Winning or losing the game, not just important for the sports team:
How the valence of program content could affect implicit and explicit
attitudes towards subtly placed brands.
Content: Research Master Thesis Student: Kim Zoon, 6148824
Supervisor: Dr. Annemarie M. Wennekers
Institution: Graduate School of Communication, University of Amsterdam Date: June 27th, 2014
1
Abstract
Objective: The current study aimed to investigate the impact of the valence of television
program content on brand placement effects, to determine whether program valence affects the implicit and/or explicit brand attitude towards the subtly integrated brand. Consequently, the influence of both implicit and explicit brand attitude on spontaneous selecting behavior was studied. Method: 203 students participated in a 2 valence (positive versus negative) x2 placement (present versus absent) x2 brand (Evian versus Sourcy) mixed experimental design, where the first two variables were between-subjects variables and the latter was a within-subjects variable. After seeing one of the manipulated television compilations, the implicit brand attitude was measured using two ST-IATs, one for each brand. Explicit brand attitude, brand memory and consumer behavior were measured as well. Results: The implicit attitude towards Evian was generally more positive than for Sourcy. This difference was larger in the positive valence conditions, independent of placement. Exploratory analyses show a
moderating role of recall. The explicit attitude towards Evian was generally higher than toward Sourcy, regardless of placement and valence. The spontaneous consumer behavior measure showed a significant influence of the explicit attitude and insignificant influence of the implicit attitude. Conclusion: Irrespectively of the absence or presence of the brand placement, there is a small indication for a valence effect on the implicit brand attitude. Furthermore, preliminary findings suggest that brand recall can have a larger influence than previously expected. What is made clear is that implicit-level processes should be taken seriously, as a lack of findings on an explicit level does not mean that nothing happens at all. Both practitioners as well as researchers should take implicit-level effects into account to get a more complete view on what is really happening.
2
On a daily basis, consumers are flooded by a large amount of commercials, advertisements and sponsored content included in mass media. Newspapers, radio
programming, websites, social network sites; they all contain commercial content. Focusing on television alone, the figures speak for themselves: The average TV consumption in the Netherlands has remained relatively stable over the past few years (a daily 3 hours, 11 minutes in 2011 compared to 3 hours, 15 minutes in 2013; Stichting Kijkonderzoek, 2012; 2014). Nevertheless, the commercial exposure has grown from 5 minutes in 1990 to 17 minutes in 2011 (Moorman, Neijens & Smit, 2011). As the advertising clutter is still rapidly increasing, it is very plausible that the 17 minutes of 2011 is an underestimation of the exposure in 2014.
Moorman et al. (2011) describe this advertising clutter as a problematic aspect of the current advertising landscape for advertisers. They advise to find new ways to get the
commercial message across. There is probably no area more creative and resourceful than the advertising industry, so for this problem, a creative solution was found: commercial content is merged into the editorial content. This results in different solutions for different media
vehicles. On television, this found shape in the form of brand placements. Brand placements are defined as “the compensated inclusion of brands or brand identifiers within media programming” (Van Reijmersdal, 2009, p. 151).
Nowadays, brand placements are well-known additions to television programming and movies. The most popular Dutch soap opera “Goede Tijden, Slechte Tijden” is a good
example. It is filmed in a set that clearly contains Rivièra Maison home furniture; the actors order meals using Thuisbezorgd.nl and eat a Kanzi-branded apple in the meantime. The widespread use of brand placements is easy to explain, as it has a lot of important advantages for the advertiser. Consumers that are watching the program or movie do not switch to another channel if they see the brand placement, as they often do with the start of a
3
commercial block. Furthermore, research has shown that brand placement is related to improved brand recall and recognition (Law & Braun, 2000), positive brand image change (Van Reijmersdal, Neijens & Smit, 2007), and changes in brand attitude (Van Reijmersdal, 2009).
The impact of these findings is undisputed. However, brand placements may even have more influence than currently assumed. Existing research on brand placement effects report findings on brand memory, image and attitude, all measured in explicit ways. A central theme in recent research on evaluative processes is the distinction between deliberate, explicit attitudes and automatic, implicit attitudes. Gawronski and Bodenhausen (2006) note that changes in implicit attitudes are still largely unexplained, because of the shortfall that several models of persuasion and attitude change, such as the classical elaboration likelihood model of Petty and Cacioppo (1986), treat attitudes as a unitary construct, instead of separating implicit and explicit attitudes. Gawronski and Bodenhausen (2006) argue that explicit and implicit attitudes are in fact constructed and changed through their own distinct mental processes, as described in their associative-propositional evaluation (APE) model.
Implicit brand attitudes are especially interesting to study, as they are better predictors of more automatic, impulsive behaviors than explicit attitudes (Strack, Werth & Deutsch, 2006). Focusing on implicit attitudes in brand placement research would thus provide a useful addition to existing research (Gibson, 2008), especially when investigating subtle forms of brand placement. It is assumed that subtle brand placements are processed in a more
automatic way than prominent brand placements, as they are not a central element of the story line. This automatic processing makes it likely to affect implicit brand attitudes, whereas explicit attitudes have been found less susceptible to such automatic processes. However, until now, implicit brand attitudes are relatively under-studied. Therefore, Van Reijmersdal
4
(2009) encourages other scholars to further investigate the implicit attitude changes in the context of brand placements.
A question that is very relevant when it comes to studying implicit attitudes in the area of brand placement is to what extent the valence of the content of the program the brand is placed in, spills over to the implicit brand attitude. This is relevant because implicit attitudes are known to be susceptible to an evaluative context (Gibson, 2008). Furthermore, it would be very useful for advertisers to know if the valence of the program has an effect on the implicit brand attitude of the subtly placed brand. Should advertisers be worried if their products are placed in a scene wherein people lose, fight, or receive bad news? Or, from a positive perspective, will advertisers benefit from a placement in a scene wherein people are happy and celebrate?
The goal of the present study is to investigate whether the valence of the television program spills over to the implicit (and explicit) attitudes toward brands subtly placed within the program. The findings of the current study would add to the academic knowledge in the field, especially to the knowledge about implicit attitude changes in the context of brand placement. Furthermore, the findings would help advertising practitioners in making choices about their coming brand placement tactics.
Concluding, the current study aims to investigate two research questions: First, does the valence (positive versus negative) of media content spill over to the implicit attitude towards the subtly integrated brand? Additionally, implicit brand attitudes are seen as reliable predictors of automatic, impulsive behavior. Changes in the implicit brand attitude caused by subtle brand placements may have an effect on the impulsive consumer behavior of these brands. The second research question is therefore: Can the valence of media content influence spontaneous consumer behavior through implicit attitude change?
5
Theoretical background The effects of brand placements on memory and attitude
Brand placement is a popular practice in today’s television programming (Van Reijmersdal et al., 2007). This is not surprising, as research has shown how successful it can be. Television watchers see the brands integrated in a natural and credible context, and are not able to really avoid it, a common problem for advertisements in commercial blocks (Van Reijmersdal et al., 2007). Research has shown that brand placement is related to improved brand recall and recognition (Law & Braun, 2000), positive brand image change (Van Reijmersdal et al., 2007) and changes in brand attitude (Van Reijmersdal, 2009). Van Reijmersdal (2009) has
formulated two laws about brand placement effects on the two most important measures: (brand) memory and (brand) attitude. She has found that brand prominence has a positive effect on brand memory. However, when viewers were aware of the deliberate brand
placement, involved with the medium vehicle or liked the medium vehicle, brand prominence was found to have a negative effect on the attitude towards the brand. The study of Van Reijmersdal, Neijens and Smit (2009) gives an extensive overview of the findings in brand placement research.
However, focusing on the findings on brand attitude change, these results do not give a complete picture of brand placement effects. The findings rely on explicit measures, leaving the more associative aspect of the attitude-construct unstudied. Attitudes are defined as the psychological tendency to evaluate a given entity with some degree of favor or disfavor (De Houwer, Thomas & Baeyens, 2001). According to the associative-propositional evaluation (APE) model (Gawronski & Bodenhausen, 2006), this general term can be rooted in two different kinds of processes, resulting in two different kinds of attitudes: an explicit and an implicit attitude. Explicit attitudes are formed by propositional processes, which are based on inferences that occur in a reflective system. More concrete, this means that individuals actively reflect on the information, a process that requires awareness. This means that an
6
explicit attitude is formed during a propositional process of which the individual is
consciously aware. This makes it possible to measure explicit attitudes by self-assessment of the participant. However, according to the APE model, the explicit attitude coexists with an implicit attitude, which cannot be measured using these explicit measures.
Implicit-level processes contrasting explicit-level processes
In contrast to the propositional process as basis for the explicit attitude, the implicit attitude is based on an associative process that requires a low amount of cognitive capacity. It is an automatic evaluative reaction that is activated automatically upon encountering an attitude object. The association can be activated irrespectively of whether the individual that holds these associations finds them accurate or inaccurate.
An important aspect about the coexistence of the implicit and explicit attitudes is that they can be similar, but also very different from each other. An example of this can be found in associations towards minorities in society. The activation of negative associations can be high, even though the individual finds these associations false or undesirable on an explicit level. This discrepancy is caused by the difference in cognitive elaboration associated with both processes according to the APE model. The cognitive elaboration is the main
determinant of people’s success in suppressing the influence of automatic attitudes on evaluative judgments and in retrieving their explicit attitudes from memory. Moreover, the cognitive elaboration affects the complexity of propositional thinking by influencing the amount of relevant propositions one considers in addition to one’s automatic affective reaction. When there is no room for cognitive elaboration, associative processing will still take place. As associative processes are independent of truth-values, implicit attitudes do not have to be in line with explicit attitudes toward the same object. One can imagine that if people do not have the time or capacity to thoroughly think ones decisions through, one falls
7
back on the implicit attitude and behaves more instinctively, for instance, with impulse buying (Strack et al., 2006).
A question circling around the topic of implicit and explicit attitudes is the extent to which people are aware of them, especially of their implicit attitudes. There is no real
consensus on this topic. The possibility that some affective reactions are below the threshold of experiential awareness is not ruled out completely yet (Gawronski & Bodenhausen, 2006), but in general it is assumed that individuals have some level of awareness of their implicit affective reactions (associative process). Compared to explicit attitudes however, people are not always willing or able to explain the causes or consequences of their implicit attitudes. Thus, even though people might be aware of subtly placed brands in TV shows, they might not know how these cues affect their implicit brand attitudes. It is therefore only logical that implicit attitudes are not measured in the same way as explicit attitudes, by self-reports, but more commonly by tasks from which the implicit attitude can be deferred. An example of an implicit attitude measure is the Implicit Association Task (IAT).
Subtle brand placement effects
In the current study, the effects of subtle brand placements on the implicit and explicit brand attitude are studied. In general, there are two ways in which brand placements can be
integrated in a TV-show: prominently and subtly. A placement is considered prominent if the brand is the focus of the attention (Gupta & Lord, 1998, in Van Reijmersdal, 2009). Subtle brand placements can appear in the background of a scene (Redker, Gibson & Zimmerman, 2013), or in the periphery of the screen.
Subtle brand placements are not processed in the same manner as prominent brand placements. Subtle and prominent placements have different effects on brand memory and brand attitude (Van Reijmersdal, 2009). Prominent brand placement have a positive effect on the brand memory, but can have a negative influence on the brand attitude. Van Reijmersdal
8
(2009) explains that this paradoxical effect of prominence can be explained by the way in which prominent placements are processed. Prominent placements are processed more deeply, in line with the propositional processing style described in the APE model (Redker et al., 2013; Gawronski & Bodenhausen, 2006). This can lead to a better brand memory, but also activate persuasion awareness and resistance, resulting in more negative brand attitudes.
For subtle brand placements, the effects are not that clear yet. Van Reijmersdal (2009) and Van Reijmersdal et al. (2007) argue that especially implicit-level processes are important when studying the effects of subtle brand placement, as subtly placed brands do not receive much attention: The attention is drawn to the story line. This is in line with the
low-elaboration aspect of implicit attitudes, as described in the APE model (Gawronski &
Bodenhausen, 2006). According to this model, it can be expected that subtle brand placements are processed on an associative level, as this kind of processing is characterized by a low intention to evaluate the object, a low amount of cognitive capacity, low cognitive elaboration and an automatic nature. To get a complete picture of subtle brand placement effects, effects on implicit attitudes should be taken into account, or rather, be a central effect measure. It is expected that subtle brand placements have an effect on implicit brand attitudes, but not on explicit brand attitudes. These expectations are formulated as follows:
H1a. Exposure to subtly placed brand logos affects the implicit attitude towards the depicted brand.
H1b. Exposure to subtly placed brand logos does not affect the explicit attitude towards the depicted brand.
However, it is more important to know under what conditions subtle brand placements have an effect, and what determines these effects to be positive or negative. Redker et al. (2013) found that a positive versus negative viewing experience had an effect on
9
the implicit brand attitude towards subtly placed brands. Redker et al. (2013) pre-selected two groups of participants: haters and likers of science fiction, to get participants with a negative and positive viewing experience. They exposed both groups to a 43-minute fragment of the movie Blade Runner, which included four subtle placements of the brand Coke. They found that the sci-fi loving participants had more positive associations with Coke than sci-fi hating participants as a result of the subtle brand placement. Corresponding with the findings of Van Reijmersdal et al. (2007), the brand recall had no influence on this effect, confirming the implicit nature of the attitude change. Redker et al. (2013) call for future research that studies the effect of this implicit attitude change on the brand choice.
The findings of Redker et al. (2013) can be explained by the process that underlies implicit attitude change. The implicit attitude has an associative character that can be seen as an automatic reaction, which requires a low amount of involvement and reflection. This associative like or dislike of an object is something that people learn through conditioning. In traditional Pavlovian conditioning, a conditioned stimulus (CS) is paired with an
unconditioned stimulus (US), to explore whether the response towards the CS changes as a result of the pairing. In evaluative conditioning (EC), a neutral stimulus is paired with an affective stimulus, to explore if the affective stimulus changes the valence of the neutral stimulus. Here, the neutral stimulus is equivalent to the CS and the affective stimulus is equivalent to the US (De Houwer et al., 2001). EC is seen as the underlying process that can affect implicit attitudes. It is defined as the “change in the liking of a stimulus that results from pairing that stimulus with other positive or negative stimuli” (De Houwer et al., 2001, p. 853).
Redker et al. (2013) use the viewing experience as US, and define viewing experience by means of individual differences in liking of the movie genre. Their results show an EC effect on the implicit attitude in the context of brand placement. However, the practical
10
relevance of their findings is somewhat low: The chance that a sci-fi hater will watch a sci-fi movie, concluding in a negative implicit attitude towards the integrated brand, is probably negligible. Therefore, the current study uses the valence of the media content as
unconditioned stimulus. It is studied if the valence (positive versus negative) of the media content spills over to the implicit attitude towards the subtly integrated brand. Based on the evaluative conditioning process, the expectations are as follows:
H2a. Exposure to positively valenced media content including subtly placed brand logos will lead to a more positive implicit attitude towards the depicted brand, as compared to placement in a negatively valenced content and to the control brand.
H2b. Exposure to negatively valenced media content including subtly placed brand logos will lead to a more negative implicit attitude towards the depicted brand, as compared to
placement in a positively valenced content and to the control brand.
Implicit attitude change influencing behavior
Ultimately, the goal of advertising in general and brand placements specifically is not to simply change the attitudes towards the brand, but to stimulate purchase of the brand. The question arises if changes in the implicit brand attitude can affect purchase behavior. And, if the implicit and explicit attitude differ from each other, under what circumstances people will act according to their implicit attitude, instead of their explicit attitude.
To answer these questions, Wilson, Lindsey and Schooler (2000) developed the dual-attention model. They argue that an attitude can change over time, but this does not mean that the old attitude is replaced. Instead, it remains in memory. This can take shape in an implicit and an explicit attitude towards the same object, coexisting in the memory. Both are
expressed under different circumstances. The implicit attitude is activated automatically, while the explicit attitude required more capacity and motivation to retrieve. This means that
11
the implicit attitude is likely to influence what Wilson et al. (2000) call “implicit responses”: uncontrollable responses, such as nonverbal behaviors. Wilson et al. (2000) suggest that as the implicit attitude is habitual and automatic, it is the response people fall back upon when they do not have the capacity, motivation or time to retrieve the explicit attitude. On the other hand, this means that when the motivation and cognitive capacity is present, the explicit attitude will be retrieved and acted on.
Independently of Wilson et al. (2006), Strack et al. (2006) describe the same process, from a more economical perspective applied to consumer behavior. They describe that one of the circumstances in which consumers deviate from the rational buyer behaving process, is during impulse buying. This concept implies that a consumer acquires the wish to purchase on the spot, for immediate gratification. This contrasts “normal” buying behavior, which requires more reflection. Strack et al. (2006) combine both types of behavior in the
reflective-impulsive model (RIM). In this model, they distinguish an reflective-impulsive system from a reflective system, similar to the associative structure contrasting the propositional structure in the APE model (Gawronski & Bodenhausen, 2006). Interestingly, Strack et al. (2006) take this one step further and include the path to different types of behavior based on this dual system. If the consumer is motivated and the situation allows him to spend sufficient time and cognitive capacity, the buying will most likely be based on the reflective system. However, if the consumer is distracted, does not have enough cognitive capacity, is overly aroused, or under influence of alcohol or narcotics, the buying behavior is most likely to be predominantly determined by the impulsive/associative system.
Both studies point in the same direction: When the motivation is low, the individual is distracted, there is a low cognitive capacity and/or limited opportunity for reflection, the decision making process is guided by automatically activated behavior. In this case, the implicit attitudes are better predictors of the behavior than the explicit attitude.
12
Gibson (2008) studied the effects of implicit attitude change on brand selection
behavior, testing this on participants that were under high cognitive load. The participants got an eight-digit number and were asked to remember it. Continuing, they were asked to choose between a can of Pepsi and a can of Coke. After making the choice, they were asked to recall the number. Consequently, the participants got the can of their choice. They found that the participants that were under cognitive load, chose the can for which they had positive implicit attitudes.
In the current study, the brand selection behavior under cognitive pressure is studied to find out if implicit attitude change caused by subtle brand placements affects this kind of behavior. In doing so, the appeal of Redker et al. (2013) to study these effects, is followed up. The expectations are formulated in the following hypotheses:
H3a. Implicit brand attitudes affect selection behavior under cognitive load.
H3b. Explicit brand attitudes do not affect selection behavior under cognitive load.
Method
In one experimental study, sports compilations were created, varying in valence and in the presence or absence of subtle brand placements. In the brand placement conditions, the Evian brand logo was placed five times in the periphery of the screen. Sourcy was used as a control brand and did not occur in the compilations. After seeing one of the four compilations, the implicit brand attitude was measured by two ST-IATs, one for each brand, and the explicit brand attitude using self-report questions.
Participants and design
A total number of two hundred and three participants completed the study. Students were recruited from the University of Amsterdam, to participate for a small money fee, or research credits. One participant answered zero comprehension questions correctly; one participant
13
was an extreme outlier on all the three questions regarding the valence of the compilation (deviated two times more than 4 SD and one time more than 5 SD of the mean). Three
participants deviated more than three standard deviations of the average IAT score. These five participants were excluded from the study. This resulted in a sample of one hundred and ninety-eight participants (68.7% female, Mage = 21.67, SDage = 2.52).1
The experiment was set up as a 2 valence (positive versus negative) x2 brand placement (present versus absent) x2 type of brand (Evian versus Sourcy) mixed design, where the first two variables were manipulated between-subjects and the latter variable within-subjects. The participants were randomly allocated to one of the four different conditions. Participation took approximately 30 minutes.
Procedure and materials
The experiment took place in the ASCoR laboratory, a venue including multiple cubicles with computers featuring the experiment in Inquisit (Millisecond Software, 2012). After entering the room, the participants signed the informed consent form and were allocated to one of the available cubicles.
Manipulation materials. The experiment started with a few questions about
demographics, followed by the manipulation material: A compilation of important Dutch sport moments, featuring a Dutch (often well-known) sportsman, -woman or team. For the purpose of creating a positive and a negative compilation, the content was centered on losing and disappointment in the negative condition and winning and happiness in the positive condition.
Sport moments were chosen as the basis of the compilation because our previous work with reality TV-shows showed no decisive results to support the expected evaluative
conditioning effect of the valence of the program (Wennekers, Vandeberg, Zoon & Van
1
The results did not differ between the complete sample and the selected case sample. Only one analysis showed a difference, this is noted in the text.
14
Reijmersdal, unpublished manuscript). For the current study, a sports content was chosen, because it was expected to increase the strength of the valence of the program content and the involvement of the participants (Moorman, Neijens & Smit, 2007). A complete overview of the content of the compilations can be found in appendix A.
Aside from the difference in positive and negative content, it was made sure that the compilations were matched on several important aspects. First, for both conditions sport items were selected starting from 2010 till most recent. Three of the four clips were from the
Olympics; the fourth was from the World Soccer Championship (2010). Second, for both conditions, a similar number of male and female sports were selected. Three of the sport items per condition featured male athletes, one female. Third, both conditions had at least one individual athlete, complemented by team sports. Fourth, the items in both conditions contained images of the sport moment and a reflection moment. For the negative condition, this was a short interview moment, for the positive condition, this was a celebration moment. Finally, both compilations lasted around eight and a half minutes.
Similar to the study of Wennekers et al. (unpublished manuscript), the inclusion of the brand placements was operationalized by placing an Evian logo five times in the compilations for a duration of four seconds, using Adobe Premiere Elements. There was no placement in the opening compilation, as this was seen as a “warm up”. Every sport item contained at least one brand placement, one item contained two brand placements. The brand placements were evenly allocated through the duration of the clip. The logos were placed in the upper left- and upper right corner. The size of the brand placement was equal for all placements and both conditions, similar to the size of the network logo.
Measurements.
Implicit attitude. After seeing the clip, the participants performed two single-target
15
toward Sourcy, which is a comparable mineral water brand that was not depicted in the compilation. Sourcy served as a control brand, to be able to test whether the valence of the compilation caused a general positive or negative spillover effect, or a specific conditioning effect concentrated on the depicted brand.
The implicit association task of the current study existed of five blocks. Participants started with a practice block of 20 trials. During these trials, they had to categorize pictures and words as being negative or positive by pressing respectively the A and L key on the keyboard as fast and accurately as possible. A mix of words and pictures was used: Five positive pictures (happy smiley, heart, dove, thumbs up, smiling child) and five positive words (Dutch translations for the words healthy, party, tasty, beautiful, fun) were included, as well as five negative pictures (sad smiley, broken heart, poison warning label, thumbs down, crying child) and five negative words (Dutch translations for the words disease, disaster, dirty, ugly, grief).
Consequently, two experimental blocks of 40 trials each followed, in which
participants had to categorize both positive and negative words, as well as pictures of logos and products belonging to the brand Evian. In one block, Evian had to be categorized using the same key as positive stimuli (L-key), whereas in the other block Evian had to be
categorized using the same key as negative stimuli (A-key). The order of the allocation of Evian to either positive or negative was counterbalanced. The assumption is that the more positive people’s implicit attitudes are toward Evian, the faster they should be in the block with Evian and positive on the same key as compared to the block with Evian and negative on the same key. After the two Evian blocks, participants performed the same task for the control brand Sourcy. The two blocks had the same set-up as the Evian-blocks. The Sourcy-blocks were counterbalanced as well. In both the Evian and Sourcy Sourcy-blocks, the same positive and negative pictures and words were used as in the practice block.
16
Spontaneous behavioral intention. Consequently, the spontaneous behavioral
intention was measured. To this aim, the participants were put under cognitive load by giving them an 8-digit number to remember (Gibson, 2008). Next, they were told that they had to make a choice between the left item and the right item, shown on the next page. As time was limited, it was asked to already place the fingers on the Q-button to choose for the left item and the P-button to choose for the right item. The participants could choose between a half liter bottle of Evian and a half liter bottle of Sourcy, the control item, similar to the work of Gibson (2008). The left-right position of Evian and Sourcy was counterbalanced.
Explicit attitude. Afterwards, the explicit attitude towards Evian and Sourcy was
measured, using five seven-point semantic differential items answering the statement “To me, Evian/Sourcy is …” disliked-liked, negative-positive, unfavorable-favorable, uninteresting-interesting, unappealing-appealing (Chung & Zhao, 2004). The items proved to construct a reliable scale for both Evian (α = .88) and Sourcy (α = .90).
Memory. Afterwards, the recall and recognition was tested posing open questions
about the recall of brand logo’s, using the open question “Did you notice something special while seeing the compilation?”, and the yes/no question “Can you remember seeing brand logos while watching the compilation?” When answered yes, participants were asked to note the brands they have seen in a follow-up question. This was followed by a recognition task including the Evian logo, Sourcy logo, six filler logos and six brand logos naturally occurring in both compilations. Participants were asked for each logo to indicate by a yes or no, if they saw the brand logos in the compilation.
Remaining measures. Consequently, the comprehension of the compilations was
tested using six multiple-choice questions regarding the content of the clips. This served both as cover for the purpose of the study, as well as a check if the participants paid attention to the clip. This was followed by questions about the manipulation checks, all on a 7-point Likert
17
scales. First of all, the valence of the clip was measured with four questions, “To what extent did you find the valence of the clip positive/negative/neutral?”, where the answers ranged from 1 (not at all) to 7 (very much) and the question “How would you describe the general atmosphere of the compilation?”, where the answers ranged from 1 (very negative) till 7 (very positive). Consequently, the entertainment value and viewing experience were measured, with respectively five and four items 7-point items. The items constituted a reliable scale for both the entertainment value (α = .86) and the viewing experience (α = .79). The items measuring these variables can be found in appendix B. The study was concluded with questions about the interest in sports, identification with the Netherlands, the frequency of buying and using Evian and Sourcy, and open questions about their ideas of the purpose of the study and the different tasks.
Results Manipulation Check
The first manipulation check that was performed concerned the valence of the positive and negative compilation, to make sure that the compilations were perceived as intended. The manipulation showed to be successful: The positive compilation scored significantly higher (M = 6.52, SD = 0.79) on the perceived positivity scale than the negative compilation (M = 2.53, SD = 1.44), t(145.56) = 23.90, p < .001, 95% CI [3.66; 4.32]. The results on the negative scale confirmed this finding. Here, the negative compilation scored significantly higher (M = 4.91, SD = 1.82) than the positive compilation (M = 1.48, SD = 0.90), t(136.56) = -16.60, p < .001, 95% CI [-3.83; -3.02]. The general scale rating the valence of the compilations
confirmed that the positive compilation was perceived as significantly more positive (M = 6.46, SD = 0.77) than the negative compilation (M = 2.82, SD = 1.25), t(155.94) = 24.53, p <
18
.001, CI [3.35; 3.93].2 From these analyses, it can be concluded that the positive compilation was indeed perceived as positive and the negative compilations as negative, as intended.
As additional measures, the perceived entertainment value and viewing experience were inspected. Although both the positive and negative compilation were generally seen as entertaining, a significantly higher entertainment value was reported in the positive condition (M = 5.56, SD = 1.19) than in the negative condition (M = 5.04, SD = 1.21), t(196) = 3.03, p < .01. There was also a significant more positive viewing experience reported in the positive (M = 5.27, SD = 0.91), than in the negative condition (M = 4.70, SD = 0.93), t(196) = 4.40, p < .001.3
The second manipulation of the current study is the presence versus absence of the brand logo of Evian. Of the 101 participants that were allocated to the either one of the conditions that included the Evian logo, 55 participants (54.5%) spontaneously recalled to have seen the brand logo of Evian. There was no significant difference between the amount of people spontaneously recalling the Evian brand in the positive (43.6%) and in the negative logo condition (56.4%), χ² (2, N = 198) = 2.98, p = .27. Of the 97 participants that were allocated to one of the conditions that did not contain the Evian brand logo, 5 participants (5.2%) spontaneously indicated to have seen the Evian brand logo.
Implicit Attitudes
The implicit attitude measure consisted of two ST-IATs, one for Evian and one for Sourcy. After recoding incorrect responses and latencies faster than 300 milliseconds and slower than 3000 milliseconds as missing, the latencies were log transformed for each brand (Greenwald,
2 There was no significant difference found between the perceived valence of the negative compilation with and
without the Evian logo, according to the positivity, negativity and general valence scale. For the two positive compilations (with and without logo), there was a marginally significant difference found on the negative scale (Mwithoutlogo = 1.66 SD = 1.12; Mwithlogo = 1.31, SD = .58), t(72.86) = -1.99, p = .051. The differences on both the
positivity scale and general valence scale were not significant.
3 The logo and no-logo conditions of both groups did not differ significantly from each other for both variables.
Furthermore, there was no significant moderation of the entertainment value and viewing experience on the relation between the valence, placement and implicit and explicit attitude.
19
McGhee & Schwartz, 1998). These log latencies were used in the analysis, but to make the score better interpretable, the untransformed means are reported. Each ST-IAT consisted of a counterbalanced negative and positive block, resulting in a positive and negative block mean log-latency score for each brand. For Evian and Sourcy individually, the mean positive block log-latency score was subtracted from the mean negative block log-latency score. This resulted in ST-IAT scores for each brand, wherein higher scores reflected stronger positive implicit associations.
To test whether the valence and subtle placement of the Evian logo affected the implicit brand attitudes, a mixed ANOVA with as within-subjects variable the type of IAT and as between-subjects variables the placement condition and the valence condition was conducted. This analysis revealed a significant main effect of the type of IAT, F(1, 194) = 8.08, p < .01, ηp2 = .04. The implicit attitude towards Evian was significantly more positive
(M = 28.92, SD = 4.84) than the implicit attitude towards Sourcy (M = 10.81, SD = 3.76). There were no significant effects found for the placement, F(1, 194) < 1, ns, or valence condition F(1, 194) = 2.50, p = .12. The interaction between the type of IAT, placement condition and valence conditions also did not yield a significant effect, F(1, 194) < 1, p = .54. All figures can be found in Table 1.
20 Table 1
The Implicit Attitude towards Evian and Sourcy in Different Conditions
Evian Sourcy Mean SD Mean SD N Placement Positive 30.79 69.66 10.67 51.00 52 Negative 30.88 61.94 21.37 53.99 49 Total 30.83a 65.70 15.86a 52.48 101 No placement Positive 32.86 78.37 -1.74 50.77 50 Negative 21.17 60.15 12.96 55.83 47 Total 27.19a 70.02 5.38b 53.51 97 Total Positive 31.81a 73.69 4.59c 51.02 102 Negative 26.12a 60.95 17.25a 54.77 96 Total 29.05a 67.70 10.73c 53.12 198
Note. Untransformed means and standard deviations are presented in the table. Higher means represent a more positive implicit attitude.
Means on the same row that do not share subscripts differ at p < .05 (subscript b) and p < 0.01 (subscript c). Difference between the means of the same brand in placement- no placement conditions and positive – negative conditions do not differ significantly.
According to the recall-analysis, 54.5% of the participants in the placement condition spontaneously recalled having seen the Evian logo. Van Reijmersdal et al. (2007) and Redker et al. (2013) found that recall did not affect implicit brand attitudes. As an explorative
analysis, it is interesting to use the difference in recall to see how it affects the current findings. Therefore, the former analysis is repeated. Only now, the participants in the logo conditions are divided into two groups: participants that spontaneously recalled having seen Evian and participants that not recalled seeing Evian. For this purpose, the placement and recall variables are combined. This results in three groups of participants: participants in the no-logo condition, in the logo condition without recall and in the logo condition with recall.
The mixed ANOVA with type of IAT as within-subjects variable and placement/recall condition and valence condition as between-subject variables remains yielding a significant difference between the different type of IATs, F(1, 192) = 7.42, p < .01, ηp2 = .04. The
implicit attitude towards Evian is significantly more positive (M = 29.05, SD = 67.70) than the implicit attitude towards Sourcy (M = 10.73, SD = 53.12). There were no significant
21
interaction effects found for the placement/recall, F(2, 192) < 1, ns, or valence condition F(1, 192) = 1.41, p = .24. The interaction between the type of IAT, placement/recall condition and valence conditions also did not yield a significant effect, F(2, 192) = 2.39, p = .10. All figures can be found in Table 2.
Table 2
The Implicit Attitude towards Evian and Sourcy, Spontaneous Recall taken into Account
Evian Sourcy
Mean SD Mean SD N
No placement Positive 32.86 78.37 -1.74 50.77 50
Negative 21.17 60.15 12.96 55.83 47
Total 27.19a 70.02 5.38b 53.51 97
Placement, no recall Positive 15.12 70.24 15.57 51.05 28
Negative 53.67 81.80 17.25 41.84 18
Total 30.21a 76.49 16.22a 47.18 46
Placement, recall Positive 49.08 65.72 4.96 51.43 24
Negative 17.64 43.05 23.76 60.46 31
Total 31.36a 55.83 15.56a 56.97 55
Total Positive 31.81a 73.69 4.59c 51.02 102
Negative 26.12a 60.95 17.25a 54.77 96
Total 29.05a 67.70 10.73c 53.12 198
Note. Untransformed means and standard deviations are presented in the table. Higher means represent a more positive implicit attitude.
Means on the same row that do not share subscripts differ at p < .05 (subscript b) and p < .01 (subscript c). Difference between the means of the same brand in no placement – placement no recall – placement recall conditions and positive – negative conditions do not differ significantly.
For explorative reasons, a closer look is given to the participants in the placement conditions, with and without spontaneous recall. The mixed ANOVA with type of IAT as within-subjects variable and placement/recall condition and valence condition as between-subject variables, is repeated excluding participants in the no-placement condition. The analysis now yields a marginal effect of type of IAT, F(1, 97) = 3.47, p = .07, ηp2 = .04. The
interaction between the type of IAT and the placement/recall condition are still insignificant, both: F(1, 97) < 1, ns. The three-way interaction between the type of IAT, the
placement/recall condition and the valence is significant, F(1, 97) = 5.03, p < .05, ηp2 =.05.4
22
When interpreting this three-way interaction, the mean scores of the implicit attitude in the different conditions give a preliminary indication that when participants see a positive valenced content, the implicit attitude towards Evian is most positive if participants recalled seeing the brand, compared to the negative condition, F(1, 97) = 2.96, p = .09, the control brand F(1, 97) = 6.46, p < .05, and the implicit attitude towards Evian of the participants not recalling the logo F(1, 97) = 3.36, p = .07. When participants are exposed to the negative compilation, it seems that an opposite effect occurs. Now, the implicit brand attitude towards Evian is most positive for the participants not recalling the brand logo. However, when comparing this to participants recalling the brand logo, F(1,97) = 1.93, p = .16, the control brand, F(1, 97) = 1.79, p = .18, and the positive condition F(1, 97) = 2.26, p = .14, none of these differences are significant.5
Explicit Attitude
To analyze whether the placement and the valence of the compilation have an effect on the explicit attitudes towards Evian and Sourcy, two mixed ANOVAs were performed, with the explicit attitude towards Evian and Sourcy as within-subject variables and the placement condition and valence condition as between-subject factors. This was done for the complete sample first, and to stay in line with the explorative implicit attitude analyses, then with a sample where a distinction was made between the participants in the placement conditions that did and did not recall seeing Evian.
For the complete sample, a significant main effect of type of brand was found, F(1, 194) = 15.76, p < .001, ηp2 =.08. The explicit attitude towards Evian was generally higher (M
= 4.91, SD = 0.97) than the explicit attitude towards Sourcy (M = 4.60, SD = 1.03). Neither the placement condition, nor the valence condition, nor the interaction between these and the
5
The use of Evian and Sourcy and the extent to which the brands are known did not correlate significantly with the implicit brand attitudes and are therefore not included in the analysis.
23
type of brand had a significant effect on the explicit brand attitude to Evian and Sourcy, all: F(1, 194) < 1, ns. Figures for every condition can be found in Table 3.
Table 3.
The Explicit Attitude towards Evian and Sourcy in Different Conditions
Evian Sourcy Mean SD Mean SD N Placement Positive 4.93 1.02 4.63 1.19 52 Negative 4.93 1.00 4.58 .91 49 Total 4.93a 1.00 4.61c 1.06 101 No-placement Positive 5.11 .96 4.72 1.06 50 Negative 4.66 .89 4.45 .95 47 Total 4.89a .95 4.59c 1.01 97 Total Positive 5.02a .99 4.68c 1.12 102 Negative 4.80a .97 4.51b .92 96 Total 4.91a .97 4.60d 1.03 198
Note. Higher means represent a more positive explicit attitude.
Means on the same row that do not share subscripts differ at p < .05 (subscript b), p < .01 (subscript c) and p < .001 (subscript d).
Difference between the means of the same brand in placement- no placement conditions and positive – negative conditions do not differ significantly.
For the second mixed ANOVA where the recall is taken into account, the significant main effect of brand remained, F(1, 192) = 13.89, p < .001, ηp2 = .07. The explicit attitude
towards Evian (M = 4.91, SD = .97) was higher than the explicit attitude towards Sourcy (M = 4.60, SD = 1.03). The interaction between brand and valence was not significant, F(1, 192) < 1.0, ns. The interaction between brand and the placement/recall variable is marginally
significant, F(2, 192) = 2.60, p = .08, ηp2 = .03. There was no three-way interaction found
between the type of brand, the placement/recall condition and the valence, F(1, 192) < 1.0, ns. Figures for every condition can be found in Table 4.
24 Table 4
Explicit Attitude towards Evian and Sourcy, Spontaneous Recall taken into Account
Evian Sourcy Mean SD Mean SD N No placement Positive 5.11 .96 4,72 1.06 50 Negative 4.66 .89 4.45 .95 47 Total 4.89a .95 4.59b 1.01 97 Placement, Positive 4.74 .99 4.73 1.19 28 no recall Negative 5.03 1.13 4.91 .97 18 Total 4.85a 1.05 4.80*a 1.10 46
Placement, recall Positive 5.17 1.01 4.53 1.20 24
Negative 4.88 .92 4.38 .82 31
Total 5.00a .97 4.44*d 1.00 55
Total Positive 5.02a .99 4.68c 1.12 102
Negative 4.80a .95 4.51b .92 96
Total 4.91a .97 4.60d 1.03 198
Note. Higher means represent a more positive explicit attitude.
Means on the same row that do not share subscripts differ at p < .05 (subscript b), p < .01 (subscript c) and p < .001 (subscript d).
Difference between the means of the same brand in no placement- placement no recall – placement recall conditions and positive – negative conditions do not differ significantly, except the two means indicated with an asterisk (*). This represents a marginal significant difference of p = .08.
The mixed ANOVA was repeated, excluding the participants in the no placement conditions. The analysis yielded a significant main effect of brand, F(1, 97) = 8.88, p <.01, ηp2
= .08. The interaction between brand and valence and the three-way interaction between brand, valence and the placement/recall variable were both not significant, F(1, 97) < 1, ns. The interaction between brand and the placement/recall variable is now significant, F(1, 97) = 5.63, p < .05, ηp2 = .06. This interaction is mostly caused by a more negative explicit attitude
towards Sourcy, when participants recalled seeing Evian.
As a control test, it was checked if the implicit and explicit attitudes towards Evian and Sourcy were correlated with each other. This was not the case, not for Evian, r(196) = .08, ns, nor for Sourcy, r(196) = -.03, ns.6
6
The use of Evian and Sourcy and the extent to which the brands are known did not correlate significantly with both of the explicit brand attitudes and are therefore not included in the analysis.
25
Implicit Choice
To analyze the determinants influencing the implicit choice, a logistic regression analysis was performed with the outcome of the implicit choice, Evian or Sourcy, as
dependent variable. Descriptive analyses show that 130 participants (65.7%) chose the Evian bottle and 68 participants (34.3%) chose the Sourcy bottle during the choice test. The implicit attitude towards Evian and Sourcy (the log-latency scores) are used as independent variables. These factors did not have a significant influence on the implicit choice (p = ns). Consequently, the explicit attitude towards Evian and Sourcy were added to the analysis. These factors both had a significant influence on the implicit choice. For every point increase on the 7-point scale measuring the explicit attitude towards Evian, participants were 66.0% less likely to choose the Sourcy bottle during the implicit choice-test. The explicit attitude towards Sourcy was even more influential: for every increase on the 7-point scale measuring the explicit attitude towards Sourcy, the chance that someone will choose the Sourcy bottle is 240.0% more likely.
As control variables, the extent to which someone already uses Sourcy and Evian, and the extent to which one knows Evian and Sourcy were added to the logistic regression
analysis. The data showed that the implicit attitudes towards Sourcy and Evian still did not have a significant effect on the implicit choice. The explicit attitudes towards Evian (p < .001) and Sourcy (p < .01) remained significant. The extent to which one knows Sourcy and Evian did not have a significant effect on the implicit choice. The extent to which one uses Evian and Sourcy did have a significant influence: The more one uses Evian, the less likely (s)he is to choose for Sourcy during the implicit choice test (for every increase of use on the 5-point scale of using Evian, the chance of picking a can of Sourcy decreases with 69.8%) and the more one uses Sourcy, the more likely (s)he is to choose for Sourcy (the chance increases with 91.0%). When these factors are included in the analysis, 21.4% of the variance in the
26
implicit choice data can be explained. All the values of the logistic regression can be found in Table 5.
Table 5
Logistic Regression Analysis Predicting Implicit Choice from Implicit and Explicit Brand Attitudes, Including Covariates
95% Confidence interval for Odds Ratio
B SE Lower Odds Ratio Upper
Included Constant 0.16 2.84 Implicit attitude Evian 2.99 4.60 -6.03 19.83 12.00 Implicit attitude Sourcy -0.36 5.24 -10.63 0.69 9.91 Explicit attitude Evian -1.08*** 0.30 -1.66 0.34 -.50 Explicit attitude Sourcy 0.88** 0.29 0.31 2.40 1.44 Evian use -1.17** 0.38 -1.91 0.31 -0.42 Sourcy use 0.65* 0.30 0.06 1.91 1.24 Evian known 1.02 1.23 -1.40 2.76 3.43 Sourcy known -0.52 0.83 -2.16 0.59 1.11
Note. Implicit choice between Evian and Sourcy is the dependent variable. R2 = .21. N = 198. * p < .05; ** p < .01; *** p < .001.
Discussion
The aim of the current study was to investigate the influence of the valence of the media content on the implicit and explicit attitude towards a subtly integrated brand. Consequently, the influence of both these types of attitudes on the spontaneous selecting behavior was studied. For this purpose, four sports compilations were created, varying in valence and the absence or presence of the subtle brand placement. The Evian brand was used in the brand placement conditions, the control brand Sourcy was not placed, nor naturally occurring in the compilations. The results for the implicit attitude, explicit attitude and spontaneous brand choice are discussed subsequently.
27
Implicit brand attitude
The results show that there is more than one process that can be discerned on an implicit level, of which a few are made visible in the current study. For the implicit attitude, a
consistent difference was found between the target brand and the control brand. Irrespectively of the valence of the compilation and the absence or presence of the brand placement, the implicit attitude towards the target brand was evaluated more positive than towards the control brand. The difference between the implicit attitude towards the target brand and control brand is bigger in the positive conditions than in the negative conditions, indicating preliminary that the implicit preference of the target brand is stronger when people see positive television content than when they see negative television content. Since the implicit attitudes toward the target brand were always measured before those of the control brand (and directly after watching the clip), this might reflect a general valence transfer of the positive valence of the clip to the target brand (irrespective of placement).
The general spillover of valence (Moorman, Neijens & Smit, 2002) is an often-applied concept when studying context effects of advertising. Advertisements are evaluated in line with the valence of the context. The positive emotions generated by the compilations are transferred to the first object participants are exposed to after the clip. In this case, this was the Evian brand, as the implicit attitude measurement of Evian was always first. The results suggest that a valence spillover also has an effect on implicit processing, instead of just on an explicit processes as showed by previous studies (Moorman et al., 2002). Further studies need to be conducted to confirm these findings. It would be a very useful insight, as it cannot solely be applied to in-program brand placements, but especially to commercials placed after
television programs. If a program is positive, it would mean that this positive affect is transferred to the first brand seen in the first commercial.
However, this valence transfer stays absent in the negative conditions, a finding that is in line with the work of Waiguny, Nelson & Marko (2013), studying the effects of a negative
28
versus positive content of advergames on the explicit and implicit attitude towards the placed brand. On an implicit level, they only found a negative valence effect when the placed brand was unfamiliar, not when the brand was familiar. Although the study of Waiguny et al. (2013) uses games instead of a television compilation and does not contain no-placement conditions, the similarity between the two findings is remarkable. A possible explanation could be that both Evian and Sourcy are relatively well-known and have a positive brand image. A positive compilation enforces this positive image, while the negative compilation is too weak to have a negative influence, as the implicit attitude is too stable to become negative. Future studies are needed to point out what the effect is of a negative valence of the context on the implicit attitude towards unfamiliar brands. Based on the work of Waiguny et al. (2013), it is expected that the implicit attitude will become negative. However, according to Redker et al. (2013), an different hypothesis can be formulated: They describe that if an unfamiliar brand is used, they expect that the implicit attitude will carry over to the explicit attitude, in line with the
reasoning of the APE model.
A second explanation for the absence of a valence-effect in the negative conditions, is that the negative conditions were not negative enough. The manipulation checks show that the negative compilations are indeed perceived as negative, but they also show that the
entertainment value is still relatively high and viewing experience relatively positive. This means that even in the negative conditions, people still enjoyed watching the compilation. This is an alternative line of reasoning that would apply to both the findings of the current study and the findings of Waiguny et al. (2013), as participants liked playing the advergames that had a negative content. A challenge for future studies will therefore be to disconnect the valence of the content of the program and the entertainment value attached to it. It is possible that the valence of the content is not that influential after all and that the entertainment value
29
or viewing experience has more influence on implicit processes, as also indicated by Redker et al. (2013).
Explorative analyses show that there was a difference between the implicit brand attitude of people with and without brand recall in the brand placement conditions. For the people that recalled seeing the brand logo, the implicit attitude towards Evian became more in line with the valence of the program content, such that the implicit brand attitude was more positive in the positive condition than in the negative condition. For the control brand, the opposite happened. The implicit attitude towards the control brand was more negative in the positive condition than in the positive condition. When there was no recall of the placement, the implicit brand attitude became more positive in the negative condition than in the positive condition, while the implicit attitude towards the control brand was not affected.
In previous studies, Van Reijmersdal et al. (2007) and Redker et al. (2013) found no differences in implicit attitudes when taking recall into account. They argue that this is in line with the implicit nature of the attitude change. However, in the current study, the brand placements were in the peripheral areas of the screen. It can be argued that recalling the brand is an indication for actually seeing the brand, without explicitly or intensively processing it. This would explain the finding that with recall, the implicit brand attitude aligns with the valence of the compilation. However, the opposite finding for the people without recall cannot be explained. Further research is needed to get a clear perspective on the role of recall on implicit attitude change in a brand placement context.
On a methodological level, there is one strength and weakness at the same time of the current study that requires attention. The most interesting difference found in the current study, was between the used brand (Evian) and the control brand (Sourcy). The implicit attitude towards Evian and Sourcy was measured using two single target IATs. For future studies, this same approach is recommended to be able to make a clear distinction of what is
30
happening between the different brands. Making use of a regular IAT-test makes this impossible. In the current study, a deliberate choice was made to place the ST-IAT of Evian always first, as this was the target brand and therefore the main focus of the study. However, with this data, it is not possible to exclude a general valence spillover. Therefore, a possibility for future studies is to counterbalance the used brand and the control brand in IAT sequence. When both brands are counterbalanced, one can control for a general valence spillover effect.
Explicit attitude
For the explicit brand attitude, a consistent difference was found between the target brand and the control brand. Irrespective of the valence and the brand placement, the explicit attitude towards the target brand was more positive than towards the control brand. Exploratory analyses show that when the target brand is recalled, the explicit attitude towards the control brand becomes less positive compared to the target brand and compared to the explicit attitude towards the control brand when there is no recall.
The results found on the explicit level are quite as expected. The explicit attitude towards the target brand is in general more positive than towards the control brand. The exploratory analyses show an effect of recall on the explicit brand attitudes. Most importantly, there is no effect of the placement found on the explicit attitude, confirming that the subtle brand placements do not affect explicit attitudes.
Implicit choice
The implicit choice was only affected by the explicit brand attitude, not by the implicit brand attitude. The extent to which one uses Evian and Sourcy also influenced that implicit choice. Contrary to the expectations, the results of the current study show that during an implicit choice test, the explicit attitude is significantly influencing the choice, while the implicit attitude is not. This unexpected finding is most likely to be caused by an inadequate operationalization of the implicit choice test. In line with the work of Gibson (2008), the
31
participants were put under cognitive pressure to make sure that they were not able to elaborate on their choice. Therefore, they were asked to remember an 8-digit number while making their choice. However, it seems that this time, the cognitive pressure was not high enough while it was operationalized in the same manner as in the paper of Gibson (2008). A possible explanation is that there was more time between remembering the number and making the choice at the study of Gibson (2008) than at the current study. For future studies, it is therefore suggested to increase the number of choices that have to be made while
remembering the number, before the actual choice test. This increase in time is expected to increase the difficulty to remember the number, decreasing the mental capacities to make a deliberate choice even more.
Concluding remarks
The current study is a good step forward studying implicit level effects in the context of brand placement. The central theme of the study was the evaluative spillover of program valence on the implicit and explicit attitude towards subtly placed brands. Future research is needed to give more insight in how and when valence can spill over to implicit brand attitudes. Because after all, the real win for the brand does not take place on the field, but in the mind of the television viewer.
32
References
Chung, H., Zhao, X. (2004). Effects of perceived interactivity on web site preference and memory: role of personal motivation, Journal of Computer-Mediated Communication, 10(1), 00. DOI: 10.1111/j.1083-6101.2004.tb00232.x
De Houwer, J., Thomas, S. & Baeyens, F. (2001). Associative learning of likes and dislikes: A review of 25 years of research on human evaluative conditioning. Psychological Bulletin, 127(6), 853-869. DOI: 10.1037/0033-2909.127.6.853
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. DOI: 10.1037/0033-2909.132.5.692
Gibson, B. (2008). Can evaluative conditioning change attitudes towards mature brands? New evidence from the Implicit Association Task. Journal of Consumer Research, 35(1), 178-188. DOI: 10.1086/527341
Greenwald, A. G., McGhee, D. E. & Schwartz, J. K. L. (1998). Measuring individual
differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464-1480. DOI: 10.1037//0022-3514.74.6.1464 Law, S. & Braun, K. A. (2000). I’ll have what she’s having: Gauging the impact of product
placements on viewers. Psychology & Marketing, 17(12), 1059-1075. DOI: 10.1002/1520-6793(200012)17:12%3C1059::AID-MAR3%3E3.0.CO;2-V Moorman, M., Neijens, P. C. & Smit, E. G. (2002). The effects of magazine-induced
psychological responses and thematic congruence on memory and attitude toward the ad in a real-life setting. Journal of Advertising, 31(4), 27-40. DOI:
33
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. DOI: 10.2753/JOA0091-3367360109
Moorman, M., Neijens, P. C., & Smit, E. G., (2011). Advertising media selection and planning. In: J. Sheth, & N. Malhotra (eds.), Wiley International Encyclopedia of Marketing, 4, 22-29. DOI: 10.1002/9781444316568.wiem04010
Petty, R. E., & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. In L. Berkowitz (ed.), Advances in Experimental Social Psychology, 19, 123-205. New York: Academic Press. DOI: 10.1016/S0065-2601(08)60214-2
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. DOI: 10.1080/01973533.2013.785400
Stichting Kijkonderzoek (2012). Jaarrapport 2011. Retrieved from https://kijkonderzoek.nl/jaarrapport-kijkcijfers.
Stichting Kijkonderzoek (2014). Jaarrapport 2013. Retrieved from https://kijkonderzoek.nl/jaarrapport-kijkcijfers.
Strack, F., Werth, L & Deutsch, R. (2006). Reflective and impulsive determinants of consumer behavior. Journal of Consumer Psychology, 16(3), 205-216. DOI: 10.1207/s15327663jcp1603_2
Van Reijmersdal, E. A. (2009). Brand placement prominence: Good for memory! Bad for attitudes? Journal of Advertising Research, 151-153. DOI:
10.2501/S0021849909090199
Van Reijmersdal, E. A., Neijens, P. C. & Smit, E. G. (2009). A new branch of advertising: Reviewing factors that influence reactions to product placement. Journal of
34
Van Reijmersdal, E. A. Neijens, P. C. & Smit, E. G. (2007). Effects of television brand placement on brand image. Psychology & Marketing, 24(5), 403-420. DOI: 10.1002/mar.20166
Waiguny, M. K. J., Nelson, M. R. & Marko, B. (2013). How advergame content influences explicit and implicit brand attitudes: When violence spills over. Journal of
Advertising, 42(2-3), 155-169. DOI: 10.1080/00913367.2013.774590
Wennekers, A. M., Vandeberg, L. Zoon, K. & Van Reijmersdal, E. A. (2014). Subtle brand placements affect implicit but not explicit brand attitudes. Unpublished manuscript. Wilson, T. D., Lindsey, S. & Schooler, T. Y. (2000). A model of dual attitudes. Psychological
35
Appendix A. Content compilations
Part Content positive compilation Content negative compilation
Sports compilation
Energizing music, winning sportsmen and -women and in positive mood
Neutral music, losing sportsmen and -women, disappointment
Sport moment 1
Dutch men ice-skating team winning gold at team pursuit Olympics 2014
Nicolien Sauerbreij disappointingly reaching the 24th position
snowboarding at the Olympics of 2014 (while winning gold at the Olympics of 2010)
Sport moment 2
Dutch women hockey team winning gold at Olympics 2012
Sven Kramer ice-skating the 10 km at the Olympics of 2010, on track for gold, but sent in the wrong lane by his coach and therefore disqualified
Sport moment 3
Dutch soccer team winning quarter finals of Brazil at World Soccer Championship of 2010
Dutch male hockey team losing the finale of the 2010 Olympics from the German team
Sport moment 4
Epke Zonderland, winning gold at the Olympics of 2012 as gymnast on horizontal bar
Dutch soccer team losing the finale of the World Championship in 2010
36
Appendix B. Items measuring entertainment value and viewing experience
Construct Items
Entertainment value “I was interested in the compilation” “I watched the compilation attentively” “I found the compilation fascinating”
“I thought it was a pity that the compilation ended” “I sympathized with the athletes in the compilation” Viewing experience “I experienced watching the compilation as … stupid-fun”
“I experienced watching the compilation as … unpleasant-pleasant”
“I experienced watching the compilation as … boring-interesting”
“I experienced watching the compilation as … being too short-taking too long”
Interest in sports “I am interested in sports”
“I would consider myself a sports enthusiast” “I like watching sports programs on television” Identification with the
Netherlands
“I identify myself with the Netherlands”
“During big international tournaments, I feel connected with Dutch athletes”
