THE Power of the poster
An experiment to the memory enhancement
and attitudinal effects of political posters
By Thom Rietberg - 6151841
Master's Thesis - Graduate School of Communication
Supervisor: Dr. Linda Bos
Abstract
Introduction
Ever since the, arguably most important, invention of the printing press was introduced by Johannes Gutenberg, marketing has gotten the opportunity to be more widespread than ever. Not longer were corporations and elitist organizations bound to word of mouth for gaining popularity. While all printed material was initially covered with ty-pography, this later shifted to more graphical imagery in the late 19th century. Painters and graphic designers started to create non-fictional objects, like trains and persons, in which art and marketing overlapped. Among the clients of these poster we could find military organizations, theaters and political parties. All with the same goals in mind: in-fluencing the viewer into a change of action or thought. Even nowadays, with emerging advanced technologies as social media and personalized advertisements, printed images are still everywhere to be seen. With every upcoming election new posters, leaflets and banners are spread to each town and city. These forms of marketing look to be tied with
Even though marketing technologies evolve rapidly, political parties still use paper posters during elections. Elaborating on previous literature about political visuals, in this study a closer look will be taken at the effects of a poster. By separating three main cues; a photograph, a logo and a slogan, an attempt to a better understanding of these elements is developed. Both the single effect of the cue is analyzed as well as the interaction between cues. In an experimental study, respondents will be randomly assigned to one of the seven conditions. Each condition has its own political poster with an unique set of cues. Using knowledge and attitudinal questions the effects of memory enhancement and attitudes due to the cues are measured. Results show that having a photograph of the politician always improves the attitude, both toward the politician and the party. Adding a logo to the poster causes the memory enhancement of the political party to increase. All in all, the three cues differ in the effects they cause. It is up to the political party to decide what goal they want to achieve and to add the cues appropriate to the goal. These results could serve as a strategic media advice for political parties or function as a guideline for future research to the effects of visuals.
politics. But why is there still value in text and image on a paper? Does it cause effects to its viewers? And if so, does it cause the intended effects?
In this master thesis I will take a closer look at political imagery. By conducting an ex-periment with several conditions, I will attempt to answer what kind of effects political visuals have on its viewers. I will mainly focus on the change of memory enhancement and attitude after seeing a political poster. This unique combination of both studying the memory enhancement and attitudes, could create a more complete answer to in what way visual elements work together best. This develops valuable information for political parties, but also for the marketing of companies or the receiver of the message. Knowing in what way companies try to influence you, can create an awareness about how vulner-able a person can be.
Designing an experiment like this means not only the field of communication and po-litical communication will be discussed. Results also tell more about the human psyche; how we react to certain elements on a daily basis and what the underlying consequences are. There is still much to explore about effects we endure everyday and going beyond the field of communication and look at psychology gives us great insights into these ef-fects. But the interdisciplinarity does not stop here, since we take an in-depth look at the print media, a media strategy advice for political parties can be made. This might create knowledge that could be used in media and marketing studies.
Academic literature on memory enhancement and attitudinal changes by visuals is plen-tiful, especially in the early eighties a vast landscape of articles has been developed. How-ever, if we look at the effects of political visuals on memory enhancement and attitude changes, the amount of literature leaves much to be desire; this combination has simply not been explored yet. There is a noticeable gap in the research when combining certain effects of visuals and by using these two, much used, effects we can add knowledge to this gap. Two of the cues used in this experiment, a photo and a text, have been research be-fore, but about the third cue, the logo, is still a lack of information in academic literature. I would argue that a logo is a very valuable brand identification and brand recognition visual for viewers and might have a big impact on the recall and attitude towards parties.
In this paragraph a theoretical framework about political visuals is provided, notably with the addition of the logo of the political party as visual. It will start by defining po-litical visuals and the historic development of these communicational cues and will end by summarizing the known effects of political visuals described by other researchers.
Political visuals
There are different forms of convincing the voter of the justness of a political party. One may be drawn to substantive debates, while others are better persuaded by a party pro-gram. Another form we have seen for about a century (Vliegenthart, 2011) are political visuals, usually published with a favorable photograph or text by its party (Boen & Van-beselaere, 2002). However, political visuals can also be directly targeted against other parties (Smith, 2001). One important form of these visuals are the posters (Dumitrescu, 2012). Taking a 20 per cent share of the campaign budget in most European countries, these visuals play a substantive role in the political communication of parties. But also other forms, like leaflets (Rosenberg, Bohan, McCafferty & Harris, 1986) or banners (Tschabrun, 2003) are ways to visually communicate with the voter. So, even in a day and age where interactive and personalized message can be created, paper posters keep to be prevalent. Usually, these visuals are displayed on designated areas in a city or to be found on a web site (Petchesky, 1987; Vliegenthart, 2011; Prior, 2014;). The visual
can contain one or multiple elements or cues, like the politicians’ face (Banducci, Karp,
Thrasher & Rallings, 2008; Prior, 2014), the logo of the political party (Bennett, 2004; Van Ham, 2001; Vliegenthart, 2011) or a textual cue, like a call to vote or a slogan (Van Ham, 2001; Tschabrun, 2003; Vliegenthart, 2011). These elements solely or combined try to persuade the viewer into voting (Boen & Vanbeselaere, 2002; Tschabrun, 2003).
Theoretical Background
Combining the gap in the literature with the continuous salience of political posters the following research question arises:
To what extent do characteristics of political advertising change the knowledge and at-titude of the receiver?
Change in visuals
As mentioned in the previous paragraph, political visuals have always played a role in the communication of political parties. Historically, visuals like posters and flyers have always been rather cheap and local instruments for the advertisement of politi-cians (Tschabrun, 2003). But since the invention of the television, political marketing has been professionalized (Blumler and Kavanagh, 1999; Brants and van Praag, 2006), shifting the political communication from political logic to a media logic. In here, the mass media decides what topics will be addressed on the agenda. Nowadays, some re-searchers suggest that we have shifted to another era, the post-modern political commu-nication. Strömbäck (2007) points out that this new era, the marketing logic, consists out of specialized campaign teams that are able to broadcast personalized messages to voters. Indeed, in an increasing amount of countries we see highly professionalized, per-manent campaigns (Blumler and Kavanagh, 1999; Strömbäck, 2008). However, some might argue that these 24/7 campaigns are the case in Anglo-Saxon countries like the United States and the United Kingdom, but not yet to be seen in smaller Western Eu-ropean countries, like The Netherlands and Denmark (Farrell & Webb, 1998). But it looks clear that the political landscape is shifting to be more professionalized. By having dedicated teams, we see more efficient and better thought out campaigns.
Strömbäck (2008) also mentions an ever growing personalization in politics. This means, relative to the party, more attention will be given to politicians and they have more aware how they should look and talk. This professionalization and personalization also carries out to the political advertisement; the political leader is presented more often (Iyengar, Norpoth & Hahn, 2004; Sellers & Schaffner, 2007; Vliegenthart, 2011), the logo of the party is used more frequently (Farrell & Wortmann, 1987; Bennett, 2004) and the message of the text is seperate from the message of the image, rather than both having a complementary message (Vliegenthart, 2011).
Memory enhancement
When discussing the effects of political visuals, one of the most common is the mem-ory enhancement when a visual is seen. Multiple research findings show that when a face is presented, the chance of recalling the name of that person and what he or she
does is significantly higher (Cohen, 1993; Dumitrescu, 2012; Prior, 2014), and even more when the face of that person is similar to that of the viewers (Schill, 2012). One explanation could be that a viewer has greater emotional involvement when looking at a photograph and is thus more drawn to the topic (Graber, 1996; Arpan, Baker, Lee, Jung, Lorusso & Smith, 2006). From previous studies we know that pictures are stored longer in the brain than verbal or textual cues (Lang, Potter & Bolls 1999; Prior, 2014) and that they aid to recall some information later in time (Graber, 1996; Schill, 2012), particularly when these images guide a more dramatic or shocking subject (Graber, 1990; Newhagen, 1998; Smith, 2001), we can assume that extreme visuals might stand out more than other ‘normal’ visuals. When a combination of texual and visual elements is used, like slogan and a logo of the political party, the memory enhancement is regard-ed as stronger (Cohen, 1993). This idea is executregard-ed by an experiment of Prior (2014), where respondents had to answer the occupation of several European leaders when only given the name of that person or the photo or a combination of both. Results suggested that the combination of textual and visual cues increases the number of correct answers. This idea of a stronger effect by combining more cues is not new; Kisielius & Sternthal (1984) and Joffe (2008) have found similar results with the combination of pictures and text.
The previous mentioned findings, especially those addressing the higher memory en-hancement due to visuals, lead us to the following hypotheses:
H1a: The image of a political candidate improve political party memory enhancement to a larger extent then textual cues.
H1b: The logo of a political party improve memory political party enhance ment to a larger extent then textual cues.
H1c: The combination of having both a text and a logo improves political party memory enhancement to a larger extent than the sole effect of these elements.
Conditional effects of memory enhancement
As we encountered in the previous paragraphs, different elements in the poster can have different kinds of effects on the viewer. However, it might be wise to look at subsets of the population as well. Prior’s (2014) findings lead to believe that women, in regard of memory enhancement, benefit more from visual cues than men do. And indeed, other researchers find similar results, Greene and Fraser (2002) have strong evidence that women are able to recognize faces faster than men do and according to Kirouac and Dore (1985) women score significant higher in decoding facial expressions of emotions. Barrett and Barrington (2005) argue that women tend to change their attitudes towards a politician more positively when a photograph is shown, especially when this photo-graph is favorable for the candidate (i.e. attractive). Other articles, like that of Kirouac and Dore (1985) suggest that further research should be focused on the difference be-tween men and women regarding to facials visuals, but for now it looks that women are better at decoding the emotions when presented with a facial stimuli.
Other characteristics of the population may a play part in the effects of political visuals. For instance, political knowledge is highly related with educational levels (Delli, Carpini & Keeter, 1996), but this gap levels out when the respondent is provided with a picture
(Prior, 2014). An explanation could be that lower educated people mainly see political
actors on television, while the higher educated read and talk about these actors more often. However, Kirouac and Dore (1985) found no evidence that a higher educational level leads to a better recognition of faces. Still, given the fact that lower educated see politicians more often than read about them, I would expect seeing a face of a politician is more beneficially for lower educated than for higher educated.
In the following hypotheses we will find out whether women and lower educated respon-dent benefit more from photographs in political communication than men and higher educated respondents do:
H2: Women gain more political knowledge recall by visual cues than men. H3: Respondents with lower education gain more political knowledge recall
Attitude change
When a political poster is designed, a political party intends to persuade the viewer into developing a more favorable image of that political party or politician which might lead to a change in voting decision (Strachan & Kendall, 2004). But before a change in behaviour is made by the viewer, he or she may develop an attitude based on the visual by processing the image (Blair, 2004). Attitude can be defined as the direct internal valuation process of a person or party (Mitchell & Olsen, 1986). This processing can be systematic (factual) or heuristic (making shortcut decisions) (Hill, 2004). Using the Elaboration Likelihood Model (Petty & Cacioppo, 1986) we can arrange these two ways of processing into a central path and a peripheral path. In the central path, the viewer has to be convinced by arguments, since the person is not distracted by external influ-ences. On the other hand, when the viewer has little to no interest while being exposed to the message, the peripheral path will be taken. Messages should then contain more superficial information, like the face or voice of a politician or a logo of the party. This is true for print media (Dumitrescu, 2011) and for online media (Riegelsberger, Sasse & McCarthy, 2003).
Visuals are particularly powerful in influencing short term attitudes (Huckfeldt & Sprague 1992; Schill, 2012), surprisingly more than the actual policies of the parties are (Smith, 2001). One example of the power of visuals over the power of text is the experiment by Rossiter and Percy (1980), they showed respondents various marketing posters for a beer brand. Some of them were very visual with just one or two words of text, whereas others showed only text. Results pointed out that the highest attitudes were formed after seeing the largest visual (image of a glass of beer). When it comes to attitudinal change in politics, different elements can affect the viewer. A common effect is when the shown candidate is regarded as attractive, the viewer rates that person as a more credible politician (Rosenberg, 1986; Graber, 1996; Smith, 2001; Banducci, 2008; Schill, 2012). Barrett and Barrington (2005) did not only measure the amount of increased positive attitudes when the respondent saw attractive candidates, he also asked them whether or not they would vote for this party (change in behaviour). Results pointed out strong evidence that a favorable looking candidate has a higher likelihood of votes than an unfavorable looking politician.
It seems that academic articles agree on the fact that showing (attractive) faces or a logo instead of textual cues with argument leads to a more positive attitude towards the party or politician.
H4a: The visual cue of a face leads to a more positive attitude towards the political party than a textual cue.
H4b: The visual cue of a logo leads to a more positive attitude towards the politician than a textual cue.
Conditional effects of attitude
In numerous studies the interaction between age and political cynicism has been stud-ied (Agger, Goldstein and Pearl, 1961; De Vreese, 2005; Zhang, Johnson, Seltzer & Bichard, 2009), each with the same outcome; when elderly respondents are asked about politicians, their political trust decreases. This might lead us to believe that the face of a politician has a different effect on the attitude towards political parties of elder respon-dents than on younger responrespon-dents. Although there is no enough evidence in regard to a political poster, I do want to research this interaction. Leading to create the following research question:
RQ1: When confronted with a image of a politician, the attitude towards a political party of older respondents decrease more than that of younger respondents.
Design of experiment
In this experiment a 2x2x2 between-subject design is used. But since a control poster (a poster with none of the elements) would not make sense, a total of seven conditions will be designed. The seven conditions will be randomly assigned to the respondents. The main effects that will be tested in this experiment are memory enhancement and atti-tude change by visual elements. The respondents will received questions based upon the images they have studied a short moment ago. These questions will measure the amount of memory enhancement and attitude change the cues could have caused.
Since memory enhancement will measure to what extent the viewer is able to recall previously gained knowledge, the visuals shown will be those of a national politician of a political party. This way, the chance that the viewer already knows who is shown is very likely. On the other hand, when measuring attitude change based on the political visuals, previous attitudes may have unwanted effects. It would be logical that the condition for these questions should involve a politician that is not too famous or, better, a fictional politician. For experimental purposes, the images will be modified using a computer program. This way, we are able to leave out certain details (like the name of the politi-cal party or the internet address) in order to measure just the effect that is researched. These images will be shown during an online survey where the respondent is asked to study the image as long as he or she likes, giving the respondent the opportunity to pro-cess the image fully. This tends to believe that the person will take the central path in the previously described Elaboration Likelihood Model.
For answering the hypotheses I will use three different cues, resulting in seven condi-tions. These cues are varied in the different conditions, in order to detect if one cues
Condition 1 Condition 2 Condition 3 Condition 4 Condition 5 Condition 6 Condition 7
Poster has a photo Poster has a text Poster has a logo
figure 2. a) Photo only, b) Photo and text, c) Pho-to, text and logo, d) Photo and logo, e) Text only, f)
Text and logo and g) Logo only Poster a Poster d Poster g Poster b Poster e Poster c Poster f
generates a stronger effect than a other cue. These cues (see figure 1) will be randomly assigned to the respondents. This also applies for measuring the attitude change. In fig-ure 2 the seven different conditions are presented.
Methods
Participants
A total of 261 respondents participated in this study of whom 213 finished until the last question, creating a sufficient amount of respondents for calculating levels of signif-icance. They were mainly recruited on social media outlets, such as Facebook, Twitter and Linkedin and executed the survey on their personal computer. The mean age of this group is 27 years old (M = 27,54, SD = 10.52) and more women (N=194) than men (N=62) participated. The survey has been filled in between the 23rd and 30th of June 2014.
Design and stimulus materials
This experiment uses a between-group design with three cues as stimuli. Since the ex-periment is a combination (or single use) of these three cues, participants were assigned to seven different groups. All questions were asked in Dutch, since the stimuli in the survey were gathered from Dutch politics. This means only Dutch participants were sought. The choice to recruit respondents largely via social media is obvious, since this is great way to reach a large amount of people with limited resources. However, this could bring some limitations regarding representativity since most respondents are in some way connected to the researcher. I expect that the sample is generally younger, higher educated and more politically engaged than average.
The link in the shared message, guided by the text to participate in the survey, leads to a survey made in Qualtrics. In here respondents were able to answer a questionnaire of sixteen to nineteen questions that lasted between 4 and 10 minutes, depending on their speed and condition they were assigned to. It is important to note that everything was shown to the participant on a digital screen, which is in contrast with the real world since only computer-owners participated in this experiment. However, we have no indi-cation that this may inflict with the results, since similar studies do not mention this may be harmful in any way (Prior, 2014).
Procedure
The first part of the survey was equal for all respondents, these included general demo-graphic questions like age (M = 27.54, SD = 10.52), sex (M = .76, SD = .22), educa-tion level on a 7-point scale (M = 6.09, SD = 1.11) and media usage (M = 14.68, SD = 4.58). Detailed information about the variable can be found in appendix 3. Besides that, questions were asked about general characteristic traits. The respondents had to answer these on a bipolar 7-point scale to what degree they see themselves as social (M = 5.54, SD = .92), tolerant (M = 5.36, SD = .981) and politically engaged (M = 4.43, SD = 1.54). We also asked what choice the participant would make when parliamentary elections were held today, this might have some correlation with the outcome, since a non-fictional party was used. Next, the Dutch political knowledge was measured using five questions with one correct answer for each question (M = 2.05, SD = 1.13). These questions were the same for every person attending the survey and are later used to detect any correlation with memory enhancement to test whether people with a higher political knowledge were able to recognize political actors better.
Dependent Variables
After this, the experimental part began. The respondents were randomly distributed over seven conditions. They would see a constructed poster with a single or combination of the following elements: an image of Lodewijk Asscher, the logo of the PvdA and the slogan of the PvdA. This resulted in seven different posters (see posters in figure 2). We selected Lodewijk Asscher of the PvdA because both attitude and knowledge of him in the Netherlands could be described as ‘neutral’ and ‘somewhat famous’. The goal of this poster was to simulate a real looking visual, one that the respondent could encounter in real life. The resemblance was close to the real world, since the same color, font-type and photographs were used as the PvdA would have. For the purpose of this research
other design aspects, like placements and size, are kept ceteris paribus among the poster.
These considerations are made based upon the design principles of the PvdA (pvda.nl). The slogan on the poster ‘Nederland Sterker & Socialer’ translates to ‘The Netherlands, Stronger and more Social’ and is the official slogan of the PvdA for 2014. Furthermore, the slogan frames the PvdA as a forward-thinking and social party.
The minimum amount of time a respondent had to study the image was six seconds, after this he or she could continue to the questions measuring the recall regarding what they saw. The poster with the photo of Lodewijk Asscher included the following ques-tions: “Who is this person?” with five answer possibilities and a ‘Don’t know’ option (M = .27, SD = .48), “What is the job title of this person?” with five answer possibili-ties and a ‘Don’t know’ option (M = .24, SD = .428) and “What is the political party depicted in this poster?” with five answer possibilities and a ‘Don’t know’ option (M = .62, SD = .47). The purpose of these questions is to what degree they are able trigger the political memory of the viewer. The poster without the image only included the last question addressing the recognition of the political party. After this, the character traits (attitudes) were measured. These personal traits have been tested before in numerous studies (Miller and Shanks, 1996) and are based on what elements voters judge can-didates. The questions involved to what extent the respondent rated the politician. The items were: “Reliable” (M = 3.26, SD = .77), “Knowledgeable” (M = 3.44, SD = .79), “Caring for people like me” (M = 2.98, SD = .85), “Integer” (M = 3.21, SD = .70), “Involved” (M = 3.32, SD = .75), “Authentic” (M = 3.09, SD = .72), “Convincing” (M = 3.19, SD = .81). The same traits are used for the attitude of the political party in the poster. Again, participants without the photograph in their poster only had to answer only the questions regarding the party.
For the questions that measured memory enhancement, participants would not see the poster guiding the questions, since they had to remember certain elements from their memory. Attitude questions however, were guided with the image, since we have learned from the previous discussed literature that attitudes are based upon what the person sees at that same time. After answering all questions, the respondents were debriefed. In the debriefing the purpose of the experiment was explained, noting that they have seen a random poster and that this image was fictional, although based on real world political visuals. If the respondent had any questions related to the experiment this could have been sent to the researcher.
The main effects that have been researched in this study are the memory enhancement and the attitude towards politicians and parties. The memory enhancement is measured
with knowledge questions that could be answered either right or wrong. Since this is a dummy variable, no factor analysis or Cronbach’s Alpha is needed. The fact whether the respondent could recall the politician could have been caused by the used experimental factors; the logo or slogan. This is the same case for the memory enhancement of the political party, with the addition of third main variable; the image of Lodewijk Asscher.
The constructed ‘attitudes’ are measured using seven 5-point scales, these exist out of numeric values between 1 (Totally agree) and 5 (Totally disagree). The reliability for the attitude towards the politicians scored a Cronbach’s Alpha of a=.902, likewise, the atti-tude towards the party scored almost equally high with a=.879. The observed internal variance (factor analysis) can also be described as sufficient, with both having an Eigen-value over the Eigen-value of 1. A complete overview of the items in the questionnaire with its scales, means and standard deviations is presented in appendix 3.
Analysis
After all the data was collected in Qualtrics, and the minimum amount of respondents were counted, the records were exported to SPSS, a statistical program. Data was re-moved for all subjects whom not had finished the full questionnaire (N= 48). In the sta-tistical program SPSS all data was merged and put together in new constructed variables. The final sample consisted out of seven group of which 29 had seen version a, 35 had seen version b, 29 had seen version c, 31 had seen version d, 26 had seen version e, 30 had seen version f and 44 had seen version g.
For each of the four dependent variables a One-Way ANOVA with a post-hoc Bonfer-roni test was executed. This analysis generates the means and standard deviations of the various groups, as well as the significance between groups. The main effects, covariates and interactions will be analyzed using a linear regression giving information about the constant value of the variable, as well as the effect of each variable. The interactions were manually constructed using the compute function in SPSS. Finally, the linear regression
Results
In this chapter the retrieved data will be analyzed. This will be done in order of the raised hypotheses; first the memory enhancements and second the attitudes. The results will be discussed in threefold. We start by presenting the mean differences of the seven main con-ditions, then based on the effects of the cues the levels of significance will be measured using an linear regression analysis. In here, the variance of the three main stimuli will be presented. This gives the opportunity to approve of disprove the hypotheses regarding the main effects. After this, the effects of the covariates and interactions will be analyzed, results needed for the hypothesis will be shown and briefly discussed.
Effects of visual elements on memory enhancement. The first set of hypotheses are about the
topic of memory enhancement. The answers measuring the recall were dichotomous and amounted to a score of one when correct and zero otherwise. Table 1 presents the main differences of the three main cues on both the memory enhancement of the politician and the memory enhancement of the political party depicted on the poster. As we can see in the overview table, the addition of a text or logo doesn’t influence the recall of a politician by a large amount. However, the combination of all three elements generates the highest score of .60 (SD=.43) out of the maximum score of 1. There were no between group significant levels found (F (6) = .055, p = .983).
table 1. Mean differences of the different conditions on memory enhancement of the political actor and political party. N=213. abcd signaling the between groups significancy
Memory enhancement
political actor Memory enhancement political party
Stimuli M (SD) M (SD) Text -- .52 (.51)d Photo .52 (.42) .54 (.51)cd Logo -- .90 (.30)a Text + Photo .52 (.48) .44 (.50)bcd Text + Logo -- .93 (.27)a
Photo + Logo .54 (.45) .85 (.37)acd
Regarding the memory enhancement of the political party, we can add the photograph as well since the photograph of the politician is not a constant factor. When we study table 1, surprisingly, three conditions deliver the highest score; logo (M=.90; SD=.30) and the logo in combination with text (M=.93; SD=.27) or a photo (M=.85;SD=.37). Re-spondents with a logo as stimuli in the poster recognized the party more often than when a logo lacked. Without any analysis measuring the significance, these results lead us to believe that the addition of a logo improves the memory of a political party. Furthermore, the combination of a text and photo lowered the recall of the party drastically (M=.44; SD=.50). As to be seen in table 1, a high number of groups differentiated significantly (F (6) = 7,227, p <.001).
The impact of each experimental element is analyzed using a linear regression. On av-erage, Lodewijk Asscher was slightly more recognized than not, corresponding with our
initial ‘somewhat famous’ thoughts (constant of b = .536, SE = .070). By adding various
elements to the poster, such as text (b = .005, SE = .079) and a logo (b = .006, SE =
.079) the general memory enhancement did not improve significantly. The proportion of
explained variance (R2) of these elements only accounted for a tenth per cent.
The baseline recognition of the party was also correct in more than half of the cases, with
a constant of b = .587 (SE = .073). Since all answered are measured in a scale to 1, the b
coefficients can be interpreted as means. Notable, by adding text (b = -.041, SE = .057)
or a photo (b = -.066, SE = .058) the change of recall of the party decreased, although not
Memory enhancement
political actor Memory enhancement political party
b SE b SE
Constant .536 .070 .587 .073
Text .005 .079 -.041 .057
Logo .006 .079 .334 .060**
Photo n.a. n.a. -.066 .058
R2 .001 .166
N = 213, * p < .05, ** p <.01.Coefficients are unstandardized.
table 2. Linear regression analysis of main effects on memory enhancement of the political actor and poltical party
significantly. Adding a logo however, significantly improved the recognition of the party (b = .334, SE = .060, p < .001) signaling the importance of a brand logo. The explained
variance of this analysis was higher than the previous, with 16,6%.
Hypothesis 1a predicted that having a photo of the face of the politician would create a higher political party memory enhancement than a textual cue. Studying the results, we can see that having a photo generates a slightly lower enhancement than text, thus sup-port for hypothesis 1a is not found.
In hypothesis 1b we assumed that a logo would have a stronger effect on memory en-hancement than a textual cue. Especially because logos are generally the symbol of the party, whereas a text can change each campaign or even within a campaign. There seems to be proof for this theory, a logo scores significantly higher on the memory enhancement of the party since more than 90% of the respondents were able to recognized the party correctly. Hypothesis 1b is supported.
From the theory we learned that combinations of cues can create stronger effects, espe-cially the text and the logo. Hypothesis 1c predicts a strong interaction between these cues. Although the text was one of the weaker cues on memory enhancement of the party, the combination with the logo indeed accumulated to the strongest effect; 93% of the respondent knew the party after they had seen the poster with the text and the logo. There seems to be enough support for Hypothesis 1c.
Taking the measured covariates into account, we see no change in the significance of the mean cues. This let us believe that, regardless of other variables, the main effects are a strong predictor. Since are experiment was randomized, we might assume that general variables as age and sex do not play a part in the effects. Looking further in appedix 1, one might argue that political knowledge, education and perceived political interest measure the same variable (Prior, 2014), however political knowledge seems to be have a more sig-nificant effect. Taking all variables into account, the explained variance is of the politican was 38,1% and that of the political party accumulated to 33,2%.
Hypothesis 2 predicts that women’s political recall is generally more improved when guided with visuals. When we analyze this interaction (counting both an image as a logo as visuals), we do indeed see an improvement of recognition. Even more telling, when
confronted with a text, men are 7,1% (b = -.071, SE = .031) more likely to answer the
party correct, but when confronted with a visual, women will answer this question 3,4% (b = .034, SE = .025) better. This effect is significant (p = .026). Thus hypothesis 2 is
approved. This change does not appear when we look at the recognition of a politician; men are in both cases more likely to answer correctly.
According to hypothesis 3 the interaction between lower education and visuals causes a higher memory improvement than with high education and visuals. If we split the low-est four educational levels from the highlow-est three levels, this interaction only grows in favour of the higher educated. Memory enhancement change of the politician is 1.4% between low and high educated and the change of the enhancement of the party is 3.1%, even more so, this last effect is significant. These results lead us to believe that visuals in combination with low education does not improve the recall of a political actor of party. Hypothesis 3 is not supported.
Effects of visual elements on attitudes. The mean attitudes on the politician’s characteristic
traits and on the political parties’ characteristic traits are presented in table 3. Comparing the seven different conditions measuring the attitude towards the politician it seems clear
Attitude towards person Attitude towards party
Stimuli M (SD) M (SD) Text -- 20.04 (4.85) Photo 25.04 (3.69)a 21.88 (3.58) Logo -- 19.48 (4.77) Text + Photo 21.44 (4.99) 20.91 (4.53) Text + Logo -- 22.04 (4.33) Photo + Logo 22.38 (4.08)ab 21.69 (4.60)
Text + Photo + Logo 21.52 (3.12)ab 19.64 (5.25)
table 3. Mean differences of the different conditions on attitudes toward the political actor and political party. N=213. ab signaling the between groups significancy
that just the photo generates the highest attitude score of 25.04 (SD= 3.69) based on a maximum score of 35. Significance between groups was present (F (3) = 5.076, p = .002)
Remarkably, in third column the results show that the attitude towards a party is the oth-er way around. In hoth-ere, not the photo but the combination of text and the logo genoth-erates the highest score (M = 22.04, SD = 4.33), albeit not as extreme differences seen in the results of the attitude towards the politician. Only showing the logo of the PvdA evokes the lowest attitudes (M=19.48, SD= 4.77). The same can be said for having all the cue on the poster (M= 19.64, SD = 5,25). Although groups differ in attitude score no signif-icance between group was found (F (6) = 1.810, p = .099).
The analysis of change in attitude shows the effects of the main cues combined with the political actor and political parties (table 4). In here, it is clear that having a photograph of the politician does influence the attitudes. Adding a text or logo always decreases the attitude towards the politician. In the case of the addition of a text is this decrease
sig-nificant (b = -2.474, SE = .763, p = .002). The text and logo explain the variance of
atti-tude towards the politician for almost ten per cent (R2 = .099). The exposure of a photo
has the same effect when measuring the attitude towards a political party, although the
change is less dramatic (b = .631, SE = .673). The R squared for attitude towards political
parties is less predictive, with 0.5%. Summarizing the main effects on attitudes, adding a photograph of the politician is always beneficial. The sole effects of the text and logo decrease the attitudes. However, the means predict that the combination of text and logo creating an interaction effect on the attitude, which is discussed in the next paragraph.
Attitudes towards
political actor Attitudes towards political party
b SE b SE
Constant 24.397 .676 20.608 .861
Text -2.474 .763** -.115 .663
Logo -1.318 .764 -.254 .709
Photo n.a. n.a. .631 .673
R2 .099 .005
N = 213, * p < .05, ** p <.01.Coefficients are unstandardized.
table 4. Linear regression analysis of main effects on attitudes toward the political actor and pol-tical party
In appendix 2 all other measured covariates and interactions are presented. Same as with the effects of memory enhancement, the main cues keep their significance, correspond-ing with the perceived strength of the main effects. When the interactions are taken into account, the addition of a photograph becomes significant when measuring the attitude
towards the political party (b = 2.179, SE = 1.086, p = .046). Hypothesis 4a predicted
this effect and is hereby supported. Showing a photograph on a poster causes an increase of attitudes, both towards the politician as to the party. Hypothesis 4b seems to be not true. The addition of a symbol, in this case the logo, causes a decrease in attitude towards
the politician (b = -1.318, SE = .764, p = .087), but no significant results were found.
Hypothesis 4b is not supported.
The second significant result of attitude found is the one of the raised research question, the age of the participant that would interact with a visual. We expected that a higher age combined with the photo of a politician would increase the cynicism which would lead to a lower attitude. And indeed, this seems to be true; an increase of age leads to a significant
lower attitude (b = -.076, SE = .073, p = .040). This could lead us to believe the theories
about age and cynicism in politics (Agger, Goldstein and Pearl, 1961; De Vreese, 2005; Zhang, Johnson, Seltzer & Bichard, 2009) are indeed true. Research question 1 is hereby supported. The explained variance of attitude towards the political party is 2,5%.
Conclusions and Discussion
This study attempts to answer whether different elements of a poster influence the view-er’s attitude or memory enhancement. By combining known theories and unexplored poster characteristics, this study may deliver some useful insights into the underlying effects of a political visual. Before any results are being discussed, one question ought to be answered: what is the goal of the poster? This study provides four scenario’s, each with different outcomes.
First, the memory enhancement for a politician is analyzed. Priming ones mind with a person’s appearance could be very valuable in a political landscape that has been shifting increasingly towards political leaders. Since the political actor part of this questionnaire was guided with a photograph (i.e. the constant factor), I was unable to measure any differences in memory enhancement because of the addition of the photo. It seemed that the addition of a logo or text had almost no impact on the recall. As expected, respon-dents that were more interested in politics or followed the news closely, were better at recognizing Lodewijk Asscher. And in accordance with the theories, women’s memory enhancement did significantly benefit from visuals in relation to the party. With visual of the actor, men recognized the actors better, the effect was stronger when corrected for the interaction with visuals.
The second dependent variable is the memory enhancement regarding the political party. When a party has trouble being noticed, this may play a important role into getting more recognized. Results point out that having a logo on the poster does significantly improve the chance of being recognized. Having a text or photo seems to distract the viewer and leads to a lower memory enhancement. The combination of text and photo makes this effect even worse, generating the lowest score. Perhaps that the logo does remind the respondent of the party, whereas all other elements generate a stronger priming effect for other topics, like the politician himself.
I expected that those with lower education would have a higher memory enhancement benefit by visuals than those with higher education. Correcting the model for education
and visuals, it seems that this is simply not true. Even more striking, higher educated score significantly better when taking visuals into account as a interaction variable.
A third goal of a poster could be changing the attitude towards a politician. This would be an obvious next step for a political candidate; first you want to get recognized and then establish a healthy positive image. Using a set of character traits as predictor of attitude, this study found that having a text guiding the image would significantly decrease the atti-tude towards the candidate. Knowing text or logo provoke less high attiatti-tude, it is quite re-markable that the combination of both a text and logo improves the attitude towards the actor. We might expect that just adding a logo or a text does not give the viewer enough context, but that the combination of both gives the viewer more knowledge which leads to a higher attitude towards the politician.
The same could be said about the fourth, and final goal a party could have. Adding a pho-to pho-to the poster is the only cue that does improve the attitude pho-towards the party, although not significantly. But when corrected for interactions the photo cue becomes significant. This effect apparently disappears when the interaction between the photo and age is test-ed, older respondent would rate the political party lower when confronted with a photo. We see the same effect occurring as with the attitude toward the politician: the combi-nation of text and a logo does also improve the attitude. It seems that when it comes to creating a higher attitude, the second best option would be printing a text guided by the logo of the party.
Summarizing, this means that if a party wants to improve their attitude towards the party or politician, they should always add a photo of the political actor as this generates the higher attitude. From the theory we know that this evokes more personal and genuine emotions. A face makes the party more personal and likely functions as good gateway for one-way communication with the receiver. Although both a photograph and a logo can be counted as a visual elements in the poster, they differ in the measured effects. It seems that a logo does remind the viewer more of the party, where the photo generates a positive attitude towards the party and the politician.
A cacophony of elements is never a good option; it seems that viewer rather sees a fo-cused poster with one of two elements. This is also true for the memory enhancement of the party. The one goal that does not seem to fit in this guideline, is the memory enhancement of the political actor. In here, the three items combined scored the higher recall, although not significant. If the memory enhancement of the politician needs to be better, there is not much a party can do; from the results of the covariates only the respondents with high political knowledge were significantly better at recognizing the politician.
A limitation of this study may be the fact that the sample only existed out of internet users. Received feedback from one of the respondents pointed out that having a differ-ent media outlet as prime media source, like the radio, could lead to a lower memory enhancement since the respondent would only recognize the voice of a politician, not the face. The same differences could be developed when having a newspaper as prime news source. Future research could have a more cross-media character. The second limitation I would find in this research is the lack of a control poster; a poster with none of the elements. This would only be useful when measuring the initial attitudes, before any of the elements interacted with the respondents. Analyzing memory enhancement would be not possible, since this was based upon what the respondent saw on the poster.
Other recommendations based on the design and results of this study are mainly about widening the research. For example, more elements of the poster could have been taken into account. Some interesting research about the effects of colors have been done in the past and combining this with politics could lead to very interesting results (Mitchell & Olsen, 1981). For example, red has always been linked to the more socialist parties, whereas blue has a strong connection to liberal ones. Color, among other cues, could give great insight into the salience of the poster. Instead of elements, a comparison be-tween parties or actors could also been made. Is there a difference bebe-tween a male and female political candidate and what kind of elements in the visual have interaction ef-fects? Because I only used a left wing party, although close to the middle of the political spectrum, we might expect some effects could have been different.
Lastly, all these found effects on memory and attitude change were short term. The posters were shown to the respondents and in matters of minutes they would answer questions that measured the memory enhancement and attitudes that could have been evoked by the three cues. Although short term effects are quite important for politics (Cappella and Jamieson, 1997; De Vreese, 2005), it is unsure what the duration and stability of these effects are. This study will benefit from a longitudinal approach, mea-suring the impact of political posters on memory enhancement and attitudes over a longer period of time.
Memory enhancement
political actor Memory enhancement political party
B SE B SE
Constant .536 .070 .587 .073
Text .006 .079 -.041 .057
Logo .011 .079 .334 .060**
Photo n.a. n.a. -.066 .058
R2 .001 .166
Covariates
Constant -.451 .411 .600 .304
Text .053 .066 -.001 .052
Logo .042 .066 .326 .055**
Photo n.a. n.a. -.057 .053
Age .003 .003 .002 .003
Sex (0=man, 1=woman) -.101 .082 -.057 .062
Education .001 .033 .030 .024
Media Usage .020 .007** .005 .006
Perceived Political Interest .052 .026* .026 .020
Perceived Tolerance .020 .041 .028 .028
Perceived Social .022 .039 -.063 .029*
Political Knowledge .128 .030** .109 .023**
R2 .381 .332
Main Effect Interactions
Constant .551 .082 .639 .123
Text -.023 .111 -.139 .149
Logo -.019 .114 .253 .107**
Photo n.a. n.a. -.053 .107
Logo x Image n.a. n.a. -.085 .126
Logo x Text .060 .159 .181 .155
Text x Image n.a. n.a. .010 .146
Text x Image x Logo n.a. n.a. -.097 .153
R2 .001 .167
Expected Interactions
Constant .525 .070 .580 .074
Text .022 .079 -.038 .057
Logo .302 .390 .555 .155**
Photo n.a n.a. .167 .154
Sex x Logo x Photo (0=man, 1=woman) n.a. n.a. .034 .025*
Sex x Logo (0=man, 1=woman) -.238 .133 n.a. n.a.
Education x logo x photo n.a. n.a. .061 .025*
Education x logo .069 .148 n.a. n.a.
R2 .031 .175
Appendix 1. All measured variables and interactions of memory enhancement on recalling a political actor or political party. N = 213, * p < .05, ** p <.01.Coefficients are unstandardized.
Attitude towards political
actor Attitude towards political party
B SE B SE
Main effects
Constant 24.397 .676 20.608 .861
Text -2.474 .763** -.115 .663
Logo -1.318 .764 -.254 .709
Photo n.a. n.a. .631 .673
R2 .099 .005
Covariates
Constant 12.105 4.671 14.085 3.718
Text -2.217 .754** -.052 .647
Logo -1.127 .755 -.228 .687
Photo n.a. n.a. .631 .653
Age .040 .036 -.065 .030*
Sex (0=man, 1=woman) .626 .926 -.226 .757
Education .446 .382 .534 .299
Media Usage .096 .088 -.034 .069
Perceived Political Interest .242 .302 -.334 .243
Perceived Tolerance 1.171 .470* .555 .345
Perceived Social -.423 .446 .776 .359*
Political Knowledge .325 .347 .078 .282
R2 .218 .119
Main Effect Interactions
Constant 25.037 .785 19.710 1.397
Text -3.608 1.045** .333 1.690
Logo -2.554 1.091* -.199 1.220
Photo n.a. n.a. 2.179 1.086*
Logo x Image n.a. n.a. -1.126 1.450
Logo x Text 2.395 1.518 2.259 1.764
Text x Image n.a. n.a. -1.310 1.642
Text x Image x Logo n.a. n.a. -3.332 1.751
R2 .119 .050
Expected Interactions
Constant n.a. n.a. 20.526 .856
Text n.a. n.a. .014 .658
Logo n.a. n.a. -.159 .703
Photo n.a. n.a. 2.793 1.261*
Age x Photo n.a. n.a. -.076 .073*
R2 n.a. n.a. .025
Appendix 2. All measured variables and interactions of attitudes on recalling a political actor or political party. N = 213, * p < .05, ** p <.01.Coefficients are unstandardized.
Appendix 3: Overview of variable scores of the questionnaire
Age: In years. (M = 27,54, SD = 10,52)
Gender: Female = 1; male = 0. (M = .76, SD = .22)
Education: ranging from 1 (no education), 2 (primary school), 3 (secondary school or equivalent
[VMBO, MAVO, LBO, VBO, LTS or LHNO]), 4 (secondary vocational education or equivalent [MBO, MTS or MEAO]), 5 (higher secondary school or equivalent [HAVO, VWO or Gymnasium]), 6 high vocational education or equivalent [HBO, HEAO, PABO or HTS]) or 7 ((under)graduate [University]). (M = 6.09, SD = 1.11)
Media usage internet: “How many days a week do you watch news on the internet?”. Items
ranging from 1 (never) to 8 (each day). (M = 6.41, SD = 2.28)
Media usage television: “How many days a week do you watch news on the television?”. Items
ranging from 1 (never) to 8 (each day). (M = 4.61, SD = 2.36)
Media usage newspaper: “How many days a week do you read news in the newspaper?”. Items
ranging from 1 (never) to 8 (each day). (M = 3.66, SD = 2.34)
Constructed media usage: sum of internet, television and newspaper usage. (M = 14.68, SD =
4.58)
Political preference: “What if there would be presidential elections today, what would you
vote?” 1 (VVD), 2 (PvdA), 3 (PVV), 4 (SP), 5 (CDA), 6 (D66), 7 (ChristenUnie), 8 (GroenLinks), 9 (SGP), 10 (Partij van de Dieren), 11 (50PLUS), 12 (Different party), 13 (I would not vote).
Perceived political interest: “To what extent do you consider yourself as politically interested?”
ranging from 1 (not at all) to 7 (completely). (M = 4.43, SD = 1.54)
Perceived tolerance: “To what extent do you see yourself as tolerant?” ranging from 1 (not at
all) to 7 (completely). (M = 5.36, SD = .981)
Perceived sociality: “To what extent do you see yourself as social?” ranging from 1 (not at all)
to 7 (completely). (M = 5.54, SD = .92)
Political knowledge. A collection of five questions measures the political knowledge.
Questions could be answered right (1) or wrong (0).
1. How many seats are there in the Eerste Kamer? [random order answer options, except ‘Ik weet niet’]: 50 (0), 75 (1), 85 (0), 100 (0), 150 (0), Ik weet niet (0).
2. Who is currently the minister voor Wonen en Rijksdienst in cabinet Rutte-II? [random order answer options, except ‘Ik weet niet’] Frans Timmermans (0), Stef Blok (1), Henk Kamp (0), Wouter Bos (0), Ronald Plasterk (0), Ik weet niet (0).
3. Of which party is Bram van Ojik leader? [random order answer options, except ‘Ik weet niet’] SGP (0), SP (0), ChristenUnie (0), GroenLinks (1), 50Plus (0), Ik weet niet (0).
4. Which of the following party is most left wing in the political spectrum? [random order answer options, except ‘Ik weet niet’] SP (1), CDA (0), D66 (0), VVD (0), PVV (0), Ik weet niet (0) 5. How many countries are there currently in the European Union? 6 (0), 15 (0), 25 (0), 28 (1), 30 (0), Ik weet niet (0).
Constructed variable political knowledge. Total correct answers divided by 5. (M = 2.05, SD =
1.23).
Memory Enhancement of politician. Question could be right (1) or wrong (0).
1. “Who is this person?” Diederik Samsom (0), Henk Kamp (0), Stef Blok (0), Lodewijk Asscher (1), Ronald Plasterk (0), Weet ik niet (0). (M = .27, SD =.45)
2. “What is the job title of this person?” Burgemeester van Den Haag (0), Minister van Sociale Zaken en Werkgelegenheid (1), Staatssecretaris van Onderwijs (0), Minister President (0), Voorzitter van de FNV (0), Weet ik niet (0). (M = .24, SD = .43)
Constructed variable memory enhancement of politician. Two questions measuring memory
enhancement of the politician divided by 2. (M = .26, SD = .41).
Memory Enhancement of the political party. Question could be right (1) or wrong (0).
1. “What is the political party depicted in this poster?” PvdA (1), VVD (0), SP (0), GroenLinks (0), Partij van de Dieren (0), Weet ik niet (0). (M = .62, SD = .47).
Attitude towards politician. Measured on a 5-point scale, ranging from 1 (Totally disagree)
to 5 (Totally agree). Question: “I think this person is:” “Reliable” (M = 3.26, SD = .77), “Knowledgeable” (M = 3.44, SD = .79), “Caring for people like me” (M = 2.98, SD = .85), “Integer” (M = 3.21, SD = .70), “Involved” (M = 3.32, SD = .75), “Authentic” (M = 3.09, SD = .72), “Convincing” (M = 3.19, SD = .81). All questions measure 1 factor. Cronbach’s Alpha = .902.
Constructed variable attitude towards politician. Seven questions measuring the attitude
added up. Minimum of 7, maximum of 35. (M = 22.49, SD = 4.29).
Attitude towards political party. Measured on a 5-point scale, ranging from 1 (Totally
disagree) to 5 (Totally agree). Question: “I think this party is:” “Reliable” (M = 2.87, SD = .88), “Knowledgeable” (M = 3.13, SD = .83), “Caring for people like me” (M = 3.01, SD = .88), “Integer” (M = 2.94, SD = .82), “Involved” (M = 3.18, SD = .88), “Authentic” (M = 2.91, SD = .87), “Convincing” (M = 2.73, SD = .90). All questions measure 1 factor. Cronbach’s Alpha = .879.
Constructed variable attitude towards political party. Seven questions measuring the attitude
References
Agger, R. E., Goldstein, M. N., & Pearl, S. A. (1961). Political cynicism: Measurement and meaning. The Journal of Politics, 23(03), 477-506.
Arpan, L.M., Baker, K., Youngwon, L., Taejin, J., Lorusso, L., & Smith, J. (2006).
News coverage of social protests and the effects of photographs on prior attitudes.
Mass Communication & Society, 9(1), 1-20.
Banducci, S.A., Karp, J.A., Thrasher, M., & Rallings, C. (2008). Ballot photographs as cues in low- information elections. Political Psychology, 29, 903917.
Barrett, A.W., & Barrington, L.W. (2005). Is a picture worth a thousand words?: Newspaper photographs and voter evaluations of political candidates. The Harvard International Journal of Press/Politics, 10(4), 98-113.
Bennett, W. L. (2004). Branded political communication: Lifestyle politics, logo campaigns, and the rise of global citizenship. Politics, products and markets: Exploring political
consumerism, past and present, 101-125.
Blair, J.A. (2004). The rhetoric of visual arguments. In C.A. Hill & M. Helmers (Eds.),
Defining visual rhetorics. Mahwah, NJ: Erlbaum.
Blumler, J. & Kavanagh, D. (1999). The Third Age of Political Communication: Influences and Features. Political Communication. 16, 209-230.
Boen, F., & Vanbeselaere, N. (2002). The Impact of Election Outcome on the Display of Political Posters: A Field Study During Communal Elections in Flanders. Political Psychology, 23(2), 385-391.
Brants, K., & van Praag, P. (2006). Signs of media logic. Half a century of political communication in the Netherlands. Javnost/The Public, 13(1), 25-40.
Cappella, J. & Hall, K. (1997). Spiral of Cynicism: The Press and the Public Good. New York:
Oxford University Press.
Cohen, G., & Burke, D. M. (1993). Memory for proper names: A review. Memory, 1(4), 249-263.
Carpini, D., Michael X., & Keeter, S. (1996). What Americans Know About Politics and Why It Matters. New Haven, CT: Yale University Press.
Dumitrescu, D. (2010). Know me, love me, fear me: The anatomy of candidate poster designs in the 2007 French legislative elections. Political Communication, 27(1), 20-43.
Dumitrescu, D. (2011). The importance of being present: Election posters as signals of electoral strength, evidence from France and Belgium. Party Politics, 1-20.
Farrell, D. M., & Wortmann, M. (1987). Party strategies in the electoral market: Political marketing in West Germany, Britain and Ireland. European Journal of Political Research, 15(3), 297-318.
Graber, D. A. (1990). Seeing is remembering: How visuals contribute to learning from television news. Journal of communication, 40(3), 134-156.
Graber, D. A. (1996). Say it with pictures. The annals of the American academy of political and social science, 85-96.
Greene, E., & Fraser, S. (2002). Observation Distance and Recognition of Photo- graphs of Celebrities’ Faces. Perceptual and Motor Skills, 95, 637–51.
Van Ham, P. (2001). The rise of the brand state: the postmodern politics of image and reputation. Foreign affairs, 2-6.
Hill, C. A. (2004). The psychology of rhetorical images. Defining visual rhetorics, 25-40.
Huckfeldt, R. and Sprague, J. (1992). Political Parties and Electoral Mobilization: Political Structure, Social Structure, and the Party Canvass. American Political Science Review, 86, 70–86.
Iyengar, S., Norpoth, H., & Hahn, K. S. (2004). Consumer demand for election news: The horserace sells. Journal of Politics, 66(1), 157-175.
Joffe, H. (2008). The power of visual material: Persuasion, emotion and identification.
Diogenes, 217, 84-93.
Kirouac, G., & Dore, F.Y. (1985). Accuracy of the judgment of facial expression of emotions as a function of sex and level of education. Journal of Nonverbal Behavior, 9, 3-7.
Kisielius, J., & Sternthal, B. (1984). Detecting and explaining vividness effects in attitudinal judgements. Journal of Marketing Research, 21, 54-64.
Lang, A., Potter, R, & Bolls, P. (1999). Something for Nothing: Is Visual Encoding Automatic?
Media Psychology, 1, 145–63.
Mitchell, A. & Olson, J. (1981). Are product attribute beliefs the only mediator of advertising effects on brand attitude? Journal of Marketing Research. 18, 318-332.
Mitchell, A. (1986). The effect of verbal and visual components of advertisements on brand attitudes and attitude toward the advertisement. Journal of Consumer Research, 12-24.
Newhagen, J.E. (1998). TV news images that induce anger, fear, and disgust: Effects on approach-avoidance and memory. Journal of Broadcasting & Electronic Media, 42, 265-276.
Petchesky, R. P. (1987). Fetal images: The power of visual culture in the politics of reproduction. Feminist Studies, 263-292.
Petty, R. E., & Cacioppo, J. T. (1986). Communication and persuasion: Central and peripheral routes to attitude change.
Prior, M. (2014). Visual Political Knowledge: A Different Road to Competence?
The Journal of Politics, 76(01), 41-57.
Riegelsberger, J., Sasse, M. A., & McCarthy, J. D. (2003, April). Shiny happy people building trust?: Photos on e-commerce websites and consumer trust. In Proceedings of the SIGCHI conference on Human factors in computing systems, 121-128. ACM.
Rosenberg, S. W., Bohan, L., McCafferty, P., & Harris, K. (1986). The image and the vote: The effect of candidate presentation on voter preference. American Journal of Political Science, 30, 108 - 127.
Rossiter, J. R., & Percy, L. (1979). Visual communication in advertising. Columbia University,
Graduate School of Business.
Sellers, P. J., & Schaffner, B. F. (2007). Winning coverage in the US Senate. Political Communication, 24(4), 377-391.
Schill, D. (2012). The Visual Image and the Political Image: A Review of Visual Communication Research in the Field of Political Communication. Review of Communication, 12(2),
118-142.
Smith, G. (2001). The 2001 general election: Factors influencing the brand image of political parties and their leaders. Journal of Marketing Management, 17(9-10), 989-1006.
Strachan, J. C., & Kendall, K. E. (2004). Political candidates’ convention films: Finding the perfect image—An overview of political image making. Defining visual rhetorics, 135-154.
Strömbäck, J. (2007). Political marketing and professionalized campaigning: A conceptual analysis. Journal of Political Marketing, 6(2-3), 49-67.
Strömbäck, J. (2008). Four phases of mediatization: An analysis of the mediatization of politics.
The International Journal of Press/Politics, 13(3), 228-246.
Tschabrun, S. (2003). Off the wall and into a drawer: Managing a research collection of political posters. American Archivist, 66(2), 303-324.
Vliegenthart, R. (2012). The professionalization of political communication? A longitudinal analysis of Dutch election campaign posters. American Behavioral Scientist, 56(2), 135-150.
De Vreese, C. H. (2005). The spiral of cynicism reconsidered. European Journal of Communication, 20(3), 283-301.
Zhang, W., Johnson, T. J., Seltzer, T., & Bichard, S. L. (2009). The revolution will be networked: The influence of social networking sites on political attitudes and behavior. Social Science Computer Review.
