Ad-brand (in)congruent creative media choice : an examination of its effects on ad and brand attitude and recall, moderated by WOM

Ad-brand (in)congruent

creative media choice

An examination of its effects on ad and brand attitude and recall, moderated by WOM

Kimberley van der Krogt 10643249

Master’s thesis

Graduate school of Communication Master’s programme Communication Science

Persuasive Communication University of Amsterdam Supervisor: Ester de Waal

1 Abstract

The current research expands on previous research by broadening the concept of creative media choices (CMC) of Dahlén (2005) with respect to perceptual novelty and incongruency; hereby including traditional advertising forms and ad-brand incongruent advertising. By means of an online experimental survey, 173 participants were exposed to an existing ad-brand congruent or ad-ad-brand incongruent Durex advertisement, in order to research and compare the effects of ad-brand congruent CMC and ad-brand incongruent CMC on ad and brand attitude and (un)aided ad and brand recall. Furthermore, the research looks into a possible moderation effect of the extent of the nature of WOM – ranging from more positive to more negative. The research shows that an ad-brand congruent CMC has a more

pronounced effect than brand incongruent CMC on attitude towards the ad, but an ad-brand incongruent CMC has a more pronounced effect on aided ad recall than ad-ad-brand congruent CMC. There is no empirical support found for the moderating role of the nature of WOM, but additional analyses show that WOM has a partial mediation effect on ad attitude.

2

Introduction

To address the issues of declining effectiveness of traditional media (Van den Putte, 2009), ad clutter and its negative consequences on attitudes (Ha & McCann, 2008; Elliot & Speck, 1998; Edwards, Hi & Lee, 2002), marketers approach consumers with alternative advertising methods in a bid to enhance attitudes and recall (Van den Putte, 2009). Examples of such alternative advertising methods are creative advertising (Ang & Low, 2000; Ang, Lee & Leong, 2007; Dahlén, 2005; Dahlén, Friberg & Nilsson, 2009; Hutter & Hoffman, 2011; Nufer, 2013; Pieters, Warlop & Wedel, 2002; White & Smith, 2001) and ad-brand

incongruent advertising (Alden, Mukherjee & Hoyer, 2000; Lange & Dahlén, 2003; Dahlén, Lange, Sjödin & Törn, 2005). However, only a small amount of empirical studies of creative advertising and ad-brand incongruent advertising have addressed the issue of its effectiveness with respect to attitudes and recall, raising the question if the alternative advertising methods can enhance attitudes and recall as suggested by Van den Putte (2009).

Within the literature and practice of creative advertising, multiple forms and

definitions exist, with creative media choices, creative marketing and guerrilla marketing as most pronounced within the literature (hereafter referred to as creative advertising (CA) as umbrella concept). Creative media choices are defined as advertisements placed on a novel medium (i.e. previously unused as a medium) with the medium in itself carrying out the message (e.g. a salsa logo placed on a fire-extinguisher). Creative media choices are found to have positive effects on brand evaluations and associations (Dahlén, 2005; Dahlén et al., 2009) and recall (Dahlén et al., 2009).

Creative marketing and guerrilla marketing are both not clearly defined within the literature, but creative marketing is explained in terms of marketing that is novel; unexpected, divergent or unique. Subsequently, the novelty and creativity is meaningful, it can add value or meaning to the ad and brand (Ang & Low, 2000; Ang et al., 2007; Smith, MacKenzie,

3 Yang, Buccholz & Darley, 2007; Smith & Yang, 2004). Creative marketing has been found to have a positive influence on attitude towards the ad (Ang & Low, 2000; Smith et al., 2007), attitude towards the brand (Smith et al., 2007) and recall (Ang et al., 2007; Baack, Wilson & Till, 2008; Till & Baack, 2005).

Guerrilla marketing is most commonly referred to as a-typical, unconventional and novel (surprising, unexpected) advertising for conventional goals with monetary costs (e.g. flash mobs, gigantic or minimalistic objects and graffiti) (Hutter & Hoffman, 2011; Nufer, 2013). While the studies of creative media choice and creative advertising have shown positive influences on attitudes and recall, Hutter & Hoffman (2011) theorize that guerilla marketing enhances the processing of the ad and subsequently induces and enhances

diffusion by means of Word-of-Mouth (WOM) due to the creative and surprising element of the ad (Hutter & Hoffman, 2011).

Although the CA concepts are similar in definitions, they differentiate firstly in the appliance of novelty. Novelty is among all conceptualizations, but is applied in a literal manner within creative media choice; the use of a novel medium (e.g. a previous unused medium). While in creative marketing and guerrilla marketing novelty is applied on an attitudinal, perceiving level. Hence, creative media choice excludes traditional media and advertising forms (e.g. television, billboards), while creative and guerrilla marketing include traditional media, if the advertisement is perceived as novel.

Secondly, the CA concepts differ in the appliance of congruity. Dahlén (2005) argues that a creative media choice has to be ad-brand congruent (e.g. overlapping associations between ad and brand). While literature on creative marketing argues that ad-brand

incongruency can add novelty and meaningfulness to ads, such as that the disclosure of the brand or its slogan can add meaningfulness or novel associations to ad or brand.

4 and brand attitude and recall (Ang et al., 2007). However, the argumentations and

expectations of both the authors regarding (in)congruency, are based on assumptions and are not empirically researched. Therefore, considering the conflicting assumptions, the academic field could benefit and be expanded from ad-brand (in)congruency research.

Thus, the literature is divided in terms of defining CA, but solely Dahlén’s (2005) definition of creative media choice is delimited and clear and hence empirically researchable. However, as previously mentioned, Dahlén’s (2005) definition is in conflict with other CA literature regarding novelty and ad-brand congruency and is therefore arguably narrow. Building on this definition with the other conceptualizations of CA provides a more thorough understanding of the phenomenon of creative advertising with theoretical foundations.

Therefore, in this study, a new definition will be held when referring to creative media choice (CMC); unconventional, novel advertising by means of a medium where the medium carries

the message in itself.

With the novel CMC definition, the definition of creative media choice of Dahlén (2005) is extended regarding novelty on an attitudinal level, thus including traditional media and advertising platforms when used in a novel matter. An example of CMC within a

traditional medium in a novel matter can be found in image 1. The billboard carries the message in itself; it shows that the pimple is being cleansed and removed. The image and disclosure of the brand add relevancy to the ad and brand; the product cleans out pores thorough.

5

Image 1: Pond’s CMC on traditional medium; novel advertising where the medium carries the message. Retrieved from appealingstudio.com

Furthermore, the definition is extended regarding ad-brand (in)congruency, considering incongruency can arguably add value and positively enhance ad and brand attitude and recall (Ang et al., 2007). An example of this can be found in image 2. It is arguable that a crosswalk (a novel medium) and the cleaning detergent Mr. Proper separately do not share overlapping associations, hence it can be considered an ad-brand incongruent advertisement. However, when ad and medium are intertwined and provided with the disclosure (e.g. the brands’ logo), the cleaned stripe of the crosswalk carries the message of the ad in itself and adds value and meaning to the brand and ad; it cleans well and gets rid of all dirt.

.

6 With the intertwining of the CA concepts to a novel, broader definition of CMC, it raises the question to what extent the expanded definition has the positive influences on ad and brand attitude and recall and the stressed influence of WOM as theorized by the CA studies.

Furthermore, it raises the question if there will be a difference in effects between an ad-brand congruent CMC and an ad-brand incongruent CMC.

With respect to the influence of WOM, the literature is limited and often solely discusses theorizations of the effects of WOM regarding attitude. However, some research suggests that the nature of WOM (e.g. more positive or more negative) can have a positive effect on ad and brand attitude (Ferguson, 2008; Price, Feick & Higie, 1989; Sundaram & Webster, 1999). While recall has not been researched among WOM studies, WOM could arguably lead to an enhanced recall. Namely, when receiving WOM (either negative or positive), it denotes the ad for the consumer, it is arguable that when an ad is denoted, a better process of the ad will occur. Several studies argue that an enhanced ad process enhances recall (Phillips, 2000; Rajaram, Stinivas & Travers, 2001; Rosbergen, Pieters & Wedel, 1997). Although WOM so far has not been identified as moderator within the literature and with respect to CMC, this research expects that the nature of WOM will play a moderating role between CMC and ad and brand attitude and recall.

Thus, the focus of the research is to examine the differences between ad-brand incongruent CMC and ad-brand congruent CMC on ad and brand attitude and recall, with an examination of a possible moderation effect regarding the nature of WOM. By understanding this phenomena, both the academic field and practical field are enhanced. Namely, the academic field benefits insofar as this current study will build and extend on previous research, providing a greater and more thorough understanding of the relationship of ad-brand (in)congruent CMC and its effects. Practitioners can benefit in the way that the

7 findings of this study can assist with creating an effective creative advertising medium. The research question central in this study is:

RQ: To what extent does ad-brand (in)congruent CMC influence ad and brand attitude and recall and is this relation moderated by the nature of WOM?

Theoretical framework

Considering the definition of CMC proposed in the current study is novel, the theoretical foundation will build on the theorizations of the CA concepts; creative media choice, guerrilla marketing and creative marketing. Furthermore, considering ad-brand

(in)congruency intertwined with creative advertising is only briefly speculated (Ang et al., 2007; Dahlén, 2005), past literature on ad-brand (in)congruent advertising is also applied in order to explain the assumed relationship proposed in the research question and in order to formulate hypotheses. Within the literature of CA and ad-brand (in)congruent advertising the theories schemata theory (Alden et al., 2000; Ang & Low, 2000; Dahlén, 2005; Dahlén et al., 2005; Dahlén et al., 2009; Hutter & Hoffman, 2011; Phillips, 2000) and the guerrilla effect (Hutter & Hoffmann, 2011; Nufer, 2013) are most pronounced and most relevant for appliance to CMC.

Schemata theory in CA studies suggests that by priming ad schemata by means of textual cues (Dahlén, 2005; Dahlén et al., 2009), the creative ad will go around the general subject of schema-bases suspicion (Hoch, 2002; Stafford & Stafford, 2002). By means of priming, the creative ad challenges the consumers ad schemata. Hence, the ad is novel and surprising and therefore challenges and is inconsistent with the mold of ad schema (Ang & Low, 2000; Campbell & Goodstein, 2001; Goodstein, 1993; Stoltman in Ang & Low, 2000). By challenging the ad schemata of consumers, positive and negative ad and brand attitudes can occur. Namely, due to unfamiliarity of the ad, negative evaluations towards the

8 ad and brand can occur (Campbell & Goodstein, 2001; Goodstein, 1993). However, the challenging of ad schemata can also increase interest (Fiske & Taylor, 1991; Goodstein, 1993; Stafford & Stafford, 2002) and positively influence evaluations towards the ad and brand (Goodstein, 1993; Stafford & Stafford, 2002; Taylor, Wilson & Miracle, 1994). Hence, with respect to ad schemata theory, CMC could arguably challenge ad schemata and enhance or decrease ad attitude and brand attitude.

Schemata theory applied within studies of ad-brand (in)congruent advertising suggest that incongruent advertising challenges existing brand schemata. Namely, incongruent

information is unexpected and novel with respect to existing brand knowledge (Heckler & Childers, 1992; Lee & Mason, 1999). Due to the conflict with brand schema, the theory suggests that consumers are motivated to process the information and resolve the ad-brand incongruency (Alden et al., 2000). Subsequently, positive outcomes can occur; the increase in process can result in an enhanced recall (Dahlén et al., 2005; Heckler & Childers, 1992; Phillips, 2000) and the need for resolution of the incongurency and the sense of satisfaction and accomplishment after solving could be transferred to the evaluations of the ad and brand (Dahlén et al., 2005; Peracchio & Meyers-Levy, 1994; Phillips; 2000).

However, when consumers are reluctant towards the puzzle, are incapable of solving the puzzle, or the ad does not match the expectations of the brand schemata of the consumer, the ad can be perceived as irritating by the consumer. An unsolved incongruent puzzle can have a decreasing effect on ad and brand attitude (Phillips, 2000) and recall (Lange & Dahlén, 2003). Hence, with respect to brand schema in comparison with ad schemata, an ad-brand incongruent CMC can arguably likewise result in positive or negative ad and ad-brand attitude, but it can also enhance and decrease recall.

Findings of CA and ad-brand (in)congruent advertising studies are coherent with the diverse assumptions of ad and brand schemata theory. Namely, CA studies have shown that

9 creative advertising positively influences ad and brand attitudes (Ang & Low, 2000; Ang et al., 2007; Smith et al., 2007) and recall (Ang et al., 2007; Baack et al., 2008; Dahlén et al., 2009). Hereby the effect is more pronounced for ad attitude than for brand attitude (Ang & Low, 2000). However, research of Till & Baack (2005) likewise found positive effects on ad and brand recall, but they did not find significant effects on ad and brand attitude. Thus, although not all research indicated significant effects on ad and brand attitude, the significant findings support the positive effects of the theorizations of ad schemata. Hence, it is arguable that ad-brand congruent CMC will positively influence ad and brand attitude and recall, with a possible more pronounced effect on ad attitude.

For the ad-brand (in)congruent studies the findings are also contradicting, while the set-up and experimental stimuli remained consistent. Namely, Lange & Dahlén (2003) found that ad-brand incongruent versus congruent advertising decreases ad recall, but enhances brand recall. Furthermore, they did not find significant effects on ad and brand attitude. While Dahlén et al (2005) indicated that ad-brand incongruency versus congruency decreases ad attitude, but enhances brand attitude. Thus, taken the significant findings into account, ad-brand incongruency decreases ad attitude and ad recall, while it enhances brand attitude and brand recall.

When taking the CA studies with ad schemata theory and the ad-brand incongruent studies with brand schemata theory into account, some significant findings of the CA studies and incongruency studies are notable. Namely, with the ad-brand incongruent advertising it is notable that the findings stress brand related outcomes; it enhances brand attitude and brand recall. This can suggest that the challenging of brand schema by ad-brand incongruency subsequently has a more pronounced effect for brand related outcomes. On the other hand, with the CA studies it is notable that Ang & Low (2000) found a more pronounced effect on ad attitude, which could suggest that the challenging of ad schemata also is especially

10 effective for ad related outcomes. Hence, it is arguable that an ad-brand congruent CMC has a more pronounced effect on ad related outcomes such as ad attitude and ad recall, while an ad-brand incongruent CMC likely has a more pronounced effect on brand related outcomes such as brand attitude and brand recall.

However, the ad-brand incongruent studies also suggest a decreasing effect on ad attitude and ad recall, but it has to be taken into account that an ad-brand incongruent CMC is an aggregation of creative advertising and incongruent advertising. Therefore, it is arguable that ad-brand incongruent CMC benefits from being a form of as well CMC marketing as incongruent advertising. Hence, an ad-brand incongruent CMC will likely have a positive effect on ad attitude and recall due to creativity, but due to decreasing effect of incongruency the effect will be less pronounced than for ad-brand congruent CMC. In sum, the next

hypotheses can be formulated:

H1a: Ad-brand congruent CMC has a more positive influence on attitude towards the ad than ad-brand incongruent CMC

H1b: Ad-brand incongruent CMC has a more positive influence on attitude towards the brand than ad-brand congruent CMC

H1c: Ad-brand congruent CMC has a more positive influence on ad recall than ad-brand incongruent CMC

H1d: Ad-brand incongruent CMC has a more positive influence on brand recall than ad-brand congruent CMC

While schemata theory is mostly relevant for explaining the relationship between ad-brand (in)congruent CMC and ad and brand attitude and recall, the guerrilla effect can subsequently assist relevantly to schemata theory by providing a possible argumentation on how the

11 relationship between ad-brand (in)congruent CM and ad and brand attitude and recall can be established and influenced with respect to WOM. Subsequently, schemata theory can assist the guerrilla effect in explaining possible effects during and after the guerrilla effect with respect to ad and brand attitude and recall.

Namely, the guerrilla effect is explained in terms of that a guerrilla campaign goes beyond the scope of ad familiarity and perceptions and expectations (e.g. arguably

challenging schemas), which will draw attention and will therefore be processed more thoroughly, referred to as the surprise effect. Expanding on the scope of schemata theory, guerrilla marketing suggests that due to this surprise effect diffusion by means of WOM can occur, hereby further enhancing attention towards the ad and processing of the ad (Hutter & Hoffman, 2011; Nufer, 2013).

Hereby it is arguable that the mentioned surprise effect occurs in coherence with schemata theory of the CA studies and ad-brand incongruent studies; the ad is perceived as surprising and novel, hereby challenging ad and brand schema. Therefore, resulting into a possible negative ad and brand attitude due to unfamiliarity of the ad (Campbell &

Goodstein, 2001; Goodstein, 1993) or due to incapability of resolving the incongruence

(Philips, 2000). A positive ad and brand attitude is also likely due to breaking and

challenging the ad mold (Goodstein, 1993; Stafford & Staffor, 2002; Taylor et al., 1994) or due to a sense of satisfaction after solving the incongruency (Dahlén et al., 2005; Peracchio & Meyers-Levy, 1994; Philips, 2000).

Subsequently is it arguable that the effect of surprise – the resulting negative or positive ad and brand attitude – can rub off on the nature of WOM of the sender (e.g. positive or negative responses leading to positive or negative WOM) and subsequently on the

receiver. Furthermore, by receiving positive or negative WOM, an ad is likely to be denoted for the recipient. By denoting the ad by means of WOM a better process of the ad can occur,

12 leading to an enhanced ad and brand recall (Philips, 2000; Rajaram et al., 2001; Rosbergen et al., 1997). Thus, when receiving WOM, positive or negative, it can moderate the recipient own surprise effect and ad process of the ad-brand (in)congruent CMC and its effect on ad and brand attitude and recall, hence the nature of WOM is a possible moderator for ad-brand (in)congruent CMC on ad and brand attitude and recall.

However, there are no empirical studies conducted to research WOM as a moderator and therefore no empirical evidence to support the assumptions mentioned above.

Nonetheless, there are some empirical studies which have found that the nature of WOM can effect ad and brand attitude. Namely, studies have found that negative reviews and negative WOM leads to a negative attitude towards the product (Lee, Park & Han, 2008) and ad and brand (Sundaram &Webster, 1999). Furthermore, the negative attitude can be enhanced by the growing amount of negative reviews (Lee et al., 2008). On the other hand, receiving positive attitudes from peers has a positive influence on the recipient’s ad and brand attitude (Price, Feick & Higie, 1989; Sundaram & Webster, 1999). When the findings and

theorizations of WOM are applied to a possible moderation of the nature of WOM, it is arguable that the nature of WOM will moderate these relationships by further strengthen or weaken them depending on the type of WOM. Hence, a positive WOM will strengthen a positive effect and weaken a decreasing effect, while a negative WOM will strengthen a decreasing effect or weaken a positive effect.

Thus, considering that it is argued that ad-brand congruent CMC will lead to a more pronounced effect on ad attitude and ad recall and an ad-brand incongruent CMC will have a more pronounced effect on brand attitude and brand recall, a positive WOM will further strengthen this positive effect. On the other hand, a negative WOM will decrease a positive ad attitude and ad recall for an ad-brand congruent CMC and will strengthen the decrease for ad-brand incongruent CMC on ad attitude and ad recall, considering the expectation

13 incongruency decreases ad attitude and ad recall. Hence, the next hypotheses can be

formulated:

H2a: A more positive WOM will more enhance the positive influence on attitude towards the ad for ad-brand (in)congruent CMC than a more negative WOM

H2b: A more positive WOM will more enhance attitude towards the brand for ad-brand (in)congruent CMC than a more negative WOM

H2c: A more positive WOM will more enhance the positive influence ad recall for ad-brand (in)congruent CMC than a more negative WOM

H2d: A more positive WOM will more enhance attitude brand recall for ad-brand (in)congruent CMC than a negative WOM

Nature of WOM

more negative versus more positive

Ad-brand congruent CMC

ad and brand attitude and recall Ad-brand incongruent CMC

Figure 1: Conceptual framework

Method

In order to research the influence of ad-brand (in)congruent CMC on ad and brand attitude and recall with the nature of WOM as moderator, an online survey (Qualtrics) experiment

14 with 2 conditions was conducted (ad-brand incongruent CMC ; ad-brand congruent CMC). To generate a higher response, a €15 gift voucher of choice was raffled among the

participants of the main study. By means of 3 pre-tests the experimental stimuli was chosen for the main study. For the pre-tests as well for the main survey the participants were acquired via a convenience sample via Facebook or via personal contact.

The main survey consisted of N=265 participants between 18 – 35 years old. This age group was selected due to that this age group is considered to be most prone to creative expressions. Furthermore, this age group is a group that is considered to have influence and whom are adapters and whom create trends (Dahake, 2013), which is arguably a relevant factor for WOM. 183 Participants completely fulfilled the survey (dropout rate of 82 participants, 30.94%). Of these 183 participants 4 participants were filtered out due to age (>35 years old) and 6 for unfulfilled answers. Thus, 173 participants (n=173) remained suitable for statistical analyses, hence the final sample. The socio-demographics showed that 76.3% was female with an average of 24.7 years old (M=24.73, SD=3.64). The participants mainly had their residence in the Netherlands (75,7%) and their highest education currently enrolled or completed was distributed as followed: High School (3.5%), MBO (2,3%), HBO (11%), Bachelor degree (32.4%), Master’s degree (41%), PhD (7,5%), Professional Degree (1.2%) and other (1,2%).

Pre-tests

In total the results of 3 pre-tests were used to select the experimental stimuli for the main survey. Participants of the pre-tests were asked to name their associations for a medium and a brand without being actual exposed to the advertisement, as in accordance with Dahlén (2005) and Dahlén et al. (2009). The qualitative interpretation of the results of all the pre-tests was as followed; an advertisements was listed as ad-brand incongruent when there was no overlap in association between ad and brand, it was listed as congruent when there was an

15 overlap in associations between ad and brand (Dahlén, 2005; Dahlén et al., 2009). Hereby an overlap was considered congruent when the associations were very similar to completely similar (e.g. bumps and bulges). Due to unconvincing results for ad-brand congruency of the first pre-test derived from four advertisements – only one advertisement had an overlap in 1 association for 3 participants out of 30 – a second (n=26) and third pre-test (N=10) was conducted.

The second and third pre-test showed that the Durex ads as shown in image 3 and 4 were found to be ad-brand congruent (image 3) and ad-brand incongruent (image 4). Namely, for the ad-brand congruent advertisement participants were questioned to name their

associations with blind guide paths and with Durex Pleasuremax. In the first and second pre-test ribs was listed 9 times, knobs 6 times and relief 5 times as associations for the blind guide paths. Of the participants in the third pre-test, 6 out of 10 (N=10) mentioned for Durex Pleasuremax the ribs and 2 out of 10 mentioned knobs and 1 of 10 mentioned relief / pattern. For the ad-brand incongruent advertisement participants were questioned about their associations with Durex and their associations with turnstiles in the first and second pre-test. As expected, there was no overlap in associations and therefore the ad was considered ad-brand incongruent.

16

Image 4: ad-brand incongruent CMC: Durex with posters on turnstiles

Procedure

In the main study the participants were randomly assigned to the two conditions, with 86 participants (49.7%) exposed to the congruent ad and 87 participants (50.3%) to the

incongruent ad. The participants were introduced to the experiment with “On the next page

an advertisement will be shown, please imagine that you have seen this while you were outdoors”. After the advertisement participants were proceeded to the survey with questions

regarding WOM, ad and brand attitude and recall and their socio-demographics.

Observed variables

Dependent variables

Ad attitude was measured on a seven-point semantic differential scale consisting of five

items: “bad/good”, “dislike/like”, “unpleasant/pleasant”, “unfavourable/favourable” and “negative/positive” (Ang & Low, 2000; Dahlén, 2005; Dahlén et al., 2009; Till & Baack, 2005). The principal component analysis (PCA) (varimax) shows that the 5 items form a single uni-dimensional scale: only one component has an eigenvalue above 1 (EV=4.50,

17

R2=.90). Hence, all items load on 1 dimension, with “unfavourable – favourable” as strongest (factor loading is .97). The reliability of the scale is good (α=.97, M=22.11, SD=9.00).

Brand attitude was measured on a seven-point semantic scale with 6 items

“bad/good”, “dislike/like”, “unpleasant/pleasant”, “unfavourable/favourable”,

“negative/positive”, “bad quality/good quality” (Till & Baack, 2005). The PCA (varimax) for brand attitude shows that the 6 items also form a single uni-dimensional scale: only one component has an eigenvalue above 1 (EV=5.10, R2=.85). Hence, all items load on 1

dimension with “dislike – like” as strongest (factor loading is .96). The reliability of the scale is good (α=.96, M=33.16, SD=7.3).

Ad and brand recall was measured aided and unaided. With the unaided ad and brand

recall, participants were asked to recall and name the brand or ad features. For aided brand recall a list was shown with multiple options: Trojan, Sloggy, Victoria Secret, Albert Heijn, Durex or none of all, the participants had to tick the right box. The other brands were selected on the argumentation that they could have been related towards the ads shown. For aided ad recall also a list was given with the correct and possible related ad prompts (Till & Baack, 2005), which consisted of the options blind guide path, walk trough gates, toilet, trains, blind guide dog and none of all.

Moderator

Nature of WOM was measured with a newly created scale on a 7-point Likert-scale

containing the 4 items “Can you indicate to what extent your intention would be to share the

advertisement (for example via social media) with friends?”, “Imagine that you share the advertisements with comment(s) to your (close) friends. Can you indicate the extent you think you would talk positive or negative about the advertisement?”, “Can you indicate the extent you think your (close) friends will share this advertisement (for example via social media) with you?” and “Imagine that your (close) friends share the advertisements with

18

comment(s) to you. Can you indicate the extent you think your (close) friends would talk negative or positive about the advertisement?”.

The PCA (varimax) shows that the 4 items form an uni-dimensional scale, only one component has an eigenvalue above 1 (EV=2.93, R2=.73) with “can you indicate to what extent you think you would talk negative or positive about the advertisement, very negative / very positive” as strongest loading factor (factor loading is .89). The reliability of the scale is good (α=.88; M=16.16, SD=6.43).

Control variables

Ad and brand familiarity. Participants were asked to indicate (yes or no) if they were familiar

with the advertisement and the brand Durex.

Gender. Participants were asked to tick the right option male or female with the

question “What is your sex?”

Age. Participants were asked to fill in their age at the end of the survey.

Education. Participants were asked to indicate the highest level of education currently

enrolled in or completed, with the options primary school, high school or equivalent, MBO, HBO, Bachelor Degree (BS), Master Degree (MS), Doctoral Degree (PhD), Professional Degree (MD, JD, etc.) or other.

Country. Participants were asked in which country their residence was and tick the

correct country from the list.

Surprise. The perception of surprise was measured on a 7-point Likert-scale with

“Please indicate to what extent you agree with the statements. Do you perceive the

advertisement as..” “unsurprising / surprising” (Dahlén, 2005), “unusual / usual - common”,

“unoriginal / innovative”. The item unusual / usual - common was recoded into a positive formulation in order to have the 3 items all in the same directions for analyses. The PCA (varimax) shows that the items form an uni-dimensional scale (EV=2.61, R2=.65), with

19 unsurprising / surprising as highest loaded factor (load factor is .84). The scale had a

reasonable reliability of α=.78 (M=15.39, SD=4.14).

Results

Co-variates check

To control for alternative explanations of the expected differences between experimental groups, the dependent variables were tested for the control variables ad and brand familiarity, gender, age, education, country and surprise. Brand familiarity showed a weak negative association with unaided brand recall (phi=-.34, p <.001), ad familiarity a weak association with aided brand recall (phi=.15, p=.042). The findings indicate that as brand familiarity increases, unaided brand recall tends to 1 (i.e. correct answer) and when ad familiarity increases, aided brand recall also seems to increase.

Another positive correlation was found between gender and ad attitude (F(1.17)=5.98,

p=.015), women (76.30%, M=4.24) were more likely to have a less positive attitude than men

(23.70%, M=5.01).

Age indicated a significant weak negative correlation with ad attitude (rs=-.22,

p=.004), seeming to indicate that when age increases, ad attitude decreases. Furthermore,

there was a significant difference between groups in age for unaided brand recall

(F(16,15)=2.93), p<.001) and aided brand recall (F(16,16)=3.51, p <.001). Further analyses showed weak negative significant correlations, indicating that when age increases, unaided brand recall (rs=-.16, p=.039) and aided brand recall (rs=-.281, p<.001) decreases.

Education tested significant on unaided ad recall (χ2=80.32, p=.003, df=49),

indicating that when the level of education of the participant increases, the amount of correct named ad elements increases.

The groups differed significantly on surprise with a strong correlation for ad attitude (rs=.52, p <.001) and a moderate correlation for brand attitude (rs=.37, p <.001). Thus, as the

20 perception of surprise increases, the attitude towards the ad and brand also seem to increase. For the control variable country it was found that the groups did not significantly differ on the dependent variables.

In sum, the control variables ad and brand familiarity, gender, age, education and the perception of surprise can have an influence on the dependent variables ad and brand attitude and recall. Therefore the mentioned control variables were run with the analyses. However, within the analyses it was found that not all control variables tested significant, indicating that the control variables do not always have an influence on the dependent variable(s). Thus, only the significant control variables of the analyses are reported.

Hypotheses

To analyse the assumptions of H1 and its sub hypotheses, univariate analyses of variance tests (One-way ANCOVA) were performed. With ad and brand attitude, (un)aided brand recall and (un)aided ad recall as dependent variables in each separate analysis and type of CMC (ad-brand incongruent vs ad-brand congruent) as between subjects independent variable. To make the two groups of CMC suitable for analyses, a new variable was created that merged the groups together. Hereby, brand congruent CMC was coded as 1 and ad-brand incongruent CMC coded as 2. Furthermore, within the analyses the control variables ad and brand familiarity, gender, age, education and surprise were taken into account.

For the dependent variable ad attitude, the ANCOVA analysis showed that there is a significant moderate main effect of CMC on ad attitude when controlled for ad familiarity, gender and surprise (F(1.17)=37.26, p<.001, η2=.18). The Bonferroni test indicated that this effect on ad attitude is more pronounced for ad-brand congruent CMC than for ad-brand incongruent CMC (Mdifference=1.36, SD=.22, p<.001, 95% CI [.92, 1.80]) and therefore

confirming H1a. Participants exposed to the ad-brand congruent CMC had the most positive attitude towards the ad (M=5.35, SD=1.35) and the participants exposed to the ad-brand

21 incongruent CMC the least positive attitude towards the ad (M=3.51, SD=1.73). However, it should be noted that the assumption of equal variances in the population has been violated (Levene’s F(1.17)=5.32, p=.022).

For the dependent variable brand attitude no significant effect was found (F(1.17)=2.70, p=.103, η2=.02) and therefore H1b is rejected. Thus for the dependent variables ad and brand attitude, these results show that an ad-brand congruent CMC has a more positive effect on ad attitude than ad-brand incongruent CMC, indicating that

incongruency, as in coherence with the argumentations, seems to decrease the positive effects of CMC on attitude towards the ad. However, the different types of CMC do not effect and significantly differ from each other on brand attitude.

For the dependent variables of (un)aided ad and brand recall of H1c and H1d, the analyses indicated no significant effect for unaided brand recall (F(1.16)=2.29, p=.133, η2

=.01), aided brand recall (F(1.16)=.86, p=.355, η2=.01) and unaided ad recall

(F(1.16)=1.91), p=.169, η2=.01). However, a moderate main effect was found for aided ad recall (F(1.17)=25.63, p<.001, η2=.13), while the control variables tested non-significant. The Bonferroni test showed that ad-brand incongruent CMC has a more pronounced effect on aided ad recall (Mdifferendce=.35, SD=.068, p<.001, 95% CI [.21, .48]) than ad-brand congruent

CMC. Considering that aided ad recall was measured on recognizing the used medium in the ad (e.g. blind guide path and walk trough gates), this finding indicates that the participants of the ad-brand incongruent CMC better recall the used medium of the ad (M=.90, SD=.05) than the participants exposed to an ad-brand congruent CMC (M=.55, SD=.05) when aided.

Although yet again it has to be mentioned that the assumption of equal variances in the population has been violated (Levene’s F(1.17)=118.97, p<.001). Thus, H1d is rejected and H1c is partially not coherent with the expectations, considering it was expected that ad-brand incongruent will decrease the positive effects of CMC on ad recall (Dahlén et al., 2005)

22 instead of enhance ad recall. However, it has to be taken into consideration that this only applies for aided ad recall.

To answer H2a and H2c, univariate analyses of variance tests (Two-way ANCOVA) were run. Considering only a significant main effect was found for ad attitude and aided ad recall only H2a and part of H2c were analysed, hence the analyses were run with as

dependent variables ad attitude and aided ad recall. The independent variables within the test were the type of CMC (ad-brand congruent vs ad-brand incongruent) and WOM (ranging from very positive – very negative) as between subjects independent variables in order to create an interaction effect. The variables ad and brand familiarity, gender, age, education and surprise were taken into account within the analyses.

The analyses show no interaction effect between CMC and WOM on ad attitude (F(15.13)=.33, p=.991, η2=.04) and on aided ad recall (F(15.13)=.95, p=.509, η2=.10). However, the analysis on ad attitude indicates that WOM instead of acting as a moderator, it rather acts as a predictor; WOM has a strong main effect on ad attitude (F(24.13)=7.40,

p<.001, η2=.59), when controlled for surprise.

To analyse the unexpected finding more thoroughly, a linear regression analyse was run to examine the direction of the relation. The regression analysis shows that the more positive WOM, the more the positive the attitude towards the ad is (b*=.77, t =14.90,

p=<.001), when controlled for ad and brand familiarity, gender, age, education and surprise.

Taken everything into account, hypotheses 2a, 2b, 2c and 2d are rejected, WOM does not seem to moderate ad-brand (in)congruent CMC on ad and brand attitude and recall, but interestingly acts as a predictor for ad attitude.

Additional analyses were performed to closer examine the main effect of WOM as a predictor for ad attitude. Namely, this relation possibly indicates that the nature of WOM instead of a moderator, acts like a mediator between ad-brand (in)congruent CMC and ad

23 attitude. Therefore, a closer inspection was given to WOM and its effects by means of linear regression analyses as in coherence with the three steps of Baron & Kenny (1986). The analyses were executed with the control variables ad and brand familiarity, gender, age, education and surprise.

Firstly, the linear regression analysis with the type of CMC (ad-brand congruent versus ad-brand incongruent) as independent variable and attitude towards the ad as dependent variable indicated a significant effect (F(7.17)=18.51, p<.001). Thus, the

regression model can be used to predict attitude towards the ad with 66% of the proportion in variance explained of CMC (R2=.66, b*=-1.36, t=-6.10, p<.001, 95% CI [-1.80, -.92). The model indicates that the ad-brand incongruent CMC on average predicts lower scores on ad attitudes than congruent CMC, when controlled for ad and brand familiarity, gender, age, education and surprise.

The second step of Baron & Kenny (1986) was executed with as independent variable the type of CMC (congruent versus incongruent) and WOM as dependent variable. A

significant effect of CMC on WOM was found (F(7.17)=17.28, p<.001). The regression model can be used to predict WOM with 65% of the proportion in variance explained (R2=.65, b*=-1.05, t=-5.22, p<.001, 95% CI [-1.45, -.66]). Thus, the model indicates that ad-brand incongruent CMC significantly predicts a more negative WOM than ad-ad-brand

congruent CMC, when controlled for ad and brand familiarity, gender, age, education and surprise. This finding seem to indicate a logical mediation effect; when the ad-brand incongruent CMC generates a more negative WOM, it will also result in a less positive attitude.

However, the third linear regression analysis with CMC and WOM as independent and ad attitude as dependent, indicated a partial mediation instead of a full mediation.

24 ad attitude with 85% of the proportion in variance explained (R2=.85). Furthermore, the model indicates that the type of CMC is within the model significant on ad attitude (b*=-.53,

t=-3.14, p=.002) as well for WOM on ad attitude (b*=.79, t=12.99, p<.001). Thus,

considering the effect of CMC is still significant, WOM does not fully mediates the relationship, but partially. This indicates that CMC partially has a direct influence on ad attitude, but also partially has an indirect influence on ad attitude due to mediation of WOM. Hereby the beta’s indicate that the incongruent CMC predicts a lower ad attitude and that when there is a more positive WOM, ad attitude becomes less negative.

Furthermore, an additional analysis was run to examine the influence of ad-brand (in)congruent CMC on the control variable surprise by means of a one-way ANCOVA. With the type of CMC as independent and as dependent variable surprise, the analysis indicated a weak main effect (F(1.17)=20.10, p<.001, η2=.11), when controlled for age. The Bonferonni test indicated that the ad-brand congruent ad has the most pronounced effect on surprise (Mdifference=.88, SD=.20, p<.001, 95% CI [.49, 1.26). The ad-brand congruent ad is perceived

as most surprising (M=5.61, SD=1.11) and the ad-brand incongruent CMC as the least surprising (M=4.66, SD=1.52). Hereby the Levene’s F-test was not violated (F(1.17)=1.20,

p=.275), hence there are equal variances. Thus, an ad-brand congruent CMC has a more

positive influence on the perception of surprise than ad-brand incongruent CMC, indicating that an ad-brand congruent ad is more perceived as a novel, surprising and unusual than an ad-brand incongruent CMC.

To test if surprise plays a more key role than expected, surprise was tested for moderation on ad attitude and aided ad recall with a two way ANCOVA. The analyses were run with as dependent variables ad attitude and aided ad recall and as independent variables the type of CMC (congruent versus incongruent) and the perception of surprise (very

25 gender, age and education were also taken into account.

The analyses did not show a significant interaction effect on ad attitude

(F(13.14)=.59, p=.858, η2=.05) and aided ad recall (F(13.14)=.44, p=.954, η2=.04). However, the ANCOVA indicated that surprise seems to act as a predictor on ad attitude

(F(18.14)=2.08, p=.010, η2=.22), when controlled for ad familiarity and gender.

An additional linear regression was run to examine the direction of the relationship between surprise and ad attitude. The analysis shows that when the perception of surprise increases (e.g. more surprising) it significantly predicts a higher attitude towards the ad (R2=.56, b*=.48, t=7.21, p<.001), with 56% of the proportion in variance explained. In sum, ad-brand congruent CMC enhances the ad attitude and the perception of surprise more than ad-brand incongruent CMC, while ad-brand incongruent CMC enhances on aided ad recall more than ad-brand congruent CMC. Furthermore, the analyses indicate that when WOM is more negative, the attitude is less positive and when the ad is seen as more surprising, the ad attitude increases. An important note that has to be taken into

consideration that the effects found were mostly weak and moderate and sometimes violated the assumption of equal variances, therefore decreasing the validity and reliability of the findings, hence the findings are most likely not generalizable to the research population or other ad-brand (in)congruent CMC’s.

Conclusion and discussion

The aim of this study was to research the extent of the effect of ad-brand (in)congruent creative media choice (CMC) on ad and brand attitude and recall and if this relation was moderated by the nature of WOM. Whilst there were theories and findings present within the academic field, the studies of creative advertising (CA)(e.g. creative media choice, creative marketing and guerrilla marketing) and incongruent advertising related to the research question and subject were contradicting and not yet intertwined. Furthermore, while the

26 literature on guerrilla marketing offered a theoretical insight with respect to WOM, WOM and subsequently the nature of WOM was never empirically researched as a moderator. Therefore, an online experimental survey was conducted to give insight into the differences in effects between ad-brand congruent CMC and ad-brand incongruent CMC on ad and brand attitude and recall and the moderation effects of the extent of nature of WOM – ranging from more negative to more positive – on these dependent variables.

In coherence with the findings of past research, the study found a positive effect of ad-brand (in)congruent CMC on ad attitude (Ang & Low, 2000; Ang et al., 2007; Smith et al., 2007). More specifically, an ad-brand congruent CMC had a more pronounced effect on ad attitude than ad-brand incongruent CMC, which is in coherence with H1a and with the findings of Dahlén et al. (2005), whom indicated that incongruency decreases ad attitude compared to congruency.

In contrast with past research and the expectations of H1b, no significant effect was found for brand attitude (Ang & Low, 2000; Ang et al., 2007; Dahlén et al., 2005; Smith et al., 2007). A possible explanation for this unexpected finding is the strong position of Durex in the market which likely results in a positive attitude towards the brand, regardless of exposure to ads. The means of both groups seem to confirm this, there seems to be a ceiling-effect of brand attitude for as well the congruent group (M=5.38, SD=.13) as the incongruent group (M=5.68, SD=.13). Considering past research did indicate significant effects on brand attitude for as well creative advertising and ad-brand incongruent advertising, it raises the question if Durex was a suitable choice for the comparison between ad-brand incongruent CMC and ad-brand congruent CMC. Therefore, a suggestion for future research could be to compare strong brands and weak brands for incongruency and congruency in order to research if the choice of a strong brand could have influenced the findings.

27 significant effect was found for (un)aided brand recall and unaided ad recall (Ang et al., 2007; Lange & Dahlén, 2003; Till & Baack, 2005). However, a significant main effect was found for aided ad recall, but in the opposite direction of the assumptions of the hypothesis 1C (Lange & Dahlén, 2003). Namely, this study found that ad-brand incongruent CMC has a more pronounced effect on aided ad recall than ad-brand congruent CMC, instead of an expected decrease in effect.

A possible explanation for the non significant findings for (un)aided brand recall is yet again the ceiling effect. The means of unaided brand recall seem to indicate that for as well the congruent group (M=.95, SD=.21) as for the incongruent group (M=.90, SD=.30) the participants recalled the brand mostly correctly. For aided brand recall the ceiling effect seems even stronger for as well the congruent group (M=1.00, SD=.00) and incongruent group (M=.98, SD=.15). This could be due to the strong correlation found between ad and brand familiarity and (un)aided brand recall; familiarity with ad and brand suggests existing schemas. Existing schemas are arguably easily retrievable and active, hence the brand is already prominent and stored within the schemas, so the ad does not enhance or has an influence on (un)aided brand recall.

By means of a newly created scale measuring the nature of WOM, a possible

moderator was empirically tested for ad-brand (in)congruent CMC and ad and brand attitude and recall. The study showed that WOM does not have a function as a moderator on ad and brand attitude and recall. However, the analyses did indicate that the nature of WOM rather acts as a partial mediation between ad-brand (in)congruent CMC and ad attitude. Namely, it was found that ad-brand incongruent CMC significantly predicts a more negative WOM and subsequently an enhanced negative attitude than ad-brand congruent CMC. Therefore, the finding does offer partial empirical evidence for the theorizations of guerrilla marketing, which hypothesized WOM acts as a mediator (Hutter & Hoffman, 2011; Nufer, 2013).

28 Within the literature of creative media choices (Dahlén, 2005; Dahlén et al., 2009), guerrilla marketing (Hutter & Hoffman, 2011; Nufer, 2013) and creative advertising (Ang & Low, 2000; Ang et al., 2007; Till & Baack, 2005), surprise and novelty are pronounced within the definitions and theorizations. Within the current research the perception of surprise likewise was pronounced within the analyses. Namely, it was found that surprise is

influenced by ad-brand (in)congruent CMC, but can also act as a predictor for ad attitude. More specifically, an ad-brand congruent CMC was considered to be most surprising, while an ad-brand incongruent CMC as the least surprising. Furthermore, as a predictor it was found when the CMC was seen as more surprising, the ad attitude increased.

However, the finding that an ad-brand congruent CMC is more surprising is an

unexpected finding. Namely, it implies that an ad-brand congruent CMC is perceived as more novel, surprising and unusual than ad-brand incongruent CMC, while it is arguable that the opposite is more logical. Hereby it raises a critical question if the perception of surprise can be attributed to the (in)congruency or due to the choice in advertisements, considering the incongruent ad made use of posters of female bums. It is arguable that the use of women in underwear is a common form of advertising and hence less surprising.

A strength of this research is the more diverse research group in comparison with past research. Namely, past research mainly focused on (undergraduate) students (Ang & Low, 2000; Ang et al., 2007; Dahlén, 2005; Dahlén et al., 2009; Smith et al., 2007; Till & Baack, 2005), while this research took age, education and country of residence into account. The co-variates check also indicated that the socio-demographics age, gender and education

significantly differed between experimental groups. More specifically it was found that women and older participants were less likely to have a positive attitude towards the ad. This could indicate that women and older people are less sensitive towards CMC and subsequently (in)congruency. Future research should examine more closely what the effect is of gender and

29 age for ad-brand (in)congruent CMC in order to provide a more thorough understanding of an effective creative advertisement and a suitable target audience.

Furthermore, a strength and a weakness at the same time is the choice of existing advertisements. Namely, it can be considered a strength due to an examination of actual advertisements instead of manipulated advertisements, this offers an empirical insight to a direct case. However, it can also be considered a weakness due to that the comparison between ad-brand congruent CMC and ad-brand incongruent CMC is not analysed as

constant as possible. When the ads are kept constant with a slight differentiation, the findings are more likely to be considered more valid and reliable.

A critical note of this research is that the significant findings of the analyses were often weak or moderate and often violated the assumption of equal variances. Therefore, the validity and reliability of this research is decreased and it is unlikely that the findings can be generalized to the research population (18 – 35 years old) and to other ad-brand (in)congruent CMC’s.

Concluding, this research offers new insights related to the field of creative

advertising. Namely, by intertwining ad-brand incongruency and the different types of CA, the research offers a less narrow definition of creative media choice and therefore a novel understanding of ad-brand (in)congruent CMC. By showing that an ad-brand (in)congruent CMC can have a positive influence on ad attitude and aided ad recall, it offers insights for the practitioners to get one step closer into an effective creative advertisement.

30 Reference list

Ang, S. H., & Low, S. Y. M. (2000). Exploring the dimensions of ad creativity. Psychology

and Marketing, 17(10), 835-845.

Ang, S. H., & Lee, Y. H. (2007). The ad creativity tube: conceptualization and initial validation. Journal of the Academy of Marketing Science, 35, 220-232. doi: 10.1007/s11747-007-0042-4

Alden, D. L., Mukherjee, A., & Hoyer, W. D. (2000). The effects of incongruity, surprise and positive moderators on perceived humor in television advertising. Journal of

Advertising, 29(2), 1-15.

Baack, D. W., Wilson, R. T., & Till, B. D. (2008). Creativity and memory effects recall, recognition, and an exploration of nontraditional media. Journal of Advertising,

37(4), 85-94.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of

Personality and Social Psychology, 51, 1173-1182.

Campbell, M. C., & Goodstein, R. C. (2001). The moderating effect of perceived risk on consumers’ evaluation of product incongruity: Preference for the norm. Journal of

Consumer Research, 28, 439-449.

Dahake, P. S. (2013). Guerrilla marketing – an untraditional methods to place a brand in young consumers’ perception. International Journal on Marketing Management, 6(2), 37-45.

Dahlén, M. (2005). The medium as a contextual cue: Effects of creative media choice.

Journal of Advertising, 34(3), 89-98.

Dahlén, M., Friberg, L. & Nilsson, E. (2009). Long live creative media choice. Journal of

31 Dahlén, M., Lange, F., Sjödin, H., & Törn, F. (2005). Effects of ad-brand incongruency.

Journal of Current issues and Research in Advertising, 27(2), 1-12. doi:

10.1080/10641734.2002.10505178

Edwards, S. M., Li, H., & Lee, J. H. (2002). Forces exposure and psychological reactance: Antecedents and consequences of perceived intrusiveness of pop-up ads. Journal of

Advertising, 31, 83-95.

Elliot, M. T., & Speck, P. S. (1998). Consumer perceptions of advertising clutter and its impact across various media. Journal of Advertising Research, 38(1), 29-41. Ferguson, R. (2008). Word of mouth and viral marketing: Taking the temperature of the

hottest trend in marketing. Journal of Consumer Marketing, 25(3), 179-182. Fiske, S. T., & Taylor, S. E. (1991). Social Cognition. New York: McGraw-Hill. Goodstein, R. C. (1993). Category-based applications and extensions in advertising:

motivating more extensive ad processing. Journal of Consumer research, 20, 87-99. Ha, L., & McCann, K. (2008). An integrated model of advertising clutter in offline and online

media. International Journal of Advertising, 27(4), 569-592. doi: 10.2501/S0265048708080153

Heckler, S. E., & Childers, T. L. (1992). The role of the expectancy and relevancy in memory for verbal and visual information: What is incongruency? Journal of Consumer

Research, 18(3), 475-492.

Hoch, S. J. (2002). Product experience is seductive. Journal of Consumer Research, 29(12), 448-454.

Hutter, K. & Hoffman, S. (2011). Guerilla marketing: The nature of the concepts and propositions of future research. Asian Journal of Marketing, 1-16.

Lange, F., & Dahlén, M. (2003). Let’s be strange: brand familiarity and ad-brand incongruency. Journal of Product & Brand Management, 12(7), 449-461. doi:

32 10.1108/10610420310506010

Lee, J., Park, D., & Han, I. (2008). The effect of negative online consumer reviews on product attitude: An information processing view. Electronic Commerce Research

and Applications, 7, 341-352.

Lee, Y. H., & Mason, C. (1999). Responses to information incongruency in advertising: The role of expectancy, relevancy and humor. Journal of Consumer Research, 26(9), 156-169.

Nufer, G. (2013). Guerrilla marketing – Innovative or parasitic marketing? Modern

Economy, 4, 1-6.

Peracchio, L. A., & Meyers-Levy, J. (1994). How ambiguous cropped objects in ad photos can affect product evaluations. Journal of Consumer Research, 21(6), 190-204. Pieters, R., Warlop L., & Wedel, M. (2002). Breaking through the clutter: Benefits of

advertisement originality and familiarity for brand attention and memory.

Management Science, 48(6), 765-781.

Phillips, B. J. (2000). The impact of verbal anchoring on consumer response to image ads.

Journal of Advertising, 29(1), 15-24.

Price, L. L., Feick, L. F., & Higie, R. A. (1989). Preference heterogeneity and coorientation as determinants of perceived informational influence. Journal of Business Research,

19(3), 227-242.

Rajaram, S., Srinivas, K., & Travers, S. (2001). The effects of attention on perceptual implicit memory. Memory and Cognition, 29(7), 920-930.

Rosbergen, E., Pieters, R., & Wedel, M. (1997). Visual attention to advertising: A segment-level analysis. Journal of Consumer Research, 24(3), 305-314.

Smith, R. E., & MacKenzie, S. B., Yang, X., Buchholz, L. M., & Darley, W. K. (2007). Modeling the determinants and effects of creativity in advertising. Marketing Science,

33

26(6), 819-837.

Smith, R. E., & Yang, X. (2004). Toward a general theory of creativity in advertising: Examining the role of divergence. Marketing Theory, 4(1), 31-58.

Stafford, T. F., & Stafford, M. R. (2002). The advantages of atypical advertisements for stereotypical product categories. Journal of Current Issues and Research in

Advertising, 24(1), 25-37.

Sundaram, D. S., & Webster, S. (1999). The role of brand familiarity on the impact of Word-Of-Mouth communication on brand evaluations. Advances in Consumer Research

(26), 664-667.

Taylor, C. R., Wilson, R. D., & Miracle, G. E. (1994). The impact of brand differentiating messages on the effectiveness of Korean advertising. Journal of International

Marketing, 4(2), 31-52.

Van den Putte, B. (2009). What matters most in advertising campaigns: The relative effect of media expenditure and message content strategy. International Journal of

Advertising, 28, 669-690.

White, A., & Smith, B. L. (2001). Assessing advertising creativity using the creative product -semantic scale. Journal of Advertising Research, 41(6), 27-34.