Abstract
This study explores whether visual presentation style (i.e., camera shot variation) may influence perceptions of Presidential candidates during televised debates. In a field experiment, 341 randomly assigned college students viewed or listened to one of seven broadcast network or cable/satellite channel conditions during the first televised Presidential debate between Democratic nominee Hillary Clinton and Republican nominee Donald Trump. A feeling thermometer and 19 leadership and personal traits were assessed during the time leading up to the debate and again immediately afterward. No differences were found based on presentation style, likely reflecting lack of variance in visual differences between the networks and broadcast channels. However, participant political ideology was significantly related to postdebate feelings toward each candidate. Judgments of Clinton’s professional and personal traits increased as was also the case albeit less significantly for Trump. Reasons for these changes in perceptions of the candidates following the debate are explored.
Candidate debates in the United States date back to at least 1858 when Abraham Lincoln and Stephen A. Douglas traversed the seven Illinois congressional districts speaking in front of hundreds of citizens with the debates often lasting more than 3 hours. In just over a century and a half, audiences to all forms of political competition have exponentially increased due to mass media exposure through television and the Internet. Likewise, the influence of television on voter’s attitudes toward Presidential candidates has grown steadily since the early 1960s as candidates and political advisors discovered its effectiveness in reaching voters through paid programming as well as through relatively unmediated unpaid programming such as interviews, speeches, and debates. Of these, presidential debates are perhaps the most influential because they present a relatively unmediated communication in a competitive context.
Debates are important political events because a great majority of voters often lack a deep understanding of critical issues and candidate positions (Shaw, 1999) and therefore rely on media to understand the political agenda (McCombs & Shaw, 1972). At the same time, the world has arguably become more complex policy-wise and citizens may have diminished time or abilities to evaluate their candidates and the positions they espouse. Although citizens tend to defer to their partisan preferences (Benoit, 2013), their opinions may also be influenced by contemporaneous events. Issues such as economic downturns or external threats may lead to voters discounting partisanship to decide on the candidate best equipped to address the issue. Beyond this, candidates have three primary ways to make their points: acclaims, attacks, and rebuttals (Benoit, Blaney, & Pier, 2000). Debates are ideal venues to take advantage of these mechanisms.
The 2016 Presidential campaign featured two very different but well-known entities. While the eventual Republican Party nominee Donald Trump was an unlikely outsider, he was already famous for being an outspoken billionaire real estate mogul and reality television show host that had only dabbled in politics with his “birther” criticisms of President Barack Obama. On the Democratic side, early favorite Hillary Clinton was already a national political fixture known for her obsessive attention to political and policy detail. As former First Lady, New York Senator, and Secretary of State, Clinton stood out on her record in politics and for being the first female presidential candidate from a major political party.
As a result, the 2016 presidential debates were highly important yet arguably aberrant due to the nature of this exceptional election. While observers have noted that the steady stream of political advertising and the 24-hour news cycle can overwhelm (Tedesco & Kaid, 2000), the general election debates provide the American public perhaps the only mechanism available to obtain relatively unmediated, and hence trustworthy, information. In other words, neither stagecraft nor “pitch-and-spin” (Norton & Goethals, 2004) can obscure candidate characteristics in the relatively unpredictable environment of a debate. Therefore, debates can be particularly important in close races such as the 2016 presidential election because voters, especially those remaining undecided, evaluate candidates both on their leadership and personal traits.
Believing that a candidate is qualified, intelligent, or competent would suggest that voters perceive that she or he has leadership traits allowing not only the ability to direct an effective legislative agenda but also be capable of dealing with and thinking through the complex policy issues that often arise unexpectedly and must be dealt with competently. Voters may also sense that a candidate is caring, sincere, and honest, traits related to expectations that the candidate will feel compassion and be trustworthy. Successful campaigns develop appeals relating to both the hearts and the minds of the voters (Brader, 2006) by highlighting a range of positive leadership traits (Abelson, Kinder, Peters, & Fiske, 1982; Kinder, Peters, Abelson, & Fiske, 1980).
In this article, we consider whether differences in visual presentation style (i.e., how the broadcast networks and cable/satellite channels presented the content in terms of visual imagery) influenced perceptions of the candidates during the debates after statistically controlling for the effects of political ideology (i.e., liberal vs. conservative) and predebate attitudes.
Visual Presentation Style
Although the debates were carried on numerous media outlets including broadcast networks and cable/satellite television channels, the content in terms of dialogue is uniform. However, visual presentation style decisions are made by program directors at each cable channel or network. So, while audience members may hear the same debate on all outlets, what they see may vary depending on the channel or network selected. Technical directors choose among camera shots that may convey different information although the audio remains the same across channels. This is important because as Messaris (1994) has argued,
images reproduce many of the informational cues that people make use of in their perception of physical and social reality. Our ability to infer what is represented in an image is based largely on this property, rather than on familiarity with arbitrary conventions. (p. 165)
In other words, whether audience members (or for that matter, technical directors) are aware of it, shot selection within a debate focuses attention on certain aspects of the participants’ performance.
Previous research suggests that camera angle and shot variation can have an impact on audience perceptions of the candidates (Cho, Shah, Nah, & Brossard, 2009; Gong & Bucy, 2015; Shah, Hanna, Bucy, Wells, & Quevedo, 2015). Wicks (2007) found that C-SPAN and PBS were more prone to use split screens during the third 2004 presidential debate, while the other networks were more likely to employ numerous switched camera shots. Specifically, viewers may be influenced through the reaction shots of the candidates (Gong & Bucy, 2016; Seiter & Weger, 2005; Seiter, Weger, Jensen, & Kinzer, 2010) as well as that of the host and the audience (Haumer & Donsbach, 2009; Nabi & Hendriks, 2003).
The Present Study
In the first debate in 2016, ABC provided the camera shots that were utilized by all of the channels or networks. 1 Technical directors at each network selected from a wide range of shots (e.g., wide, medium, and close-ups as well as audience reaction shots, etc.). The methodology employed in this study is similar to Wicks’s (2007) field experiment which examined audience response to the third presidential debate between President George W. Bush and Senator John Kerry during the 2004 election. However, the present study significantly increased the number of participants to enhance power. In addition, the number of broadcast networks and cable/satellite channels was increased to evaluate the potential for different visual presentation styles.
The study of visual presentation styles utilized in the 2004 election revealed that participants felt more positively (i.e., warmer) on a feeling thermometer toward John Kerry following his debate performance on C-SPAN and PBS, which featured the split screen also known commonly as the podium watch format (Wicks, 2007). Furthermore, although C-SPAN viewers perceived Kerry more favorably after the debate, they did not perceive him as more friendly. Perceptions of Bush’s strength declined for viewers of each network from before the debate to after the debate. Analysis revealed a significant difference between viewing the debate on C-SPAN with viewers perceiving him to be weaker following the debate. Hence, participants watching the split screen format appear to have come away from the debate with different impressions and perceptions of the candidates than did viewers of other formats.
However, the electoral contexts of 2004 and 2016 could not be more different. The 2004 presidential election was a relatively low-information affair with traditional candidates and an incumbent president. On the other hand, the 2016 election between Trump and Clinton was between two historically groundbreaking and very different candidates in a charged political environment. Because the situations between these two elections were so different, and because it was unclear to what extent networks would vary in visual production style, we looked not only at this variable but also political ideology and predebate attitudes of the participants as noted earlier.
Research Focus
Differences in debate presentation style may influence how voters perceive candidates. A large body of scholarship has addressed the impact of images on the impression of voters with respect to selecting a President (Grabe & Bucy, 2009; Jamieson & Waldman, 2004; Lammers, 1982; Stewart, Salter, & Mehu, 2009). Likewise, significant research has demonstrated that visual presentation style such as edits and cuts can influence processing of media messages (Lang, Geiger, Strickwerda, & Sumner, 1993; Lang, Zhou, Schwartz, Bolls, & Potter, 2000; Wicks, 2007).
Understanding the influence of presentation style has become an important concern in the past two decades as presentation techniques have progressed and evolved. Between 1960 and 1996, broadcast networks and cable channels typically presented a common feed with identical images and camera shots to viewers. With addition of Fox News and MSNBC in 1996 as alternatives to CNN, these new cable channels began looking for ways to differentiate themselves. By 2004, these and existing media outlets began using their own technical directors who selected various camera angles, reaction shots, wide shots, close-ups, and split screens. Thus, while debate content remained constant, visual presentation style varied among broadcast networks and cable/satellite channels potentially leading the formation of different perceptions of candidates.
A growing body of research has endeavored to understand the influence of different visual presentation style on viewer perceptions of Presidential candidates during debates. Analysis of visual presentation style during the third Bush–Kerry debate in 2004 suggested significant presentation style variation among broadcast networks and cable channels (Wicks, 2007). As noted earlier, this study was intended to understand more about how citizens may receive differential messages in a multichannel environment by employing more participants to enhance explanatory power (see Figures 1-2). Given the distinctive differences between these two candidates, the 2016 election seemed an optimal time to expand this inquiry and led to the primary research question.
Example of split screen or podium watch format used throughout the debate.
Example of switched screen image presenting both candidates sparing during debate.
Method
Research Design
A field experiment was conducted at the University of Arkansas, Fayetteville, to study the first debate on September 26, 2016. The 90-minute debate was sponsored by the Commission on Presidential Debates and was held at Hofstra University. It was moderated by NBC News anchor Lester Holt, and focused on achieving prosperity, America’s direction, and securing America. The candidates answered specific questions on jobs, race relations, taxes, and the prospect of cyberattacks. ABC News provided the pool cameras for the event, but each network was free to select from the range of shots and camera angles provided.
Participants and Procedure
Participants were recruited from approximately 2,000 undergraduate students in more than 100 communication, political science, and psychology course sections. Students were contacted through in-class announcements and e-mails in late August and early September 2016 and asked to take part in a study about the 2016 Presidential debates. Participating students in this opportunity sample received extra credit in their respective courses.
A total of 610 participants filled out an online predebate survey prior to the debate (between August 29 and September 26, 2016) and were randomly assigned to one of seven rooms, each presenting a different network condition (ABC, Fox News, MSNBC, 2 CNN, CBS, C-SPAN, and NPR, 3 each room ranged from 42 to 57 participants). To ensure random assignment, student identification was verified when participants reported to their assigned room. Participants were asked to bring beverages and were provided free snacks as they entered their assigned rooms. Immediately prior to the debate, moderators read instructions to the participants asking them to remain seated and refrain from using cell phones or other electronic devices during the debate. Immediately after the debate, participants filled out a pencil-and-paper postdebate survey, then were thanked and dismissed. Due to cross-departmental collaboration on this opportunity for data collection, both the predebate and postdebate surveys contained a battery of questions for studies unrelated to the present research. Only the measures germane to the present study will be discussed herein.
The debate was viewed by 362 participants. 4 Of these, 21 participants did not properly label their postdebate survey (i.e., could not match predebate and postdebate data), leaving a usable sample of 341 participants. The participants were composed of 217 females and 123 males. 5 The mean age was 19.53 and ranged from 18 to 53 years with only 10 participants older than 24 years. The racial composition was composed of 283 Caucasians, 22 African Americans, 15 Hispanics, 9 Asians, 4 Native American, with the remaining 7 self-identifying as “other.”
Approximately 77% reported being registered voters with 173 self-identifying as Republicans, 94 as Democrats, and 75 as “other” (including no party affiliation). Aside from categorical party identification, participants also reported their overall, social, and economic political ideologies (separately) on a 7-point Likert-type scale (1 = very liberal; 7 = very conservative). Each ideology type was normally distributed with overall and economic ideologies both slightly right of center (Ms = 4.30 and 4.75, respectively) and social ideology slightly left of center (M = 3.83). Chi-square (sex, party, race) and analyses of variance (age, ideology) show no significant difference in demographics between rooms (psex = .59; pparty = .29; prace = .85; pideology = .15; page = .16).
Measures
On both the predebate and postdebate surveys, participants were asked to rate their favorability toward the candidates using a feeling thermometer with a scale ranging from unfavorable/cold (0) to favorable/warm (100). Participants also evaluated the candidates on 19 traits using a 7-point Likert-type scale anchored with 1 (not at all) and 7 (extremely). The traits were qualified, sophisticated, honest, believable, successful, attractive, friendly, sincere, excitable, aggressive, 6 strong, active, competent, intelligent, caring, trustworthy, agreeable, generous, and humorous. For each candidate, reliability among the 19 items was sufficiently high (αs > .95) to average the items into a composite trait index that acts as a single measure indicating an overall judgment of the candidates’ traits (i.e., the primary dependent variable for Research Question 1B). Finally, as noted above, demographics included participants’ overall political ideology.
Results
Analytic Strategy
One-way analyses of covariance were conducted to assess if production techniques chosen by the different networks had a significant impact on postdebate feeling thermometer ratings and trait judgments, statistically controlling for participants’ overall political ideology and predebate attitudes toward the candidates (i.e., predebate feeling thermometers and trait ratings). For each feeling thermometer analysis, the respective candidate’s predebate feeling thermometer was entered as a covariate; likewise, for each trait index analysis, the respective candidate’s predebate trait rating was entered as a covariate. By using this approach, the analysis asks the theoretical question, “Did a person watching on Network A change his or her attitude more than a person watching on Network B, given he or she had the same initial attitudes/ideology?”
Visual Presentation Style (Research Question 1A, Research Question 1B)
The research question explored the potential influence of visual presentation style on favorability toward and/or trait judgments of the candidates. For both Hillary Clinton and Donald Trump, favorability toward the candidates did not significantly differ by network, Clinton: F(6, 269) = 1.18, p = .32, η2 p = .03; Trump: F(6, 264) = .46, p = .82, η2 p = .01 (Research Question 1A). The trait index was analyzed using the same approach as the feeling thermometer analyses and also revealed no significant difference between networks, Clinton: F(6, 325) = 1.31, p = .25, η2 p = .02; Trump: F(6, 325) = 1.20, p = .31, η2 p = .02 (Research Question 1B).
The political ideology covariate was significant for postdebate feelings toward both Clinton: F(1, 268) = 22.85, p < .001, η2 p = .079 and Trump: F(1, 263) = 4.79, p = .029, η2 p = .018. Additionally, political ideology was a significant covariate for postdebate trait ratings for Clinton: F(1, 324) = 15.14, p < .001, η2 p = .045 and was marginally significant for postdebate trait ratings of Trump: F(1, 324) = 3.23, p = .073, η2 p = .010. As expected, liberal-leaning participants were more favorable toward Hillary Clinton and less favorable toward Donald Trump than conservative-leaning participants (and vice versa), and the same direction of patterns were found for ratings of the candidates’ traits.
The predebate attitudes covariate was significant for postdebate feelings toward both Clinton: F(1, 269) = 95.08, p < .001, η2 p = .262 and Trump: F(1, 263) = 32.48, p < .001, η2 p = .110. Additionally, this covariate was significant for postdebate trait ratings of each candidate, Clinton: F(1, 324) = 127.32, p < .001, η2 p = .282; Trump: F(1, 324) = 339.47, p < .001, η2 p = .512.
To assess if any trait(s) in particular drove changes (or lack thereof) in overall trait ratings, we analyzed the individual traits in the same fashion as the trait indices. When considering the effect of visual presentation style on traits judgments of Hillary Clinton, there was only a significant difference between networks (Research Question 1B) for the trait successful (η2 p = .038). 7 Post hoc pairwise comparisons revealed only a marginally significant difference between NPR and C-SPAN for this trait. 8 For Donald Trump, the only significant main effects for network (Research Question 1B) were for the traits excitable, active, and humorous (η2 p s = .041, .037, .051, respectively). However, pairwise comparisons revealed no significant differences between networks that may offer insight into these findings.
Check on Variability in Presentation Style Across Networks
We anticipated differences in attitudes toward the candidates as a function of differences in visual presentation style between the networks. However, examination of the six media outlets investigated during the first debate revealed very few differences in presentation styles; side-by-side split screen format was used 83% to 100% of the debate by all networks and broadcast channels. Computations of z scores for the proportion of split screen time to total debate time for each network showed that these differences were not significant (−1.32 > z < 1.18).
Discussion
In this study, we sought to discover whether visual presentation style of the debate would influence perceptions of candidate favorability and their traits by audience members. The study was guided by previous research revealing that visual presentation styles taken by different networks during the 2004 presidential election were capable of influencing audience member perceptions (Wicks, 2007). In other words, with this study, we engaged in a conceptual replication with the understanding that not only would candidates obviously differ from the 2004 debate and with them the types of verbal and nonverbal behavior encountered by viewers but also that network visual presentation style would exhibit production differences.
Perhaps due to the then relatively recent expansion of broadcast and cable outlets that began in the late 1990s and concomitant competition for viewers, greater variation in presentation style occurred in 2004. Despite each network or channel having similar options as was the case in 2004, directors in 2016 opted primarily for the side-by-side split screen format (see Figure 1). In turn, programs were generally indistinguishable, with the exception of electronically generated network identifiers, which in turn was reflected in the lack of differences between the rooms for viewer perceptions. What few individual main effects were statistically found were likely a function of chance.
Unsurprisingly, political ideology played important roles in favorability and trait judgments of the candidates. Hillary Clinton represented the liberal part of the political spectrum while Donald Trump offered himself as the conservative option. Trump routinely espoused policies typically not held by the conservative establishment including building a costly Southern border wall to prevent illegal immigration, attacking free trade in order to return manufacturing jobs from China and Mexico and using executive powers to jail his opponent Hillary Clinton for allegedly purging e-mail documents from a private server she had maintained as Secretary of State. While some of these positions would generally be considered antithetical to both the liberal and conservative establishments’ positions on the rule of law, they proved to be lively debate talking points. This in turn was reflected in the influence political ideology had on favorability ratings and trait evaluations for Hillary Clinton in comparison with Donald Trump, as participant political ideology was substantially more powerful in evaluating Clinton than Trump.
In sum, while visual presentation style differences appear to have had little influence, the personal beliefs represented by political ideology appear to influence candidate evaluation. Further research should focus on specific media measures such as time participants spent watching cable news, reading online news, and using social media to isolate these factors’ individual effects on prepresidential and postpresidential debate judgments. Finally, future research may seek to isolate situations in which visual presentation styles do vary to determine whether the findings reported in previous research (e.g., Wicks, 2007) represent a fruitful path for future researchers.
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.