842 (FierceOnlineVideo, 2012). Each had videos on their own campaign websites as well, and many of these were posted and reposted on other sites.

Importantly, many people view these Internet video ads. One estimate suggests that in 2008 users spent upward of 14 million hours viewing roughly 1,000 Obama and Obama-related videos on YouTube. These videos were viewed some 50 million times (Work: Obama for America, 2012). A more recent report claimed that in 2012 as many as 55% of registered voters watched some form of online video related to the campaign or politics in general. The report further notes that approximately equal numbers watched “humorous or parody videos online dealing with political issues” (37%) and political ads (36%; Smith & Duggan, 2012). Yet another reported that from mid-2011 through election day, there were approximately 150 million views of official Obama or Romney campaign videos on YouTube (Krashinsky, 2012).

In fact, relative to television the Internet ad campaign may be becoming more relevant with each passing election cycle. In 2012, most Americans did not see televised political ads related to the presidential election. One report claimed that 96% of all television ad money (national broadcast and cable) devoted to the presidential campaign in 2012 was spent in just 10 states (Associated Press, 2012; Campaign 2012, 2012). The fact that presidential campaigns focus most of their resources on a relatively small number of key states is not necessarily a new development, but in previous election cycles money and ads were spread out over a larger number of states. This has implications for how we think about the efficacy and effects of televised political ads.

The point here is that the Internet ad campaign cannot be ignored. While this may be obvious to most, much of the discussion about presidential campaign ads still centers around television ads. But the presidential ad campaign on television is restricted to a few media markets in a few states and the Internet is where the campaign is increasingly being carried out. This makes the present study especially relevant. More ads are “aired” on the Internet than on television, and they are more accessible. It may be the case that more of the scholarly focus on the presidential ad campaign would be better directed toward the web.

A final point about the 2012 ad presidential campaign germane to this research is that it was the most negative in recent memory (Franklin Fowler, 2012; Slack, 2012). By one estimate, only 14.4% of pro-Obama and 20.4% of pro-Romney ads, which include those sponsored by the candidate, their party, or other groups, could be considered strictly positive ads. The remaining ads were either negative (58.5% for Obama and 49.2% for Romney) or contrast ads. This makes 2012 an excellent choice for research examining the effect of negative political ads on individual attitudes.

Televised Negative Ads, Internet Ads, and Political Humor Effects Research

Relative to the amount of research devoted to televised political advertisements, little has been written about the ads that candidates, parties, and interest groups produce for the Internet (see Kaid, 2006). One early study focused on the 2004 campaign cycle, providing a brief but comprehensive overview of advertising on the web. It included a discussion of possible advantages and disadvantages of web advertising, types and content of web ads, their potential effects, and news coverage of political ads on the Internet (Kaid, 2006).

Another focused on sponsorship and issue content of humorous television and web ads during the 2004 presidential campaign. Examining ads produced by citizens (rather than candidates and their allies), this study found that web ads tended to be both longer and funnier than television ads (Kaid & Postelnicu, 2008). Ads aired throughout the 2012 Republican nomination season by the likes of Ron Paul, Rick Santorum, and others, tended to fit this mold. Web-only ads do, in other words, seem to be at least somewhat different than televised ads.

The paucity of research on Internet ads is especially evident with regard to the effects of these ads. One early study showed that the medium might matter in this regard. Kaid (2002) showed that independents who were exposed to web ads prior to the 2000 election were swayed in favor of Al Gore, while those independents who viewed the same spots on television favored George W. Bush (other differences were present as well; see also Kaid, 2003). Another experimental study of the effects of web advertising in 2004 suggested that among young viewers, John Kerry’s ads were somewhat successful in raising evaluations of the candidate (Kaid & Postelnicu, 2005).

Are Internet messages similar to other audiovisual messages (e.g., televised messages) in terms of their persuasive power? While some research (e.g., Karson & Korgaonkar, 2001) suggests that there may be critical differences in how Internet and other messages (including television) are processed, but there is some reason to believe that online video may be able to sway opinion and attitudes. Practitioners clearly believe that this is the case, as an Internet search of the term “persuasion and Internet video” makes abundantly clear. Some academic research (e.g., Baumgartner (2007, 2008) has found that online humorous messages (in particular, humorous video clips) have effects similar to that of televised humor. So while there may be a “medium effect” (Innis, 1964; McLuhan, 1962; Meyrowitz, 1994) that renders online video or ads different than their televised counterparts in terms of their ability to persuade, we can safely look to research on the latter for guidance as a starting point in understanding the former.

In this regard, there is a wealth of scholarship. Political scientists and communication scholars have been investigating televised political ads for several decades, and some consensus seems to exist that negative ads are more memorable, generate more knowledge of the campaign, and may increase cynicism (Lau, Sigelman, & Rovner, 2007). More germane to the present study is research that examines the effects of viewing negative television ads on candidate evaluations and vote intention. Here too there seems to be some agreement that there is (at minimum) a “modest tendency for negative campaigns to undermine positive affect for the candidate they target” (Lau, Sigelman, & Rovner, 2007, p. 1182). This includes a lowering of evaluation and vote intention (see also Faber, Tims, & Schmitt, 1993; Tinkham & Weaver-Lariscy, 1993; West, 1993).

On the other hand, research suggests that negative political ads may have unintended effects as well (Lau, Sigelman, & Rovner, 2007). In particular, these ads may actually lower evaluations of the candidate who sponsors them, undermining to some extent the original purpose of the ads (see Garramone & Smith, 1984; Jasperson & Fan, 2002; Merritt, 1984; Sonner, 1998). However evidence in this regard is less conclusive. For example, one recent study found that negative ads “decreased the perceived warmth of the source while simultaneously increasing the perceived competence” (Carraro & Castelli, 2010, p. 617). It should also be noted that very few studies focus on the effects of negative political ads on both sponsor and target (Lau, Sigelman, & Rovner, 2007).

Because the ads selected for this experiment were both negative and humorous, it may be appropriate to look beyond research on the effects of negative political ads to research on the effects of humorous messages. Recent research shows that exposure to political humor has a message consistent effect on evaluations of the target of the humor. Because most political humor contains negative messages about its targets (Niven, Lichter, & Amundson, 2003, 2008), this translates into a negative effect on perceptions of the target (Baumgartner, 2008; Baumgartner & Morris, 2006; Morris, 2009).

Evidence regarding the effect of humor on evaluations of the source of the message is more mixed. One early experimental study of the effects of a radio drama parody found the humor had a negative effect on source evaluations (Berlo & Kumata, 1956). A later experiment involving public speaking showed that other-disparaging humor reduced evaluations of the speaker’s character (Hackman, 1988). Eisend’s (2009) meta-analysis of advertising literature suggests that humor reduces source credibility and provides no evidence that it increases liking of the source.

However, other research suggests that humor may have a positive effect on evaluations of the source. An early experiment found that written satire may increase both source liking and credibility (Powell, 1978). Some advertising research suggests that humorous ads may be positively related to brand attitude or the source of the ad (e.g., Gelb & Zinkhan, 1986). A recent experiment involving stand-up comedians as message source found that humor was associated with greater source liking, although the message itself was somewhat discounted (Nabi, Moyer-Gusé, & Byrne, 2007). Two early reviews of advertising literature found that while individual findings tended to vary, humor appeared to have a positive effect on perceptions of the source of the humor. This positive effect included increased source liking and in some cases credibility (Markiewicz, 1974; Sternthal & Craig, 1973). A later review suggested that humor may or may not enhance source credibility but does enhance source liking (Weinberger & Gulas, 1992).

The weight of evidence on the effects of negative political ads and political humor on the target of the message suggests the following hypothesis:

Hypothesis 1: Viewing humorous negative political advertisements on the Internet will have a negative effect on individuals’ evaluations of the target of the advertisements.

Expectations with regard to the effects of the humorous negative Internet ads on the sponsor of the ads are more tentative, inasmuch as there are two conflicting streams of research to draw from. Evidence from research on negative political ads suggests that there may be a negative effect resulting from exposure to the ads. Humor effects research findings seem to be contradictory. This may be due to the fact that the methodologies of each, in particular the types of sources used in each experiment, were different. However, some evidence suggests that if a negative ad is also perceived as being funny, there may be a positive effect on the sponsor (or source) of the ad. Accordingly, I hypothesize:

Hypothesis 2: Viewing humorous negative political advertisements on the Internet will have a positive effect on individuals’ evaluations of the source of the advertisements.

Data, Experimental Stimuli, and Method

To test these propositions I used a posttest-only experimental design (Campbell & Stanley, 1966) in an online survey conducted during the 2012 presidential election campaign. I obtained e-mail addresses for all 23,480 undergraduate students at a large Southeastern public university and then randomly assigned them into a control group and four experimental groups. On October 15, 2012 I e-mailed each student, asking for their help in the “Political Ads & the 2012 Campaign” project. While participation was voluntary, I offered to donate US$.10 to Habitat for Humanity for each completed survey.¹

Each e-mail directed respondents to one of five websites, corresponding to their assignment into groups. Control group participants were taken directly to the survey. Those in the four experimental groups were directed to one of four websites that asked them to a “short video clip” prior to taking the survey. The first experimental group was asked to view a clip that combined two anti-Obama ads and the second was asked to watch two anti-Romney ads. The third experimental group viewed the anti-Obama ads followed by the anti-Romney ads, while the fourth was presented with anti-Romney followed by anti-Obama ads.

I sent two additional e-mails reminding students about the project and asking again for their help. The survey was closed on October 24, with a total of 1,352 responses. While disappointing, the low response rate (5.8%) was somewhat expected given survey fatigue among most university students (Porter, Whitcomb, & Weitzer, 2004). As is common in much survey research (O’Rourke & Lakner, 1989), females (66.9%) were slightly overrepresented in the sample. The distribution of partisan identification was less representative of young adults nationally, both in the larger sample (28.7% Democrat, 36.9% Republican) and within groups. This is reflective of the fact that the sample was drawn from a rural, Southeastern university. While difference of means tests showed there were no significant differences between experimental groups on key political and demographic variables, the pro-Republican orientation of the sample clearly limits the generalizability of the findings. Table 1 summarizes response rates by experimental condition.

Table 1.

Response Rate by Experimental Condition.

Experimental Group	N
Control	439
Viewed anti-Obama ads	238
Viewed anti-Romney ads	222
Viewed all ads (anti-Obama, anti-Romney)	240
Viewed all ads (anti-Romney, anti-Obama)	213
Total N	1,352

There were any number of ads produced during the 2012 campaign, but the principal criterion for use in this experiment was how funny the ad was. I enlisted the aid of students in my (fall 2012) “Campaigns and Elections” class and that of a graduate assistant to help survey the ad landscape. All were asked to find ads that appeared to have been produced to be funny. From those that were brought to my attention, I preselected the four that I considered to be the funniest. Two of the ads targeted Obama and two Romney. The two anti-Obama ads were joined to make a single anti-Obama clip, and the same was done with the anti-Romney ads.

The first ad targeting Barack Obama was produced by the conservative group Right Change. Released on June 19, 2012, the ad was a parody of the Dos Equis beer “The Most Interesting Man in the World” commercials, titled “The Most Arrogant Man in the World.” Although there were numerous postings of the ad on various websites, the main posting on YouTube boasted almost 1 million views. The second ad attacking Obama was titled “Run, Joe, Run” and was posted by the conservative political action committee (PAC) American Crossroads on August 17, 2012. The ad is a short compilation of gaffes by Vice President Joe Biden. It too was originally posted on YouTube, where the main posting attracted roughly one-half million views.

Both of the ads targeting Mitt Romney were sponsored by Obama for America and could be found on the President’s campaign website as well as his YouTube channel. Easily the best known of all four ads included in this experiment was Obama’s “Big Bird” ad. The ad took aim at Mitt Romney’s reference to Sesame Street’s Big Bird character—as an example of how government funding of the Public Broadcasting System is wasteful spending—during the first presidential debate. Released on October 9, 2012, the ad drew approximately 3.5 million views on YouTube. This ad, unlike the other three, actually aired on television in an extremely limited run² and was shown numerous times on cable and network news programs. However, given the fact that Internet viewership of the ad was so high, I felt justified including it in the design.

The second Romney attack ad included was titled “Romney Economics: The Infomercial,” and was released on August 28, 2012. Modeled as an infomercial, the narrator presented a satirically upbeat view of the (alleged) low spots of Romney’s economic record. This ad was the least viewed of all four, drawing only 35,000 views (transcripts of all four ads can be found at http://www.ecu.edu/polsci/faculty/other/baumgartner_sscr-transcripts.pdf).

Two points about the anti-Obama ads are worthy of note, each of which may confound an examination of their effects. First, “Run, Joe, Run” attacks the Obama administration but does so only indirectly by attacking the vice president. The ad never mentions the president himself. This is in direct contrast to both of the anti-Romney ads which attack Romney himself. This may have implications for how strongly the ads affect evaluations of Obama.

Second, both anti-Obama ads were sponsored by outside groups rather than by Romney himself. While both were obviously supportive of Romney and critical of Obama, the fact is that neither is associated directly with Romney. On one hand, it might be the case that the average viewer does not see or care about who the actual “source” of the ads. On the other hand they might, and in fact there is some evidence to suggest that even when an attack ad is aired by a noncandidate group on behalf of a candidate, it may have unintended negative effects on favorability ratings of that candidate (Jasperson & Fan, 2002). In the end, the fact that I am examining the effect of ads on a “source” that differs from the actual source may be problematic. Both of these factors obviously introduce potentially confounding effects to the experiment. In the end, I considered this to be a worthwhile trade-off given my desire to find humorous ads. And for the sake of convenience, I will continue to refer to potential or actual effects of the anti-Obama ads on Romney in terms of “sponsor” or “source” effects, in spite of the fact that technically, Romney was neither.

Manipulation checks of the four ads were administered immediately subsequent to their viewing. Respondents were first asked, “Do you consider the ads you just watched to be negative ads?” (yes, no). Those who responded “yes” were then asked, “how negative?” Response choices here included 1 = somewhat, but not very negative, 2 = negative, and 3 = very negative (those who said “no” in the previous question were coded as “0”). In all, 86.4% of those who viewed the ads (n = 913) considered them to be negative (M = 1.92, SD = 1.00). There were no significant differences within this variable between groups. This suggests that it is fair to characterize the ads as negative.

Equally important were questions asking respondents to rate how funny the ads were. For each ad viewed I asked respondents, “On a scale of 1 (NOT AT ALL FUNNY) to 5 (EXTREMELY FUNNY), how funny (if at all) did you think the ad [ad name] you just watched was?” All were viewed as being at least somewhat funny. The least funny was “Romney Economics” (M = 2.24, SD = 1.28), while the other three seemed to be perceived as equally funny. Table 2 presents summary information about the experimental stimuli.

Table 2.

Experimental Stimuli.

Ad Name, Sponsor	Run Time	Views^a	How Funny
“Most Arrogant Man,” Right Change	1 min, 05 s	1.0	2.63 (SD = 1.43), n = 691
“Run, Joe, Run,” American Crossroads	1 min, 01 s	.5	2.64 (SD = 1.39), n = 691
“Big Bird,” Obama for America	31 s	3.5	2.57 (SD = 1.37), n = 675
“Romney Economics,” Obama for America	1 min, 26 s	.035	2.24 (SD = 1.28), n = 675

Note. SD = standard deviation.

^a Figures are approximate number of views, in millions, from the main listing on Youtube.com (as of November 2012).

To test for the effects of these ads on the targets and sponsors of the ads, I relied on the following questions, asked of both Barack Obama and Mitt Romney. The first question asked respondents, “On a scale of 1-10, how do you feel about the following candidates for president? The higher the number, the more favorable you feel, the lower the number, the less favorable you feel.” Respondents were instructed to select a “no answer” option if they had not heard about the candidate.

In addition to this general thermometer score measure, I asked respondents to rank each candidate on a number of characteristics. This question was phrased, “On a scale of 1 (not well at all) to 10 (extremely well), how well do the following words or phrases describe [candidate name]?” A table presented respondents with the following traits: “really cares about people like me,” “honest,” “inspiring,” “knowledgeable,” “trustworthy,” “provides strong leadership,” “experienced,” “competent,” “easy to like,” “qualified,” and “shares my values.” I then created an index of scores from all 11 traits (resulting values ranged from 11 to 110).³

A third question asked, “How would you rate the manner in which [candidate name] (and his supporters) have been conducting his campaign?” This question was designed to measure whether the use of humor had a positive effect on perceptions of the campaign itself. Response choices included 4 = completely fair or appropriate, 3 = mostly fair or appropriate, 2 = not so fair or appropriate, and 1 = not fair or appropriate at all. Those who answered “I don’t know, or I haven't thought about it too much” were recoded as missing data.

The primary independent variables of interest were based on whether the individual had seen any of the ads. Those who viewed the anti-Obama ads were coded as “1” and all others were coded as “0.” A similar strategy was followed for those who viewed the anti-Romney ads. Those who viewed all four ads, regardless of the order in which they viewed them, were collapsed into a single variable and coded as “1” while all others were coded as “0.” This resulted in three dummy “viewed ads” variables, leaving the control group (those who viewed no ads) as the reference category.

I controlled for several other variables as well. Partisan identification was measured by asking respondents, “Generally speaking, do you consider yourself a Republican, a Democrat, an independent, or something else?” This variable was coded on a scale of 1–5, where 1 = strong Democrat, 3 = independent, and 5 = strong Republican. Those who responded “I don't know, or I haven't given it much thought” were recoded as missing data. Ideology was measured in a similar fashion: “In general, how would you describe your political views?” Response choices ranged from 1 = very liberal to 5 = very conservative (3 = moderate), and those who answered “I don't know, or I haven't given it much thought” were again coded as missing data.

I measured interest in the campaign by asking respondents, “On a scale of 1 to 10, with 1 being not at all and 10 very closely, how closely would you say you have been following news about the 2012 presidential election?” I also measured how much respondents had heard about the two candidates was by asking, “On a scale of 1 (nothing) to 10 (a great deal), how much would you say you have heard about the following presidential candidates, either from family, friends, neighbors, and so on, or from various media reports?” Respondents were presented with the names of both Obama and Romney. This variable was included in an attempt to control for how malleable attitudes toward each candidate may have been. Finally, I included measures of gender (0 = female, 1 = male), age, and race (0 = White, 1 = African American, non-White Hispanic, Asian, or other). In the following section, I present my findings.

Findings

In order to assess whether the ads had any effect on evaluations of Obama and Romney, I ran ordinary least squares regression on the candidate thermometer and trait index scores. Evaluations of how the candidates ran their campaign were analyzed using ordered logistic regression. Results are presented in Table 3. Columns 2 through 4 show results of the analyses of the effects of the ads on evaluations of Obama, while Columns 5 through 7 presents similar results for Romney evaluations.

Table 3.

Ad Effects: Obama.

	Obama			Romney
	Thermometer	Trait Index	Fair Campaign	Thermometer	Trait Index	Fair Campaign
Saw anti-BO ads	−.23 (.18)	−2.45 (.87)**	−.02 (.17)	−.53 (.20)**	−3.07 (.81)***	−.37 (.18)*
Saw anti-MR ads	.29 (.18)	.31 (.89)	.03 (.18)	−.41 (.19)*	−2.25 (.79)**	−.36 (.17)*
Saw both ads	−.19 (.15)	−2.16 (.72)**	−.35 (.14)*	−.22 (.16)	−2.39 (.65)***	−.37 (.14)**
Heard about Romney	−.08 (.04)*	−.42 (.19)*	−.07 (.04)	.13 (.04)***	.87 (.17)***	.14 (.04)***
Heard about Obama	.11 (.04)**	.82 (.19)***	.09 (.04)*	−.12 (.04)**	−.64 (.17)***	−.10 (.04)**
Followed news about campaign	.12 (.03)***	.73 (.14)***	.00 (.03)	.14 (.03)***	.81 (.13)***	.01 (.03)
Partisan ID (Dem. = 1, Rep. = 5)	−1.53 (.08)***	−6.28 (.40)***	−.63 (.08)***	1.45 (.09)***	4.90 (.37)***	.89 (.08)***
Ideological ID (lib. = 1, cons. = 5)	−.97 (.08)***	−3.79 (.39)***	−.24 (.08)**	.92 (.09)***	3.12 (.35)***	.54 (.08)***
Gender (0 = female, 1 = male)	.12 (.13)	−1.02 (.63)	−.38 (.12)**	−.30 (.14)*	−.03 (.57)	−.12 (.12)
Age	−.00 (.00)	−.02 (.01)*	−.00 (.00)*	−.00 (.00)**	−.03 (.01)***	−.01 (.00)***
Race (0 = White, 1 = non-White)	1.32 (.16)***	6.29 (.78)***	.83 (.16)***	−.67 (.17)***	−3.41 (.70)***	−.08 (.16)
Constant	11.43 (.38)***	66.54 (1.82)***	—	−2.72 (.40)***	3.46 (1.64)*	—
Adjusted R²	.619	.561	—	.552	.497	—
Cut 1	—	—	−4.88 (.38)	—	—	1.63 (.37)
Cut 2	—	—	−2.69 (.36)	—	—	3.92 (.38)
Cut 3	—	—	.56 (.36)	—	—	7.29 (.43)
Pseudo R²	—	—	.128	—	—	.196
LR χ²(11)	—	—	355.40***	—	—	551.62***
Log likelihood	—	—	−1,208.69	—	—	−1,131.20
N	1,235	1,235	1,168	1,235	1,235	1,156

Note. Anti-BO = anti-Barack Obama; anti-MR = anti-Mitt Romney; ID = identification; Dem. = Democrat; Rep. = Repubican; lib. = liberal; cons. = conservative; LR = logistic regression. “Thermometer,” “trait index,” cell entries are unstandardized ordinary least squares regression coefficients with standard errors in parentheses. “Fair campaign” cell entries are ordered logit coefficients with standard errors in parentheses.

* p < .05. ** p < .01. *** p < .001 (two-tailed).

Results show that partisan identification and ideological orientation were, predictably, highly significant (p < .001) in every model. Having heard about either candidate was also a significant factor in virtually every model. Finally, results suggest that race was a significant factor in most cases, with non-Whites favoring Obama.

The evidence with regard to target effects is mixed. Obama seemed to have been largely immune to any ill effects on his evaluations as a result of ads aired by Romney’s supporters. Columns 2 through 4, Row 1, show that while his trait index scores were lowered by slightly over 2% (2.45 of 100 points), the effects of viewing the anti-Obama ads were insignificant in two of the three models. This was not the case for evaluations of Romney. Columns 5 through 7, Row 2, show that viewing the Obama ads targeting Romney had a significant negative effect on Romney thermometer and trait index scores as well as perceptions of how fairly he and his allies were conducting their campaign. Romney’s thermometer scores were lowered by one-half point (approximately 5%), and his trait index scores were lowered by 3%. Predicted probabilities for the ordered logit analysis for “fair campaign” reveal that 48.6% of those who viewed the anti-Romney ads were likely to respond that Romney and his allies were conducting their campaign not so fairly or appropriately or not at all fairly or appropriately, as opposed to the 39.4% who did not view the ads.

My first hypothesis stated that, consistent with research on negative political advertisements and political humor, viewing humorous negative political advertisements would have a negative effect on individuals’ evaluations of the target of the advertisements. The analyses presented in Table 3 (four of the six models, including all three Romney models) at least partially support this proposition.

It is the effect of these ads on source evaluations that was the primary interest of this research. Here the evidence clearly disproves the second hypothesis, although evaluations of Romney seem to have been more affected (in the opposite direction as hypothesized) than those of Obama. Columns 2 through 4, Row 2, show that Obama’s evaluations were unaffected by the viewing of his Romney attack ads. So while his evaluations were not raised, he was immune to any “backlash” effects of the humorous ads the campaign aired attacking Romney.

This was not the case for Romney. Columns 5 through 7, Row 1, show that in spite of the fact that they were cast in a humorous light, the ads attacking Obama had a significant and negative effect on Romney thermometer and trait index scores. Here thermometer scores were lowered by approximately 4%, while his trait index scores were lowered by roughly 2%. In both cases, the effects were small, but negative. Predicted probabilities for the “fair campaign” analysis show that that 48.1% of those who viewed the anti-Obama ads were likely to respond that Romney and his allies were conducting their campaign not so fairly or appropriately or not at all fairly or appropriately, as opposed to the 39.3% who did not view the ads. In general, the strategy to attack Obama resulted in some unintended and negative consequences in terms of evaluations of Romney, in spite of the fact that the ads were crafted to be funny.

The second hypothesis stated that as the result of viewing humorous negative political ads, individuals’ evaluations of the source of the advertisements will be higher. This expectation was not borne out. In fact the models show that while Obama may have been immune from any negative effects of attacking Romney with these humorous ads his evaluations were not raised. The same cannot be said of Romney, whose evaluations were actually lowered as the result of respondents viewing the ads aired by his supporters. Humor, in other words, did not mitigate the backlash effect of these ads.

Discussion

The humorous negative ads viewed by subjects in this study had varied effects on Mitt Romney and Barack Obama. In terms of their effects on the targets of the ads, evaluations of Obama were unaffected by the humorous negative attacks by Romney in two of the three models. On the other hand the anti-Romney ads had a negative effect on Romney’s thermometer and trait index scores as well as perceptions that he and his supporters were was conducting a “fair campaign.”

With regard to the main focus of this research, the effect that humorous negative ads had on the source of the ad, the evidence is similarly mixed. As the result of viewing ads attacking Obama, respondents’ reported more negative thermometer and trait index scores for Romney. Obama seemed to be insulated from this negative backlash effect.

What explains the difference in findings on the effects of the ad sets on both targets and sources? Why were evaluations of Obama less likely to be affected? It should first be remembered that the sample demographic (young adults) were generally more supportive of Obama. It might also be the case that as the incumbent president attitudes toward Obama were less likely to be influenced simply because he was, by this point, a known quantity. Baumgartner and Morris (2006) saw a similar dynamic in their experimental research on the effects of The Daily Show on evaluations of the incumbent President George W. Bush and his challenger, John Kerry in 2004.

Alternatively, we can look to the humor itself for possible explanations. Obama may have been more insulated from the negative effects of attack ads than was Romney because the humor in his ads was more lighthearted. In particular, the “Big Bird” ad was far more lighthearted than was “The Most Arrogant Man,” which may have affected the way in which respondents’ viewed the negativity-humor mix of each ad. Another possible explanation may have to do with the fact that Obama may have been insulated from the negative effects of respondents’ viewing ads attacking him because one of the ads actually attacked Vice President Biden rather than Obama.

Finally, each of the two ads attacking Obama were sponsored and aired by third-party groups. This may have made the difference in how Romney was perceived: he was not the actual source of the ad. Further, he was not showing himself to be lighthearted by poking fun at his opponent, but rather as one who was engaged in the low politics of mudslinging by standing by and letting others attack his opponent.

Like many experiments in the social sciences this research suffers from a lack of generalizability. Subjects were selected from a single university and opted into the study, making the sample less representative of college students generally. Moreover, there is no guarantee that subjects in any of the experimental groups did not previously see these ads, nor did the design control for the fact that subjects may have been exposed to other ads. However, the assumption is that random assignment of subjects to the experimental groups eliminated any systemic bias from the sample in this regard. On the other hand, the experimental design and the fact that students viewed the online clips on their own computers and in their own time renders the research high in internal validity.

Future research might attempt to isolate the humorous versus the negative effects of similar ads by conducting a two (humorous ad, nonhumorous ad) by two (negative ad, nonnegative ad) experiment. Such an experiment would help further by using ads from a non-presidential (in other words, a lower salience) race. It would also be beneficial to understand the effect of these ads on older populations rather than only young adults, whose attitudes are more malleable.

Does humor mitigate unintended negative effects of negative advertisements? The evidence presented in this study suggests that it might be possible under certain circumstances. In particular, incumbents using lighthearted humor to attack their opponents may not see an increase in their favorability, but they may not see a decrease either. Challengers may not see the same benefit.

This is especially important to understand because Internet advertising is likely to constitute an increasing share of the advertising effort of presidential (and other) candidates in the future. The low costs of “airing” Internet ads makes it likely that campaigns will continue to produce even greater numbers of ads, and the nature of the medium allows for these ads to be longer, edgier, and in many respects more creative. Many of Ron Paul’s Internet ads during the 2012 Republican nomination season are good evidence of this. This in turn makes it more likely that candidates, their organizations, and supporters will experiment by using humor to make it appear as if their attacks are “only a joke.” This research represents a first attempt at understanding how effective such a strategy might be and also adds to our general knowledge of the effects of an increasingly large and important body of Internet political ads.

As a first step toward understanding the effects of Internet-only ads on their targets and sources, future research should more carefully consider the actual source of the ads. It may be the case that the ad source really matters, in particular, that respondents viewed the interest group ads as more underhanded. This is especially important to consider, given that in future campaigns we will likely see more online ads from external groups due to the Citizens United ruling and the formation of Super PACs rather than the candidates themselves.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Associated Press. (2012). 2012 campaign ad blitz: 1 million ads, $1 billion spent. Retrieved November 4, 2012, from http://www.kob.com/article/stories/S2823006.shtml

Baumgartner

J. C.

(2007). Humor on the next frontier: Online political humor and its effects on youth. Social Science Computer Review, 29, 319–338.

Baumgartner

J. C.

(2008). Editorial cartoons 2.0: The effects of digital political satire on presidential candidate evaluations. Presidential Studies Quarterly, 38, 735–758.

Baumgartner

J. C.

Morris

J. S.

(2006). The “Daily Show” effect: Candidate evaluations, efficacy, and the American youth. American Politics Research, 34, 341–367.

Berlo

D. K.

Kumata

(1956). The investigator: The impact of a satirical radio drama. Journalism Quarterly, 33, 187–198.

Campaign 2012. (2012). Mad money: TV ads in the 2012 presidential campaign. Washington Post. November 14, 2012. Retrieved December 22, 2012, from http://www.washingtonpost.com/wp-srv/special/politics/track-presidential-campaign-ads-2012/

Carraro

Castelli

(2010). The implicit and explicit effects of negative political campaigns: Is the source really blamed? Political Psychology, 31, 617–645.

Eisend

(2009). A meta-analysis of humor in advertising. Journal of the Academy of Marketing Science, 37, 191–203.

Faber

R. J.

Tims

A. R.

Schmitt

K. G.

(1993). Negative political advertising and voting intent: The role of involvement and alternative information sources. Journal of Advertising, 22, 67–76.

10.

FierceOnlineVideo. 2012. In the online video race, Obama’s got Romney beat. October 31, 2012. Retrieved December 7, 2012, from http://www.fierceonlinevideo.com/special-reports/online-video-race-obamas-got-romney-beat

11.

Franklin Fowler

(2012). Presidential ads 70 percent negative in 2012, up from 9 percent in 2008. Wesleyan Media Project. May 2, 2012. Retrieved January 11, 2013, from http://mediaproject.wesleyan.edu/2012/05/02/jump-in-negativity/

12.

Garramone

G. M.

Smith

S. J.

(1984). Reactions to political advertising: Clarifying sponsor effects. Journalism Quarterly, 61, 771–775.

13.

Gelb

B. D.

Zinkhan

G. M.

(1986). Humor and advertising effectiveness after repeated exposures to a radio commercial. Journal of Advertising, 15, 15–20, 34.

14.

Hackman

M. Z.

(1988). Audience reactions to the use of direct and personal disparaging humor in informative public addresses. Communication Research Reports, 5, 126–130.

15.

Innis

H. A.

(1964). The bias of communication. Toronto, Canada: University of Toronto.

16.

Jasperson

A. E.

Fan

D. P.

(2002). An aggregate examination of the backlash effect in political advertising: The case of the 1996 U.S. Senate race in Minnesota. Journal of Advertising, 31, 1–12.

17.

Kaid

L. L.

(2002). Political advertising and information seeking: Comparing exposure via traditional and Internet channels. Journal of Advertising, 31, 27–35.

18.

Kaid

L. L.

(2003). Effects of political information in the 2000 presidential campaign: Comparing traditional television and Internet exposure. American Behavioral Scientist, 46, 677–691.

19.

Kaid

L. L.

(2006). Political web wars: The use of the Internet for political advertising. In Williams

J. P.

Tedesco

J. C.

(Eds.), The Internet connection: Perspectives on the web in campaign 2004 (pp. 67–82). Lanham, MD: Rowman & Littlefield.

20.

Kaid

L. L.

Postelnicu

(2005). Political advertising in the 2004 election: Comparison of traditional television and Internet messages. American Behavioral Scientist, 49, 265–278.

21.

Kaid

L. L.

Postelnicu

(2008). Air amusement versus web wit: Comparing the use of humor in 2004 political advertising on television and the Internet. In Baumgartner

J. C

Morris

J. S.

(Eds.), Lauging matters: Humor and American politics in the media age (pp. 117–130). New York, NY: Routledge.

22.

Karson

E. J.

Korgaonkar

P. K.

(2001). An experimental investigation of internet advertising and the elaboration likelihood model. Journal of Current Issues & Research in Advertising, 23, 53–72.

23.

Krashinsky

(2012). How YouTube has transformed the 2012 presidential election. The Globe and Mail. November 2, 2012. Retrieved December 7, 2012, from http://www.theglobeandmail.com/news/world/us-election/how-youtube-has-transformed-the-2012-presidential-election/article4871244/

24.

Lau

R. R.

Sigelman

Rovner

I. B.

(2007). The effects of negative political campaigns: A meta-analytic reassessment. The Journal of Politics, 69, 1176–1209.

25.

Lee

K. A.

(2012). Obama ad ruffles feathers on Sesame Street. Daily News. October 9, 2012. Retrieved November 12, 2012, from http://articles.nydailynews.com/2012-10-09/news/34346860_1_obama-ad-campaign-ads-obama-campaign

26.

Markiewicz

(1974). Effects of humor on persuasion. Sociometry, 37, 407–422.

27.

McLuhan

(1962). The Guttenberg galaxy: The making of typographic man. Toronto, Canada: University of Toronto.

28.

Merritt

(1984). Negative political advertising: Some empirical findings. Journal of Advertising, 13, 27–38.

29.

Meyrowitz

(1994). Medium theory. In Crowley

Mitchell

(Eds.), Communication theory today (pp. 50–77). Cambridge, England: Polity.

30.

Morris

J. S.

(2009). “The Daily Show with Jon Stewart” and audience attitude change during the 2004 party conventions. Political Behavior, 31, 79–102.

31.

Nabi

R. L.

Moyer-Gusé

Byrne

(2007). All joking aside: A serious investigation into the persuasive effect of funny social issue messages. Communication Monographs, 74, 29–54.

32.

Niven

Lichter

S. R.

Amundson

(2003). The political content of late night comedy. Harvard International Journal of Press/Politics, 8, 118–133.

33.

Niven

Lichter

S. R.

Amundson

(2008). Our first cartoon president: Bill Clinton and the politics of late night comedy. In Baumgartner

J. C

Morris

J. S.

(Eds.), Laughing matters: Humor and American politics in the media age (pp. 151–170). New York, NY: Routledge.

34.

O’Rourke

Lakner

(1989). Gender bias: analysis of factors causing male underrepresentation in surveys. International Journal of Public Opinion Research, 1, 164–176.

35.

Porter

S. R.

Whitcomb

M. E.

Weitzer

W. H.

(2004). Multiple surveys of students and survey fatigue. New Directions for Institutional Research, 21, 63–73.

36.

Powell

(1978). Topic salience and responses to the source of satirical messages. Southern Speech Communication Journal, 44, 60–72.

37.

Slack

(2012). RIP positive ads in 2012. Politico.com. November 4, 2012. Retrieved December 3, 2012, from http://www.politico.com/news/stories/1112/83262.html

38.

Smith

Duggan

(2012). Online political videos and campaign 2012. Pew Research Center’s Internet & American Life Project. November 2, 2012. Retrieved January 21, 2013, from http://pewinternet.org/Reports/2012/Election-2012-Video/Key-Findings.aspx?view=all

39.

Sonner

B. S.

(1998). The effectiveness of negative political advertising: A case study. Journal of Advertising Research, 38, 37–42.

40.

Stampler

(2012). Obama spent more on online ads than it cost to build the Lincoln Memorial. Business Insider. November 5, 2012. Retrieved December 22, 2012, from http://www.businessinsider.com/infographic-obama-romney-final-ad-spend-2012-11

41.

Sternthal

Craig

C. S.

(1973). Humor in advertising. Journal of Marketing, 37, 12–18.

42.

Tinkham

S. F.

Weaver-Lariscy

R. A.

(1993). A diagnostic approach to assessing the impact of negative political television commercials. Journal of Broadcasting & Electronic Media, 37, 377–400.

43.

Weinberger

M. G.

Gulas

C. S.

(1992). The impact of humor in advertising: A review. Journal of Advertising, 21, 35–59.

44.

West

D. M.

(1993). Air wars: Television advertising in election campaigns, 1952-1992. Washington, DC: Congressional Quarterly.

45.

Work: Obama for America. 2012. Blue state digital. Retrieved December 22, 2012, from http://www.bluestatedigital.com/work/case-studies/barack-obama