U n i v e r s i t y o f A m s t e r d a m
Master’s Thesis
Mirthe Enklaar - 10211454
Graduate School of Communication
Master’s programme Communication Science – Persuasive Communication University of Amsterdam
Supervisor: Dr. Lisa Vandeberg Date: June 23, 2015
June
2015
08
Fall
Is our evaluation of advertised
brands unconsciously influenced by
the valence of news messages?
The effects of positive and negative news messages on people’s implicit
and explicit attitude toward brands and advertising
Abstract
On account of the increasing importance of online advertising, the aim of this study was to investigate whether the valenced context of online ads influences people’s implicit and explicit brand attitudes, and to explore if differences between these attitudes could be explained by persuasion knowledge. The online experiment showed that the valence (being the evaluative direction) of the advertisements’ context does not influence implicit and explicit attitudes toward the advertised brand. In addition, the main effect of context does not differ for people who are high or low in persuasion knowledge. This implies that it does not matter for implicit and explicit brand attitudes whether the context of online advertising is positive, negative or neutral.
Introduction
You might be familiar with annoying advertisements such as those online pop-up advertisements that prevent you from reading a website. Being annoyed may negatively affect people’s attitudes toward the advertised brand, which in this case is a kind of attitude people are aware of. However, people’s brand attitudes in response to advertisements can also be changed in ways of which people are not aware. Yet, it is not clear what influences this last kind of “subconscious” attitude.
The effectiveness of advertising can be influenced by its context (Bronner, Bronner & Faase, 2007), which is about the editorial surroundings of the advertisement, such as news messages and other ads. Research has shown that affect evoked by the context may
automatically spill over to the advertisement (Bronner & Neijens, 2006; Moorman, Neijens & Smit, 2002). This means that when someone is exposed to a negative context, effects on attitude toward the advertised brand become negative as well.
In this study, the focus is on the effects of a valenced (positive, negative or neutral) context that surrounds an online advertisement and its influence on people’s subconscious brand attitudes. In addition, persuasion knowledge of the participants is taken into account. This concept can be described as the general knowledge people have about persuasion and how they use this knowledge to interpret, evaluate, and respond to persuasion attempts (Friestad & Wright, 1994). It might be possible that this knowledge differentially influences people’s implicit and explicit brand attitudes (Gibson, Redker & Zimmerman, 2014).
The status quo of the literature leads to the research question of this study: “Does the context of online advertisements have an influence on people’s implicit and explicit attitude
toward the advertised brand, and how is this relationship moderated by persuasion
Until now, the various context studies do not provide a clear understanding of the
nature of the automatic responses towards advertising, especially not in an online setting (Gibson et al., 2014; Moorman et al., 2002). Most research in the field of communication is focused on the advertising effects of television and print by making use of explicit
measurements, being people’s self-reported responses towards brands (Conner & Sparks, 2004). However, such explicit measurements are not able to give a complete overview of a person’s mental content (Nisbett & Wilson, 1977). In general, people do not pay attention to online ads (Drèze & Hussherr, 2003; Matthes, Schemer & Wirth, 2007; Redker, Gibson & Zimmerman, 2013) and because of that, it is hardly impossible to use self-reports solely to evaluate these online ads. Though, this does not mean that people will not be influenced by context. To get a complete overview of the processing of the online context of advertising, this study made use of explicit measurements in combination with implicit measurements. Implicit measurements provide the ability to get more insight in the unconscious processes of advertising. They are aimed at examining people’s automatic affective reactions toward stimuli without asking them to introspect (Nisbett & Wilson, 1977). Implicit measurements might be very useful especially within the online context of advertising because (a) of the assumed automatic nature of the responses towards unattended advertising, (b) the avoidance people have to online ads (Resnick & Albert, 2014) and (c) likeability of the context in which a brand is placed influences viewers’ implicit but not explicit attitudes toward the brand (Redker et al., 2013).
As described above, there is a lack of knowledge about the automatic processes that underlie advertising processing (Greenwald & Banaji, 1995; Grimes & Kitchen, 2007). This study expands the existing scientific knowledge on implicit attitudes and persuasion
Furthermore, it can be assumed that advertisers want their ads to be highly effective. It might be possible that ads surrounded by a negative context have destroying effects for brands in terms of consumer attitudes. Because of this, it is of direct importance to practitioners to know to what extent context may have such problematic effects. This research can help marketers and advertisers to get more insight in the influence of the
advertisements’ context in order to set up the most effective ads in terms of positive attitudes. It is relevant for them to know how this can be achieved.
Theoretical framework
Nowadays, we spent an enormous amount of time online. In our so-called Internet age, online advertising continues to increase at a high rate. Internet ad spending grew faster than that of any other media (Ha, 2008; Lappe, 2013). It is expected that online ads will account for 33 per cent of total advertising revenue in the next four years (PWC, 2015). These findings show an important shift in the focus of advertising from the traditional media toward more non-traditional online channels (Gartner, 2013). The question is whether the effectiveness of such online ads is influenced by its surroundings.
Context of advertisements: positive or negative contexts
Research has shown that the effectiveness of advertising can be influenced by its context (Bronner et al., 2007). Context refers to the editorial surroundings of the advertisement, such as news messages. In the scientific literature, it is generally accepted that differences in the media context of an advertisement may influence advertising effects, even when other relevant factors like audience characteristics are held constant. This occurs through a principle called “priming”, which suggests that the processing of the stimulus (“the prime”) influences the interpretation and reaction of the following stimulus (“the target”) (Bronner &
as within a positive context, they may experience positive feelings, which subsequently influence their brand attitude.
Moreover, Coulter (2013) showed that a positive or negative context affects the way people form attitudes. This can be explained with evaluative conditioning (De Houwer, 2007). By pairing an ad (the conditioned stimulus - CS) with a positive or negative context (the unconditioned stimulus - US), the advertiser hopes that the positive evaluation of the US will transfer to the CS (De Houwer, 2007; Gibson et al., 2014; Redker et al., 2013). This results in a response, an attitude, as suggested by the affective response theory (Bronner et al., 2007; Moorman et al., 2002). This theory assumes that a positive or negative context might evoke feelings that transfer to the evaluation of the advertisement (Bronner et al., 2007; Coulter, 2013; Moorman et al., 2002). The affective response theory is somewhat similar to the affective priming hypothesis of Murphy and Zajonc (1993), which asserts that positive and negative affective reactions can be evoked with minimal stimulus input and virtually no cognitive processing. This implies that a prime, the context, can evoke positive and negative feelings that influence processing of the ad. In other words, ads that are placed within a positive context might be evaluated more positively than ads that are surrounded by a negative context.
Various theories and mechanisms have been proposed for the effect of context on attitudes. However, these studies were aimed at context effects of television and print advertising on explicit attitude toward the ad, such as the study of Moorman et al. (2002). Their study showed that when people were exposed to a magazine they liked, the ads in the magazine were evaluated more positively. Extending these results towards the online context of this study, positive news messages will be evaluated more positively and liked better than negative news messages, which subsequently has a positive influence on people’s brand attitudes.
Implicit and explicit attitudes toward the advertised brand
Attitude toward the brand is defined as “a predisposition to respond in a favorable or unfavorable manner to a particular brand stimulus during a particular exposure occasion” (MacKenzie & Lutz, 1989 p. 49). In other words, it is an evaluative or affective response to the brand and does not refer to cognitive or behavioral responses.
In this study, both explicit and implicit brand attitudes were taken into account, because individually they are not adequate enough to understand the context phenomenon. Research has shown that more than half of the online visitors do not pay attention to online ads (Drèze & Hussherr, 2003). Because people often neglect advertising, traditional cognitive models of ad processing are not appropriate to examine the effects of online ads (Yoo, 2008): if consumers do not pay attention to the ad, they cannot consciously process and report about it. In this case, effects are more automatic or implicit.
Information can be processed in two different ways: by means of an explicit
conscious (controlled) process and a more implicit unconscious (automatic) process (Chaiken & Trope, 1999; Nairn & Fine, 2008). According to dual process theory, these different processes can result in different outcomes with respect to people’s attitudes (Gawronski & Bodenhausen, 2006; Greenwald & Banaji, 1995).
To start, explicit attitudes are attitudes of which people are aware. These attitudes are also defined as “deliberative, self-reported evaluations” (Gawronski & Bodenhausen, 2006, p. 692). Most research in the field of communication and advertising is focused on explicit attitudes, and therefore a lot of models of persuasion have been proven to be successful in explaining the influence of message cues, such as context, on self-reported brand evaluations. Traditionally, marketers were interested in explicit attitude change, but research has shown that advertising that links products with positive stimuli can affect implicit attitude as well (Nairn & Fine, 2008). Moreover, explicit measurements “are unable to reflect the mental
processes on an “automatic” level, simply because people cannot explain things they are not aware of” (Nisbett & Wilson, 1977, p. 248). These kinds of measurements are not capable of tapping into people’s automatic-subconscious responses about which they have no insights. Here, implicit measurements might be a solution.
Implicit attitudes on the other hand, are “automatic affective associations that are triggered when someone encounters the relevant attitude object” (Gawronski &
Bodenhausen, 2006, p. 693). This means that people do not need to fully process information to be influenced by it. In other words, a large part of a consumer’s processing of ads will be done at a pre-attentive level rather than at a full attentive level (Drèze & Hussherr, 2003).
As explained by dual process theory, such automatic processing likely affects implicit but not explicit attitudes. In addition, implicit attitudes are of importance because these attitudes “have validity, enhance the predictability of behavior over and above that provided by explicit attitudes” (Nairn & Fine, 2008, p. 457). It is important to notice that implicit attitudes can be influenced without influencing explicit attitudes.
Research has shown that context affect may spill over to implicit brand attitudes. Redker et al. (2013) for example showed that brand placements in a liked movie have a positive influence on implicit brand attitudes, whereas seeing the brand in a movie people dislike leads to negative implicit brand attitudes. In other words, movie generated affect transferred to implicit attitudes. However, no such effects were found for explicit brand attitudes. Extending these findings to an online context, a negative context of the ad might lead to negative brand attitudes whereas a positive context of the ad will improve people’s attitudes toward the advertised brand. So, the effects of the valenced context (being the evaluative direction of context) might spill over to the evaluation of the ad. In other words, context may contribute to someone’s implicit attitude; which subsequently strongly
influences judgment and behavior (Nairn & Fine, 2008). Therefore, the first hypothesis and research question are:
H1: The valence of context has an influence on peoples’ implicit attitude toward the advertised brand, meaning that a positive (negative) context leads to a more positive (negative) implicit attitude toward the advertised brand than a neutral context.
RQ1: Does the valence of context have an influence on explicit attitude toward the advertised brand?
The moderating role of persuasion knowledge
Persuasion knowledge is the general knowledge people have about persuasion, which affects the way in which they interpret, evaluate, and respond to persuasion attempts (Friestad & Wright, 1994). Persuasion knowledge is of importance because ads become less effective as viewers become more skeptical of them, which is now the case with traditional ads (Friestad & Wright, 1994). Research has shown that people automatically process product placement (Cowley & Barron, 2008). However, when the product placement is recognized as
advertising, explicit brand attitudes suffer (Gibson et al., 2014). This is because the product placement is recognized as a persuasive attack. Originally, it might have generated positive feelings but it is now viewed as a manipulative attempt to alter people’s brand attitudes (Gibson et al., 2014). Moreover, research has shown that the higher someone’s persuasion knowledge, the greater their skepticism toward advertising, which results in less positive attitudes toward the ad and brand (Boerman, van Reijmersdal & Neijens, 2012; Friestad & Wright, 1994; Hibbert, Smith, Davies & Ireland, 2007). This is probably due to the
controllable, intended and cognitive resourceful nature of people’s thoughts (Nosek, 2007). The recognition of a manipulative attempt is a feeling of which you are aware. In other words, it is an explicit feeling.
However, the decrease in attitudes might not be the case with implicit attitudes. Redker et al. (2013) showed that when people are exposed to a liked character in a movie, implicit brand attitudes increase whether persuasion knowledge was activated or not. An explanation might be that this type of attitude is known for a smaller amount of cognitive load, reduced controllability, lack of intention, reduced awareness and high efficiency of processing. In other words, implicit attitudes do not suffer from deliberative thoughts whereas explicit attitudes do (Gibson et al., 2014; Redker et al., 2013). Thus, it is assumed that persuasion knowledge moderates the effect of context on explicit brand attitudes but not implicit attitudes, as reflected by the second hypothesis:
H2: Persuasion knowledge has a moderating role in the relationship between context and explicit, but not implicit, attitude toward the advertised brand, meaning that a valence effect of context will be found for people with low persuasion
knowledge compared to those with high persuasion knowledge.
Figure 1
Conceptual model
Persuasion knowledge
Explicit attitude toward the advertised brand Context of an online
ad (positive, negative,
Method Design
An effect of valence of the context in which a brand is advertised on explicit and implicit attitude toward the advertised brand was expected. To show the potential of such a causal relationship, an experiment with an online questionnaire was conducted.
This experiment has a 2 (exposure: yes vs. no) x 3 (context: positive vs. neutral vs. negative) mixed-level design, meaning that the context manipulation only occurred for the participants that were exposed to the advertisement of the target brand, see Table 1. In total, there were four conditions and the participants in the control condition (no exposure to the target ad) were exposed to a neutral news message with a filler ad. This was done to measure if exposure to the target ad does have an effect in general (a so-called exposure effect). The people that were exposed to the target ad faced either a positive, negative or neutral news message. This resulted in three experimental conditions that were of main interest.
Table 1
Experimental design
Context Positive Neutral Negative Target ad Filler ad n = 50 X n = 55 n = 55 n = 56 X N = 216 Sample
The sample consisted of a total of 216 participants aged 18 to 30 years (M = 22.78; SD = 2.12) and 67 per cent (n = 146) of the participants were females. The participants had to be 18+ because from that age on people can “recognize advertising, understand its selling and
persuasive intent, and use this understanding in evaluating the advertised product or service” (Nairn & Fine, 2008, p. 447; Mangleburg & Bristol, 1998; Friestad & Wright, 1994). Before data gathering, the ethical board committee of ASCoR approved the online experiment. Subsequently, the link of the online questionnaire was disseminated throughout the personal network of the researcher: via Facebook and e-mail, in other words by means of a convenience sample. After agreement on the informed consent, the experiment started. Participation was voluntary and the participants could participate in a raffle to win a €25, - gift voucher from H&M.
Stimulus materials
The stimulus materials consisted of a manipulated screenshot of a web page of the Dutch news website NU.nl. All of the participants were exposed to this web page, which contained a news message and an advertisement. The news message was in Dutch and emotionally charged in either a positive, negative or neutral way. To make the valence of these news messages more salient, existing news messages were slightly edited. As in the original format of the website, an advertisement was placed next to the news message. This ad was created by using existing photographical materials of the target brand and was the same in each condition, except for the control condition. The participants in this condition were exposed to an ad of beauty brand Dove, see “Pretest”.
In total, there were four different web pages: (1) a positive news message about the Netherlands winning 24 Olympic medals in 2014 with the target ad, (2) a negative news message about the GermanWings airbus crash in March 2015 with the target ad, (3) a neutral news message about the most common baby names in the Netherlands in 2014 with the target ad and (4) the same neutral news message about baby names with a filler ad (control
Pretest
A pretest was conducted for two different goals. Firstly, to choose a brand that was perceived as neutral by this target group. A neutral brand attitude was needed because brands that evoke very strong (positive or negative) attitudes could influence the results, meaning that a positive attitude toward the advertised brand might be caused by the pre-existing positive attitude toward that brand (Gibson, 2008). Second, to make sure that the positive news message was perceived as being positive, the negative one as negative and the neutral one as being neutral. Twenty undergraduates aged 18 to 26 years (M = 22.6; SD = 1.81; 60.0% female) that suited the criteria of the main study participated in the pretest. To make sure that a neutral brand was used for the ad in the main study, the participants were asked to answer questions concerning their explicit brand attitude and familiarity with twelve brands from four different product categories, see Table 2. Familiarity was measured with two questions, namely “Do you use this product?” and “Do you buy this product?” on a three-point scale with 1 = “never”, 2 = “sometimes” and 3 = “often”. The brand attitude scale of Chung and Zhao (2003) was used to measure (explicit) attitude toward the brand, by asking the participants to indicate their attitude toward the brand on a seven-point semantic differential with bad-good,
unpleasant-pleasant, unfavorable-favorable and uninteresting-interesting.
The pretest showed that Evian was perceived as the most neutral brand (M = 3.23; SD = .94) and was therefore used as the advertised brand. In addition, Evian was somewhat comparable in terms of brand familiarity with another brand from the same product, namely Bar-le-Duc (see Appendix B, Table B1). This comparability was necessary because of the second round of the ST-IAT (see “Measures”).
Table 2
Twelve brands that were pretested, classified by product category
Crèmes Detergents Toothpaste Water
Dove Nivea Vaseline Ariel Omo Witte Reus Aquafresh Colgate Oral B Bar-le-Duc Evian Spa
Subsequently, the participants were asked to read nine news messages of which three were positive, three negative and three were neutral. The news messages were no longer than 200 words and the ones that were perceived as being most positive, negative and neutral were selected for the experiment (see Appendix B, Table B2).
Procedure
After the participants faced the informed consent and agreed to take part in the study, they were randomly exposed to one of the four web pages. After being exposed for as long as they needed, the participants performed two single target IATs (ST-IATs) to measure their
implicit attitude toward the advertised and a filler brand. Here, they had to categorize words and images as being positive or negative and belonging to the brand in a response-time environment (see “Measures”). This test was followed by questions about the purpose of the study, the extent to which the participants attended to the news message and the loading time of the ST-IATs to check if delays in response times could be caused by the functioning of the ST-IATs. Subsequently, the participants were asked to answer some comprehension
questions about the content of the news message, to indicate if they have seen an
advertisement, if they were able to recall and recognize the advertised brand and if they could indicate whether the news message was positive, neutral or negative. The latter question about their perception of the valence of the news message was used as a manipulation check.
Then, questions concerning their explicit attitude toward the target and filler brand, explicit attitude toward the advertisement, persuasion knowledge and purchase intention followed. Finally, the participants answered several demographic questions after which they were thanked and debriefed.
Measures
Independent variables
There were two independent variables of which valence of the context (the news messages) was one. Context was valenced in either a positive, negative or neutral way, as indicated by the pretest.
The second independent variable was persuasion knowledge and this was measured following the same procedure as Boerman et al. (2012) did by asking the participants whether they agreed with the statement: “I think advertising in general is …” (honest, trustworthy,
convincing, biased, not credible) on a 7-point Likert scale with 1 = “completely agree” and 7
= “completely disagree”. High scores correspond to more critical feelings toward advertising, whereas low scores exemplify less distrust. In other words, the higher the persuasion
knowledge score, the more skeptical that person is toward advertising in general.
To create a scale of these three items, a Principal Component Analysis (PCA) was conducted. However, the PCA showed that the five items did not represent a one-dimensional scale: there are two components with an Eigenvalue above 1, of which one involved the positive directed items such as honest, trustworthy and convincing (EV = 2.28; R2 = 45.50)
and one involved the negative directed items of persuasion knowledge such as not credible and biased (EV = 1.59; R2 = 31.81). As expected, the latter items did not load on the same
factor as the first three items. Even after recoding these two negatively directed items, the construction of one reliable scale was not possible. This is probably due to their opposed
(negative) direction that might have confused the participants. For this reason (and after consulting Boerman), only the first three items were used to create a reliable of persuasion knowledge (α = .89; M = 3.40; SD = 1.10). This scale suggests that the higher the score, the less skepticism toward advertising in general. Finally, two groups were created by means of a median split (Mdn = 3.33) of which one group was low and the other group was high in persuasion knowledge.
Dependent variables
The dependent variables that were of interest in this study were two different kinds of consumer responses.
First, implicit attitude toward the advertised brand was measured with two single-target IATs (ST-IAT; Wigboldus, Holland & Van Knippenberg, 2004). IATs (Implicit Association Test; Greenwald; McGhee & Schwartz, 1998) assess automatic associations among different concepts by asking people to categorize stimulus items as quickly as possible on two different categories, by pressing the “e” and “i” keys on their keyboard (Ranganath & Nosek, 2008). The logic of this response-time measurement is that if constructs are closely related in memory, the keyboard response when those concepts share a response key should be faster. So, if Evian is more strongly associated with positive constructs, the keyboard response should be faster when this brand is categorized on the same key as positive words and images rather than with negative ones (Gibson et al., 2014).
The performed ST-IAT consisted of five rounds, starting with a trial round. This trial round consisted of twenty items in which the participants were asked to categorize words and images as being either positive or negative, by pressing “e” (positive) and “i” (negative) as quickly as possible (De Houwer, Custers & De Clercq, 2006). Five Dutch positive words: good (goed), awesome (geweldig), nice (leuk), happy (blij), laughter (lachen) and five
positive images: sun, thumbs up, happy smiley, heart, laughing child were used, as well as five negative words: bad (slecht), terrible (vreselijk), stupid (stom), sad (verdrietig), cry
(huilen) and images: thunder, thumb down, sad smiley, broken heart, crying child.
In the following two rounds, the participants were instructed to include brand names and images in their categorization, in which the brand was devoted to the same key as the positive stimuli in one round (“e”-key) and to the negative stimuli (“i”-key) in the other round, see Table 3. Each round consisted of twenty randomly presented items, which resulted in ten “e”-key and ten “i”-key responses (e.g. by showing ten positive, five negative and five brand stimuli in the negative block). The assumption behind this test is that the more positive people’s implicit attitude toward the advertised brand (Evian), the faster they respond in the round when this brand was linked to positive stimuli, compared to the round that was linked to negative stimuli. After these two rounds, the participants performed the same task for filler brand Bar-le-Duc. The function of these last two rounds is to provide a comparison brand to make sure that any effects in the first two blocks could be ascribed to the advertised brand and not to brands or moods in general.
The second dependent variable was explicit attitude toward the advertised brand and this was measured using self-report. The attitude scale of MacKenzie and Lutz (1989) was used to measure participants’ overall likability of the brand in a reliable way. The subjects had to indicate on a three-item 7-point semantic differential what they thought of the brand in terms of bad-good, unpleasant-pleasant, unfavorable-favorable.
To construct a scale, the same process as described with persuasion knowledge has been carried out. The PCA showed that the three items loaded on one factor with an Eigenvalue above 1 (EV = 2.81, R2 = 70.41). The statement “I think Evian is
unfavorable/favorable” had the highest correlation (component value = .91). The scale is highly reliable (M = 4.53; SD = 1.01; α = .84) and it can be concluded that it reliably
measures the participants’ explicit attitude toward the advertised brand. This scale suggests that the higher the mean score, the more positive the explicit attitude toward Evian.
Table 3
Category assignment and stimulus proportions across ST-IAT blocks for an exemplary
participant
Number of stimuli Block Task description Left key concepts Right key concepts Positive Negative Brand 1 Evaluative
training trials
Positive Negative 10 10 -
2 Initial block Positive + Evian Negative 5 10 5
3 Reversed block Positive Negative + Evian 10 5 5
4 Initial block Positive + Bar-le-Duc
Negative 5 10 5
5 Reversed block Positive Negative + Bar-le-Duc
10 5 5
Note. Adapted from “Reliability and validity of the Single-Target IAT (ST-IAT): Assessing automatic
affect towards multiple attitude objects”, by M. Bluemke and M. Friese, 2008, European Journal of
Social Psychology, 38(6), p. 978.
Results
Preparatory analyses
After nine days of collecting data, 246 people of ages between 18 and 30 years participated in the online experiment. However, thirty participants were excluded from the analysis for multiple reasons, to start with those participants that did not read the news message carefully. This was checked in two different ways. Firstly, by means of a timer that was implemented in the stimulus material. The participants that skipped the “news page” in less than 10 seconds were deleted (N=11), because it was impossible to read the news message and view the ad carefully within this short amount of time. Secondly, participants that answered more than
two of the five comprehension questions incorrect were deleted for further analyses as well (N=8). These participants were omitted because it cannot be assured that they have read the news message carefully. The other participants were deleted to meet the criteria for IAT research (N = 11, see “IAT criteria”).
Manipulation check
To make sure that any differences in results can be assigned to the different kinds of context, a manipulation check was carried out. To see whether the manipulation of the different news messages had succeeded, an Analysis of Variance (ANOVA) was conducted. This test revealed a strong significant difference (F (3, 212) = 75.42; p < .001; η² = .52) between the four conditions and the perceived valence of the news message, meaning that the participants in the positive condition evaluated the news message more positively than those in the negative (p < 0.001), neutral (p < .001) and control condition (p = .001). Even though there was a marginal significant difference between the neutral and control condition (p = .092), it can be concluded that the manipulation was succeeded (see Figure 2).
Figure 2
Mean scores of the manipulation check: the perceived valence of the news message
Note. The perceived valence of the news message could be indicated on a scale ranging from -3 “extremely negative” to +3 “extremely positive”
-2 -1,5 -1 -0,5 0 0,5 1 1,5 2 Positive
condition condition Negative condition Neutral condition Control
P er ce iv ed v a lence Experimental conditions
Control variables
Furthermore, analyses have been carried out with potential control variables such as
demographics, ad recall and recognition, because it was expected that these might play a role in the relationship between the ad’s context and both implicit and explicit attitude toward the advertised brand. One-way ANOVAs and bivariate correlation analyses were conducted to see if there was a relationship between these variables and attitude toward the advertised brand. A Pearson Correlation test was carried out for the potential control variables that were measured at a ratio level such as age, and Chi-square tests for those variables that consisted of two levels like gender and ad recall. However, none of the variables were significant for implicit and explicit attitude toward the advertised brand. In other words, differences in the conditions regarding the dependent variables cannot be caused by differences in background characteristics.
IAT criteria
As mentioned before, several participants were excluded to meet the standard IAT criteria. As in previous IAT research, participants who had committed 20% or more errors in the four different blocks were skipped for the tests that concerned IAT responses (N = 6). In addition, recoded latencies below 300 milliseconds and above 3000 milliseconds were coded as missing because these participants clicked too slowly or too fast to reflect their implicit attitude. The participants that had too many unusable response times (< 300 and >3000 milliseconds) were deleted (N = 5) because these might distort the results (Bluemke & Friese, 2008).
Like most social psychological research using response latencies, in IAT research log transformations are used to normalize distributions (Hummert, Gartska, O’Brien, Greenwald & Mellott, 2002). In the literature, these so-called log transformations are perceived as a
solution to the skewed distribution of reaction times (Greenwald, Nosek, & Banaji, 2003). Hence, each score of each block was log transformed, after which a log transform mean score was created. Subsequently, a log transform differentiation score was computed by subtracting the mean scores of the negative block of the mean scores of the positive blocks. This means that the higher the score, the more positive the implicit brand attitude.
All of the analyses for implicit attitude were conducted as this log differentiation score being the dependent variable implicit attitude toward the advertised brand. However, the non-log transformed data are reported for the mean scores and Figures as well because these are easier to interpret.
Hypotheses testing
Exposure
Before testing the hypotheses, it was tested whether there was an effect of exposure to the ad on the participants’ implicit attitude toward the advertised brand. This was done by means of a repeated measures-ANOVA (Analysis of Variance) with the exposure conditions to the target ad (positive, negative and neutral conditions) versus non-exposure to the target ad (control condition) on implicit attitude toward Evian and Bar-le-Duc. However, there was no significant effect on implicit attitude toward the advertised brand (F (1, 208) = .64; p = .425), meaning that exposure to the target ad does not affect implicit attitude toward the advertised brand. This might be due to the amount of participants in the groups (exposure: n = 155 and non-exposure: n = 55). However, this pattern endures even when the same test is carried out when solely the neutral condition (exposure to target ad: n = 52) and control condition (no exposure to target ad: n = 55) are used (F (1, 105) = .39; p = .532). In other words, there is no effect of exposure to the ad on implicit attitude toward the advertised brand. This means that it cannot be assured that the ST-IATs were the correct measurements for this manipulation.
In addition, another repeated measures-ANOVA with the exposure conditions to the target ad versus non-exposure to the target ad on explicit attitude toward Evian and Bar-le-Duc was carried out. However, no significant exposure effect was found for explicit attitude toward the advertised brand as well (F (1, 214) = .88, p = .349).
Context effect on implicit attitude
From now on, only the positive, negative and neutral conditions were used in the
measurements, which means N = 161 for explicit and N = 155 for implicit attitude toward the advertised brand, given the exclusion of the 11 participants to meet the ST-IAT criteria. The first hypothesis suggested that the valence of the news messages had an influence on people’s implicit attitude toward the advertised brand, meaning that a positive (negative) context leads to a more positive (negative) implicit brand attitude than neutral news
messages. This hypothesis was tested by means of a one-way ANOVA with condition as a between-subjects factor and the log transform difference of implicit attitude toward Evian as a dependent variable. According to this test, there was no significant main effect of the advertisements’ context on people’s implicit attitude toward the advertised brand (F (2, 152) = .625; p = .430). Thus, the first hypothesis is rejected. It cannot be concluded that implicit attitude toward the advertised brand is higher when people are exposed to a positive news message (M = 27.48; SD = 232.99) than to a neutral (M = 10.48; SD = 185.35) or negative news message (M = 29.24; SD = 169.35), see Figure 3. In other words, the context of online ads does not influence implicit attitude toward the advertised brand.
Context effect on explicit attitude
To see whether there was a main effect of context on explicit attitude toward the advertised brand, an ANOVA with ‘condition’ as a between-subjects factor and ‘explicit attitude toward
Evian’ as a dependent variable was carried out. The one-way ANOVA showed that the effect of a valenced context of the online ad on people’s explicit attitude toward the advertised brand was not significant (F (2, 157) = 1.82; p = .165). The participants that were exposed to the positive news message did not evaluate the advertisement significantly more favorably (M = 4.19; SD = 1.03) than the participants that were exposed to the neutral news message (M = 4.17; SD = .71) and negative news message (M = 3.80; SD = 1.32). In other words, the valenced context of online ads does not influence explicit attitude toward the advertised brand.
Figure 3
Effect of context on implicit attitude toward the advertised brand (Evian) per condition
Moderation by persuasion knowledge on implicit brand attitude
The second hypothesis stated that the relationship between the context of online ads and explicit (but not implicit) brand attitude is influenced by persuasion knowledge. A repeated-measures ANOVA was carried out with ‘condition’ and ‘persuasion knowledge’ as a
between-subjects factor and the log transform variable of ‘implicit attitude toward Evian’ as a dependent variable. As stated in the second hypothesis, there is no significant interaction
0 5 10 15 20 25 30 Im pli cit a tt it ud e to w a rd the a dv er tis ed bra nd Experimental conditions Positive condition Neutral condition Negative condition
effect of persuasion knowledge in the relationship of context on implicit brand attitude toward the advertised brand (F (2, 149) = .353; p = .703), see Table 4.
Table 4
Mean scores of the moderation of persuasion knowledge on implicit attitude toward the
advertised brand
Condition PK
Scores on implicit attitude toward the advertised brand
N M SD Positive High PK 21 5.47 196.48 Low PK 29 43.42 258.45 Negative High PK 20 43.19 211.69 Low PK 33 20.78 140.72 Neutral High PK 25 11.62 212.57 Low PK 27 9.43 160.24
Note. The median split scale for PK (PK = persuasion knowledge) was used.
Moderation by persuasion knowledge on explicit brand attitude
A one-way ANOVA was conducted with ‘condition’ and ‘persuasion knowledge’ as a
between-subjects factor and ‘explicit attitude toward Evian’ as a dependent variable. A strong significant main effect of persuasion knowledge on explicit attitude toward Evian was found (F (1, 155) = 15.03; p < .001; η² = .09), suggesting that people with high persuasion
knowledge (more skepticism toward advertising in general) express a more negative explicit brand attitude. However, this test did not show a significant interaction effect of context and persuasion knowledge on explicit attitude toward Evian (F (2, 155) = .15; p = .858),
suggesting that the effect of context on explicit attitude toward the advertised brand is not moderated by persuasion knowledge, see Figure 4.
Figure 4
Mean scores on of the moderation of persuasion knowledge on explicit attitude toward the
advertised brand (Evian)
Note. The scale of explicit attitude toward Evian ranged from 1 “unfavorable attitude” to 7 “favorable attitude”.
Interestingly, a small significant interaction effect of context and persuasion knowledge on explicit attitude toward the advertisement was found (F (2, 155) = 4.06; p = .019; η² = .05). This finding reveals that the main effect of context on explicit attitude toward the ad differs for people who are high in persuasion knowledge compared with those who are low in persuasion knowledge, see Figure 5. Post-hoc tests showed that the significant
interaction is attributable to the negative condition. Independent-Samples T Tests show a significant difference between the negative (M = 3.05; SD = 1.40) and positive (M = 3.86; SD = 1.03) condition (t (40) = 2.14; p = .039, 95% CI [.04, 1.57], d = .66) and between the negative and neutral (M = 4.10; SD = .85) condition (t (46) = -3.22; p = .002, 95% CI [-1.71, -.39], d = .91). This finding means that when persuasion knowledge is high, people in the negative condition evaluated the advertisement significantly more negative than people in the positive and neutral condition. However, attitude toward the advertisement was not of main interest in this study and hence will not be extensively discussed.
0 1 2 3 4 5 Low PK High PK E xpli cit a tt it ud e to w a rd the a dv er tis ed bra nd Persuasion knowledge Positive condition Neutral condition Negative condition
Concluding, the second hypothesis is partially accepted, suggesting that persuasion knowledge does not moderate the relationship between context of online ads and implicit and explicit attitude toward the advertised brand.
Figure 5
Mean scores of the moderation of persuasion knowledge on explicit attitude toward the
advertisement
Note. The scale of explicit attitude toward the ad ranged from 1 “unfavorable attitude” to 7
“favorable attitude”.
Discussion
The aim of this study was to investigate whether the emotional valence of the context of online ads influences people’s implicit and explicit brand attitudes and to explore if differences between these attitudes could be explained by their persuasion knowledge. In addition, a control condition was implemented in this study as well to measure an exposure effect. No implicit and explicit brand effects were found. However, interestingly, significant effects were found for the interaction of context with persuasion knowledge on explicit attitude toward the advertisement.
0 0,5 1 1,5 2 2,5 3 3,5 4 4,5 High PK Low PK E xpli cit a tt it ud e to w a rd the a dv er tis em ent Persuasion knowledge Positive condition Negative condition Neutral condition
The first hypothesis suggested that the valence of the context of online ads has an
influence on people’s implicit attitude toward the advertised brand. Based on multiple studies, it was assumed that context is automatically processed. However, this study showed no effect. Context, perceived as an online news message that surrounds an ad, has not been explained according to evaluative condition before. This study shows that evaluative
conditioning is not at work in the case of an online context: the transfer of feelings evoked by the unconditioned stimulus –context of the ad– to the conditioned stimulus –the
advertisement– did not (measurably) take place within this study. In other words, it might be possible that the ad’s online context is not suitable for evaluative conditioning.
Evaluative conditioning effects exist within product placement studies. For example, Redker et al. (2013) have shown that product placements were evaluated more positively when the viewers liked the movie than when they did not like it. The current study tried to generalize these effects to an online news context, which did not succeed. This might be the case for various reasons. First, liking a certain character, a movie or TV show is completely different than being exposed to a positive or negative context. For example, news might be positive, but someone could still not like the news message or the other way around.
Second, research has shown that negative news has a negative effect on people’s state of mind in terms of perceived security and safety (Johnston & Davey, 1997; Soroka, 2006). Johnston and Davey (1997) concluded that when people watched a predominantly negatively valenced news program, self-reported measures of anxiousness and sad mood raised. But, self-reported measures of anxiousness or sad mood are different than those of brand attitudes. Attitudes towards brands have nothing to do with people’s feelings of safety. For this reason, it might be possible that consumer’s brand attitudes were not influenced by the news
whereas news can be a threat for their feelings of safety (Bearden, Hardesty & Rose, 2001; Johnston & Davey, 1997; Soroka, 2006).
Third, it might be the case that news messages are processed differently relative to movies (Redker et al., 2013). The current study showed that it might be the case that an editorial factor, such as the valence of news messages, does not evoke the same strong kind of feelings as movies do. It may well be that news messages in a printed format do not have such strong impact on people’s state of mind. This is in accordance with a study of Dijkstra, Buijtels and Van Raaij (2005). They introduced the concept of “modality” which refers to the media’s mode of presentation (i.e. text, audio or pictures) that corresponds to human senses used for processing the presented material. The more sensory functions a medium provides, the easier it is to learn or recognize information (Kisielius & Sternthal, 1984). In other words, feelings are evoked easier by movies than by texts. Further research should evince whether this truly is the case by for example comparing the effects of an online context in terms of news messages and news videos.
The lack of effects for explicit attitude toward the advertised brand is in line with a study of Gibson (2008). He showed that evaluative conditioning does not affect explicit attitudes. In addition, research has shown that evaluative condition does not affect explicit attitudes for familiar stimuli (Cacioppo, Marshall-Goodell, Tassinary & Petty, 1992; Shimp, Stuart & Engle, 1991), which Evian and Bar-le-Duc obviously are. In the current study, all participants indicated to have heard of both brands before they took part in the experiment. This suggests that further research should consider the use of unfamiliar brands to show an evaluative conditioning effect of context on explicit attitude.
The second hypothesis stated that the relationship between the ad’s context and explicit (but not implicit) brand attitude is influenced by persuasion knowledge, suggesting that a valence effect exists for participants who are low in persuasion knowledge compared to
those who are high in persuasion knowledge. This hypothesis can partially be accepted, because, in accordance with a study of Gibson et al. (2014), the current study showed that implicit attitudes are not influenced by persuasion knowledge. Gibson and his colleagues (2014) concluded that when people were exposed to a liked character in a movie, implicit attitude increases, whether persuasion knowledge was activated or not.
The opposite of what was hypothesized in the second hypothesis, explicit brand attitudes were not influenced by the interaction of context and persuasion knowledge. In others words, the main effect of a valenced context on explicit and implicit attitude toward the advertised brand does not differ for people with high and low persuasion knowledge. However, in line with a study of Boerman et al. (2012) a main effect of persuasion knowledge on explicit brand attitude was found, suggesting that the higher someone’s persuasion knowledge, the more skepticism toward advertising and the more negative their explicit attitude is toward the advertised brand. In other words, explicit brand attitudes suffer from negative perceptions generated by the recognition of advertising (Gibson et al., 2014).
Interestingly, an interaction effect was found for explicit attitude toward the
advertisement, suggesting that the main effect of the valenced context of online ads on
explicit attitude toward the ad differs for people with high and low persuasion knowledge. The current study showed that people with high persuasion knowledge evaluated the advertised brand (Evian) more negatively when they were exposed to the negative news message than the participants that were exposed to the positive and neutral news message. This research showed that attitudes towards ads are easier influenced by context than attitudes towards brands, which is in accordance with Yoo and MacInnis (2005).
Exposure effect
As a baseline, a control condition was implemented in this study to measure whether an exposure effect occurred on both implicit and explicit attitude toward the advertised brand. However, the participants that were exposed to the target advertisement did not evaluate the advertised brand (Evian) more positively than the participants that were not exposed to the target ad, both implicitly and explicitly. For this reason, it might be the possible that the manipulation of the valenced context was not strong enough.
The fact that no exposure effect was found can be explained according to a study conducted by Acar (2007). He showed that incidental exposure does not lead to great differences in people’s overall attitudes towards advertised products. In his study, no differences in attitudes between the control and exposure conditions were found when the participants had to indicate their preferences on a Likert-scale for a lemonade (the advertised) brand. However, when subjects had to choose between lemonade (advertised product) and cranberry juice, significantly more people choose the lemonade. In other words, Acar (2007) showed that direct questions such as evaluations and preferences do not show differences, but comparison questions do. Unfortunately, no such comparison question was implemented in this study, so further research should consider the use of comparison questions instead of direct questions to measure explicit brand attitudes.
Limitations and suggestions for further research
One major limitation in the present study is the fact that the ST-IATs did not work. This might be for numerous reasons. First, the valence of the unconditioned stimulus (context) might not be extreme enough. Although the pretests have shown that the used news messages were indeed positive, negative and neutral, it might be possible that the news messages were
not positive and negative enough to evoke the expected feelings. For further research, it is advised to use news messages of which the emotional valence is crystal clear.
Second, it might be possible that familiarity with the news topics influenced the results as well. Some participants indicated that they were already familiar with the news message. This familiarity could have restrained the feelings that the news message should have evoked. For further research, it is advised to use unfamiliar or even non-existing news messages to measure the effects, in order to prevent for news topic familiarity. Moreover, a professional designer should manipulate the news page to make sure that the amateuristic character of the stimulus material cannot distract people, which might have been the case in the current study.
Third, the fact that the ST-IATs did not work might be due to the online setting of the experiment. The participants might have completed the experiment while watching TV or studying in the canteen of the university. In other words, there could have been a lot of distractions. To measure implicit attitudes in online settings, it might be possible that it is better to use Affect Misattribution Procedures (AMP; Payne, Cheng, Govorun & Stewart, 2005) instead of IATs. AMP is an affective association test and is build upon the principle of evaluative priming. With this test, participants are exposed to subliminal primes that evoke positive or negative responses (just as in IAT research). However, AMPs differ from IATs because after the subliminal primes, the participants have to indicate a random Chinese sign as being favorable or unfavorable. The assumption behind this test is that the positive or negative feeling that is unconsciously evoked by the prime is associated with the Chinese signs, meaning that an unfavorable evaluation of the Chinese sign corresponds to an unfavorable attitude toward the prime (Payne et al., 2005). Future research that takes place online should consider the use of AMP instead of IATs.
Lastly, the use of familiar or unfamiliar brands is always something to consider, because both options have its pros and cons. If you have decided to choose unfamiliar brands, evaluative conditioning can be used to describe the effects for explicit attitudes, which does not apply for familiar brands (Cacioppo et al., 1991; Shimp et al., 1991). But, if a familiar brand is chosen, the frequency of ad exposures needs to be taken into account. The fact that this study did not found exposure effects for explicit attitude toward the brand might be due to the number of ad exposures. The participants in the current study were exposed to the ad once, whereas Campbell and Kellerman (2003) found that ads of familiar brands are
evaluated more positively after at least three repetitions. This finding is in accordance with the mere exposure hypothesis of Zajonc (2001). He showed that “frequency breeds
familiarity, which positively affects attitudes” (Acar, 2007, p. 47). It might be the case that this holds true for implicit brand attitudes as well. Further research should give more insight regarding the mere exposure hypothesis and implicit attitudes after one and multiple
incidental exposures.
In other words, various suggestions for improvement are provided, hoping that future research will further examine whether the proposed mechanism holds for advertising in an online context.
Implications
This study has one implication for research on implicit attitudes and advertising. According to this study, the valenced context of online ads does not affect people’s implicit and explicit attitude toward the advertised brand. For communication professionals, this finding means that the attitude toward their brand does not depend on the valence of the context in which their ad is placed. In other words, a negative news message does not have destroying effects for their brand if the ad is placed next to a negative news message and vice versa.
However, future research is certainly needed because theories show that an effect of context on implicit attitude is possible. It is assumable that the proposed effect did not exist because of the selected setting and/or the measurement method that was not optimal. This is in accordance with the rejection of the exposure hypothesis. Another reason why future research is needed is because of the important shift in the focus of advertising from traditional media toward more non-traditional online channels such as Internet (Gartner, 2013). Marketers are interested in what leads to the most positive or favorable effects for their brand or product and based on this study, no clear advice can be given.
Conclusion
The current study is an attempt to understand differences in people’s implicit and explicit brand attitudes caused by the valenced context of online ads. In conclusion, it can be said that the valenced context of online ads does not influence people’s implicit and explicit attitude toward the brand. However, theory suggests that it is likely to assume that there is such an effect for implicit brand attitudes. This study did not show these effects because of numerous possibilities, such as the study’s design or the questioned target group.
No interaction effect of context and persuasion knowledge was found for implicit and explicit brand attitudes as well. In other words, the main effect of context on explicit and implicit attitude toward the brand does not differ for people who are high and low in persuasion knowledge. However, persuasion knowledge moderates the relation between context and explicit attitude toward the advertisement, suggesting that people with high persuasion knowledge evaluated the advertisement more negative when they were in the negative condition than people in the positive and neutral condition.
Concluding, further research is needed to track whether the differences in the
valenced context of online ads do have an influence on people’s implicit (and explicit) brand attitude.
Acknowledgements
This thesis would not have been possible without my supervisor Dr. Lisa Vandeberg. I would like to express my deepest gratitude for her excellent guidance and help.
References
Acar, A. (2007). Testing the effects of incidental advertising exposure in online gaming environment. Journal of Interactive Advertising, 8(1), 45-56.
Bearden, W. O., Hardesty, D. M., & Rose, R. L. (2001). Consumer self-confidence: refinements in conceptualization and measurement. Journal of Consumer Research,
28(1), 121-134.
Bluemke, M., & Friese, M. (2008). Reliability and validity of the Single‐Target IAT (ST‐IAT): assessing automatic affect towards multiple attitude objects.
European Journal of Social Psychology, 38(6), 977-997.
Boerman, S. C., Van Reijmersdal, E. A., & Neijens, P. C. (2012). Sponsorship disclosure-effects of duration on persuasion knowledge and brand responses.
Journal of Communication, 62(6), 1047-1064.
Bronner, F., Bronner, J., & Faasse, J. (2007). In the mood for advertising. International
Journal of Advertising, 26(3), 333-355.
Bronner, F., & Neijens, P. (2006). Audience experiences of media context and
embedded advertising-A comparison of eight media. International Journal of Market
Research, 48(1), 81-100.
Cacioppo, J. T., Marhsall-Goodell, B. S., Tassinary, L. G., & Petty, R. E. (1992). Rudimentary determinants of attitudes: Classical conditioning is more effective when prior knowledge about the attitude stimulus is low than high. Journal of
Experimental Social Psychology, 28(3), 207-223.
Campbell, M. C., & Keller, K. L. (2003). Brand familiarity and advertising repetition effects. Journal of Consumer Research, 30(2), 292-304.
Chaiken, S., & Trope, Y. (Eds.). (1999). Dual-process theories in social psychology. Guilford Press.
Chung, H., & Zhao, X. (2003). Humour effect on memory and attitude: moderating role of product involvement. International Journal of Advertising, 22(1), 117-144.
Conner, M., & Sparks, P. (2005). Theory of planned behaviour and health behavior. In M. Conner & P. Norman (Eds.), Predicting health behaviour: Research and
practice with social cognition models (2nd ed., pp. 170-222). Maidenhead, UK: Open
University Press.
Cooper, H., Schembri, S., & Miller, D. (2010). Brand-self identity narratives in the James Bond movies. Psychology & Marketing, 27(6), 557-567.
Coulter, K, S. (1998). The effects of affective responses to media context on advertising evaluations. Journal of Advertising, 27(4), 41-51.
Cowley, E., & Barron, C. (2008). When product placement goes wrong: The effects of program liking and placement prominence. Journal of Advertising, 37(1), 89-98. De Houwer, J. (2007). A conceptual and theoretical analysis of evaluative conditioning. The
Spanish Journal of Psychology, 10(02), 230-241.
De Houwer, J., Custers, R., & De Clercq, A. (2006). Do smokers have a negative implicit attitude toward smoking?. Cognition and Emotion, 20(8), 1274-1284.
Drèze, X., & Hussherr, F. X. (2003). Internet advertising: Is anybody watching?. Journal of
Interactive Marketing, 17(4), 8-23.
Dijkstra, M., Buijtels, H. E., & van Raaij, W. F. (2005). Separate and joint effects of medium type on consumer responses: a comparison of television, print, and the Internet. Journal of Business Research, 58(3), 377-386.
Friestad, M., & Wright, P. (1994). The persuasion knowledge model: How people cope with persuasion attempts. Journal of consumer research, 21(1), 1-31.
Gartner (March 6, 2013). Key findings from U.S. digital marketing spending survey 2013. Retrieved on February 10, 2015 from
http://www.gartner.com/technology/research/digital-marketing/digitalmarketing spend-report.jsp
Gawronski, B., & Bodenhausen, G. V. (2006). Associative and propositional processes in evaluation: an integrative review of implicit and explicit attitude change.
Psychological bulletin, 132(5), 692-731.
Gibson, B. (2008). Can evaluative conditioning change attitudes toward mature brands? New evidence from the Implicit Association Test. Journal of Consumer Research, 35(1), 178-188.
Gibson, B., Redker, C., & Zimmerman, I. (2014). Conscious and nonconscious effects of product placement: Brand recall and active persuasion knowledge affect brand attitudes and brand self-identification differently. Psychology of Popular Media
Culture, 3(1), 19-37.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition: attitudes, self-esteem, and stereotypes. Psychological review, 102(1), 4-27.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of
Personality and Social Psychology, 74(6), 1464-1480.
Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality
and Social Psychology, 85(2), 197-216.
Grimes, A., & Kitchen, P. J. (2007). Researching mere exposure effects to advertising theoretical foundations and methodological implications. International Journal of
Market Research, 49(2), 191-219.
Ha, L. (2008). Online advertising research in advertising journals: A review. Journal of
Hibbert, S., Smith, A., Davies, A., & Ireland, F. (2007). Guilt appeals: Persuasion knowledge and charitable giving. Psychology & Marketing, 24(8), 723-742.
Hummert, M. L., Garstka, T. A., O'Brien, L. T., Greenwald, A. G., & Mellott, D. S. (2002). Using the implicit association test to measure age differences in implicit social cognitions. Psychology and aging, 17(3), 482.
Johnston, W. M., & Davey, G. C. (1997). The psychological impact of negative TV news bulletins: The catastrophizing of personal worries. British Journal of
Psychology, 88(1), 85-91.
Kisielius, J., & Sternthal, B. (1984). Detecting and explaining vividness effects in attitudinal judgments. Journal of Marketing Research, 21(1), 54-64.
Lappe, N. (October 12, 2013) How much should you budget for marketing in 2014?
Webstrategies: driving your business success online. Retrieved on February 10, 2015 from http://www.webstrategiesinc.com/blog/how-much-budget-for online-marketing in-2014
MacKenzie, S. B., & Lutz, R. J. (1989). An empirical examination of the structural antecedents of attitude toward the ad in an advertising pretesting context. The
Journal of Marketing, 53(2), 48-65.
Matthes, J., Schemer, C., & Wirth, W. (2007). More than meets the eye: Investigating the hidden impact of brand placements in television magazines. International Journal of
Advertising, 26(4), 477-503.
Mangleburg, T. F., & Bristol, T. (1998). Socialization and adolescents' skepticism toward advertising. Journal of Advertising, 27(3), 11-21.
Moorman, M., Neijens., P.C., & Smit, E.G. (2002). The effects of magazine-inducted
psychological responses and thematic congruence on memory and attitude toward the ad in a real-life setting. Journal of Advertising, 31(4), 27-40.
Moorman, M., Neijens, P. C., & Smit, E. G. (2007). The effects of program
involvement on commercial exposure and recall in a naturalistic setting. Journal of
Advertising, 36(1), 121-137.
Murphy, S. T., & Zajonc, R. B. (1993). Affect, cognition, and awareness: affective priming with optimal and suboptimal stimulus exposures. Journal of
Personality and Social Psychology, 64(5), 723.
Nairn, A., & Fine, C. (2008). Who’s messing with my mind? The implications of dual process models for the ethics of advertising to children. International Journal of
Advertising, 27(3), 447-470.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know: Verbal reports on mental processes. Psychological review, 84(3), 231-259.
Nosek, B. A. (2007). Implicit-explicit relations. Current directions in psychological
science, 16(2), 65-69.
Payne, B. K., Cheng, C. M., Govorun, O. & Stewart, B. D. (2005). Marketing actions can modulate neural representations of experienced pleasantness. Proceedings
of the National Academy of Sciences of the United States of America, 105
(3), 1050-1054.
PWC. (June 4, 2014). Global ad trends: digital ‘to account for third of all revenue by 2018’. Global entertainment and media outlook. Retrieved on February 10, 2015 from
http://www.digitalstrategyconsulting.com/intelligence/2014/06/global_ad_tre ds_digital_to_account_for_third_of_all_revenue_by_2018.php
Ranganath, K. A., & Nosek, B. A. (2008). Implicit attitude generalization occurs
immediately; explicit attitude generalization takes time. Psychological Science, 19(3), 249-254.
Redker, C., Gibson, B., & Zimmerman, I. (2013). Liking of Movie Genre Alters the Effectiveness of Background Product Placements. Basic and Applied Social
Psychology, 35(3), 249-255.
Resnick, M., & Albert, W. (2014). The impact of advertising location and user task on the emergence of banner ad blindness: an eye-tracking study. International Journal of
Human-Computer Interaction, 30(3), 206-219.
Russell, C. A. (2002). Investigating the effectiveness of product placements in television shows: The role of modality and plot connection congruence on brand memory and attitude. Journal of Consumer Research, 29(3), 306-318.
Soroka, S. N. (2006). Good news and bad news: Asymmetric responses to economic information. Journal of Politics, 68(2), 372-385.
Shimp, T. A., Stuart, E. W., & Engle, R. W. (1991). A program of classical
conditioning experiments testing variations in the conditioned stimulus and context. Journal of Consumer Research, 18(1), 1-12.
Wigboldus, D. H. J., Holland, R. W., & van Knippenberg, A. (2004). Single Target
Implicit Associations. Manuscript submitted for publication.
Yang, M., & Roskos-Ewoldsen, D. R. (2007). The effectiveness of brand placements in the movies: Levels of placements, explicit and implicit memory, and brand choice behavior. Journal of Communication, 57(3), 469-489.
Yoo, C. Y. (2008). Unconscious processing of Web advertising: Effects on implicit memory, attitude toward the brand, and consideration set. Journal of Interactive
Marketing, 22(2), 2-18.
Yoo, C., & MacInnis, D. (2005). The brand attitude formation process of emotional and informational ads. Journal of Business Research, 58(10), 1397-406.
Zajonc, R. B. (2001). Mere exposure: A gateway to the subliminal. Current directions in
Appendix A. Stimulus materials Condition 1. Negative
Condition 3. Neutral
