Blend in or Stand out?
The moderation effects of congruity on internet banner advertisements
Blend in or Stand out?
The moderation effects of congruity on internet banner advertisements
Yasen Vladimirov Grancharov
Faculty of Economics and Business, Groningen University, Jozef Israelsstraat 17, 9718 GB Groningen, The Netherlands, University of Groningen, The Netherlands
s1734121 y.v.grancharov@hotmail.com +31 06 16176805 Date of completion 01.07.2011
Size: 5780 words
“I would like to thank Dr. Debra Trampe from the University of Groningen for her support, constructive criticism and insightful remarks during the last couple of months of my work. Also, special thanks to Yordan Terziev, a Master Student in Cognitive Sciences of New Bulgarian University for the invaluable help and cooperation during the execution of the experiment.”
Supervisor - Prof. Dr. Bob M. Fennis
Department of Marketing at Groningen University
Abstract
This paper is concerned with the ways both implicit and explicit attitudes toward the brand are affected within an online advertising environment. The author tries to unify existing theories from other media and apply them within the web-space. Partial success is found as the initially hypothesized positive relationship between ads’ intrusiveness and both types of attitudes held only for the explicit and not the implicit ones. The pivotal pair of hypothesis within the paper argued for congruence’s moderating properties over the relationship between ads’ intrusiveness and implicit and explicit attitudes toward the brand. Although literature from traditional media supported the existence of such phenomenon within the World Wide Web, the evidence from the experiment the author carried out found no proof for the proposed causal relationships.
Table of Contents:
Chapter 1: Introduction p. 4
Chapter 2: Literature Review p. 5
Chapter 3: Method p. 8 3.1 Experiment p. 8 3.2 Overview of Participants p. 9 3.3 Procedure p. 9 3.4 Independent Variable p. 10 3.5 Control Variable p. 11 3.6 Dependent Variable p. 11 Chapter 4: Results p. 12
Chapter 5: Discussion and Conclusions p. 14
Chapter 6: References p. 17
Chapter 7: Appendices p. 20
7.1 Appendix “A” p. 20
7.2 Appendix “B” p. 23
Blend in or Stand out?
The moderation effects of congruity on internet banner advertisements
Introduction
Advertising is one of the most widely spread marketing techniques throughout the world. It’s known for reaching customers and exposing them to a given idea, product or service, utilizing every single path available – the latest of which is the World Wide Web (WWW). In fact it is becoming so popular that numerous companies are redesigning their strategies in order to make full use of the opportunities the online space presents business nowadays. With the number of online ads increasing progressively ever since the introduction of the web-space, however, assessing their effectiveness becomes tricky. One of the major reasons for that is that most of the advertisements are not even noticed (Grunert, 1996). With that, researchers in the field of marketing as well as those in consumer behavior and psychology identified numerous topics in need of thorough experimentation and exploration.
Banner ads were the first form of online advertising known since October 1994 when AT&T placed an ad on HotWired.com (Zeff and Robinson, 1999). Nowadays banners have established themselves as the predominant and
most widely spread form of internet advertising (Hofacker and Murphy, 1998). The most recent figures coming from The Internet Advertising Bureau [IAB] (2009) has shown an increase in advertising revenue from 16.8 billion in 2006 to 22.7 billion in 2009. Banner ads accounted for 35 percent of those revenues, which is a 2 percent increase from 2008. Those numbers on the other hand can serve a different purpose for a researcher – they are a significant proof of the ads clutter that the internet space has turned into. With that, other problems are identified – web surfers have become very experienced (Dahlen, 2001), paying less and less attention to ads placed on web-sites (Yoo, 2005). Their low recall ability (Fang, Singh and Ahluwalia, 2007) is a hard enough evidence for that phenomenon.
through a web-site ad, despite the huge competition, becomes the key towards success. For that reason, this paper will delve into the topic of online banner ad diversity and try to explore the type of effects ads have on users. Two important characteristics of online ads - intrusiveness and congruence – will be looked into in detail. Intrusiveness represents the extent to which a banner ad disrupts the process of browsing for a given consumer in an attempt to capture his attention. Simple banner ads do not interrupt browsing, thus they represent low levels of intrusiveness. Pop-up ads that cannot be closed and require the attention of the user for a short period of time before allowing him to continue browsing are examples of high intrusiveness ads. The second characteristic of central importance for this paper – congruence – can be defined as the presence of a thematic fit between the ad and the web-site it is placed on. That being said, this scientific paper will focus on the topic of website-ad congruence and whether it possesses moderating characteristics over the relationship between the level of intrusiveness and both implicit and explicit attitudes. The study will incorporate intrusiveness and congruence as causal variables, while implicit and explicit attitudes toward the brand will be the variables affected by them.
Literature Review
through rates. A few years after the publishing of that study however, Gong and Maddox (2003) proved that click through rates is a significant predictor for banner recall only, while having no effect on brand recall, purchase intentions and attitudes toward the brand. Thus, perception rates and click through rates are not reliable and trustworthy measures of advertising effectiveness. Therefore, this study will try to address the existing gap within the literature and provide a more accurate measurement, namely – implicit and explicit attitudes toward the brand.
Chang, Zhang and Luo (2010) state that "there are many ways to measure advertising effect, such as through awareness (Hansen, Olsen, and Lundsteen, 2006), brand choice (Cobb-Walgren, Ruble, and Donthu 1995), purchase likelihood (Aaker, Stayman, and Hagerty 1986), viewing time (Olney, Holbrook, and Batra 1991), brand perceptions (MacKenzie and Lutz 1989), purchase intentions (Bellman, Schweda, and Varan 2009; Gotlieb and Sarel 1991; Kamins 1990), memory-based tests (Brennan, Dubas, and Babin 1999; Russell 1998), and so on." However, one would hardly engage in generating purchasing intentions, brand intentions, etc. if one has not been successfully persuaded to do so. Thus, the key to an efficient ad in general is its ability to create and enhance positive attitudes. Several well renowned theories have been established over the past
fifty years that provide evidence for the effects of various exposure effects - the message learning model (Hovland, Janis and Kelly, 1953), the mere exposure effect (Zajonc, 1968), Berlyne’s two-factor theory ((Berlyne, 1970), which was later elaborated by Stang (1975)) and the Elaboration Likelihood Model (Cacciopo and Petty, 1985). All of them make it clear that exposure to a given stimulus has the power to alter one’s attitude towards it, where the higher the amount of exposures, the higher the chances are that an individual will have his attitudes towards the presented stimulus affected. That being said, it stands to reason to suggest that there is an existing causal relationship between intrusiveness and attitudes toward the brand, which is strengthened with the increase of the intrusiveness level. Thus, the following hypotheses are formed:
H1: The higher the level of online ads’ intrusiveness, the more positive the implicit attitudes toward the brand depicted in the ad.
H2: The higher the level of online ads’ intrusiveness, the more positive the explicit attitudes toward the brand depicted in the ad.
recognition through application within television advertising, where the viewers’ ability of recall and recognition was tested (Murphy, Cunningham and Wilcox, 1979). For that purpose two types of ads – humorous and non-humorous - were presented to observers in differentiated contextual settings – comedy, action/adventure and documentary. The results showed that humorous ads appearing in action/adventure and documentary channels had higher unaided recall levels then situations of placement within comedy channels. Contrasting results were obtained when non-humorous ads were shown in the same three different settings – viewers’ unaided recall was higher for the comedy channel setting, surpassing the results from both the action/adventure and documentary channel. In other words, lack of thematic fit between the ad and the environment leads to results outperforming those from conditions of congruence. Moorman, Neijens and Smit (2002), however, argue in a completely opposite direction. They prove that a congruence scenario representing a fit between the ad and the environment it is placed in is going more effective than scenarios of misfit. They base their findings on the theory of cognitive priming (Yi, 1990b; Zimbardo and Leippe, 1991), according to which a customer is prone to be affected by ads that are congruent to their environment. That is because the core subject
there is a relationship between congruence on the one hand and both implicit and explicit attitudes toward the brand on the other, meaning that exposure effects are in turn altered based on the presence or the lack of congruence. More specifically, in cases of thematic congruence between the ad and its environment, a customer is expected to unconsciously process the advertisements based on the cognitive priming effect, which would result in an amplification of the intrusiveness effect on the implicit attitudes toward the brand. For scenarios of incongruence between the ad and its environment, a customer is expected to consciously process the advertisements based on the effects of thematic misfit, which would result in an amplification of the intrusiveness effect on the explicit attitudes toward the brand. Thus, the following hypotheses are formed:
H3: For scenarios of congruence between the ad and its environment, the intrusiveness’ effects on the implicit attitudes toward the brand will result in higher, more positive attitudes than those for scenarios of incongruence.
H4: For scenarios of incongruence between the ad and its environment, the intrusiveness’ effects on the explicit attitudes toward the brand will result in higher, more positive
attitudes than those for scenarios of congruence.
Having analyzed some of the relevant existing literature and created the hypotheses that this paper will attempt to provide an empirical proof for, the following conceptual model can be created, summarizing the argued relationships of causality between the specified variables:
In short, this academic work will try to combine existing theories for traditional media and provide a unified one that would not only apply for the newest of arenas on which marketers are expected to compete on, but would also incorporate a more accurate measure of testing advertising effectiveness – explicit and implicit attitudes toward the brand.
Method
Experiment
for existing moderating effects of internet site – ad congruity on the relationship between ad intrusiveness and implicit and explicit attitudes toward the brand. The hypothesized effect of congruity on the implicit attitudes is based on the theory of cognitive priming (Yi, 1990; Zimbardo and Leipe, 1991), according to which advertisements are easily processed in congruent environments because the subject and theme of the ad is already in the mind of the person viewing it, thus enhancing positive attitudes towards the ad and the brand (Moorman, Neijens and Smit, 2002).The positive effects of incongruity on explicit attitudes on the other hand are based on the established opinion within the literature that incongruity causes increased level of recall within people viewing an ad in such an environment (Murphy, Cunningham and Wilcox, 1979). The effects of the increasing levels of intrusiveness are based on the theory of mere exposure, according to which direct positive effects on both implicit and explicit attitudes are to be observed. The experiment used the work of Hafner and Trampe (2009) as a foundational basis and replicated their way of testing the dependent variables.
Overview and Participants
Seventy eight undergraduate students from the University of Groningen and the New Bulgarian University majoring in different
subjects (Male = 39, Female = 39, Mean age = 20.65, SD = .89) were used as participants for the experiment. For the ones studying at the New Bulgarian University special research credits were given in return for participation. A twenty euro Media Markt voucher for music, movies and games was the price students at the University of Groningen could win if they participated in the experiment. Participants at the New Bulgarian University and the University of Groningen were randomly assigned to a condition in a 2 (Intrusiveness: Low versus High) x 2 (Congruence: Congruence versus Incongruence) between subjects factorial design measuring the effects on both Implicit and Explicit attitudes. The level of intrusiveness varied between participants; congruence varied between participants also.
Procedure
The procedure applied for participants at the University of Groningen and at the New Bulgarian University was identical. Upon arrival, participants knew they would participate in a study that would earn them a special research credit (for students in Bulgaria) or a chance to win a twenty euro voucher (for students in the Netherlands) upon completion of a fifteen minute experiment.
attitudes toward the brand. For measuring the implicit attitudes toward the brand, participants were presented with one of four possible version of the same web-site, created specifically for this experiment. Each version represented a different possible scenario which each participant was randomly assigned to upon agreeing to participate in the experiment. The web-site required participants to read the information spread over three sub-pages under five minutes. While reading, depending on the scenario they were randomly assigned to, participants would be exposed to ads of high or low intrusiveness, the subject of which would either match the one of the web-site contents (congruence scenario) or not (incongruence scenario). Once five minutes have passed, the web-sites were blocked and participants had to turn their attention to a special software program - Authorware v7.0. Thanks to it an affective priming task was performed that tested implicit attitudes toward the brand which was depicted in the ads each participant encountered while reading the web-sites. Since telling what exactly the procedure was testing for would switch participants’ attention solely to the ads and eventually provide unreliable results, participants were intentionally misled by the supervisor. They were told that they are participating in an experiment that would present each of them with a website that is utilizing differing writing style patterns aimed at
increasing level of perception and that the task after it (the affective priming test) would test for whether participants’ perception has been affected or not.
Upon completion, participants were asked to provide explicit evaluations of information they have encountered on the website prototype. Those evaluations tested the explicit attitudes toward the brand that was depicted on the ads of the web-site each participant was exposed to. Furthermore, they were asked to provide some demographic data for the purposes of the study. Once done, a last question was asked – whether they found there was a match between the theme of the website content and the one of the ads placed on it, which each participant answered correctly with regard to the scenario they were in. After all the participants were finished with the tasks presented to them, they were all debriefed and dismissed.
Independent Variable
had to evaluate them. For the purposes of my study, the independent variable was defined as “level of intrusiveness”, where throughout experiencing exposure to the ad over three times, the participant was presented with either a low intrusiveness ad or a high intrusiveness ads. For low intrusiveness ads I used simple banner ads, while for high intrusiveness ads a pop-up no-skip duration ad was utilized, which had to be viewed for 20 seconds before allowing the participant to reach the destination of his choice. The brand “Puma” was used for all the ads alongside the following slogan: “Puma Energy = (More + Cool) ^ 2” for both the banner ad and the pop-up no-skip ad (for screenshots of the ads, please refer to Appendix C).
Control Variable
The control variable – “Congruence” – was utilized by presenting the participant with a prototype website with a congruent mix between context and ads. As already said in the previous section, all ads shared the same subject – football shoes. The control variable had two possible dimensions – congruence and incongruence. Also, while the independent variable dealt with the ads, the control variable represented the environment in which the ads were presented. Thus, based on the two values of the control variable, only two scenarios were possible. For the congruence scenario – the football ads were presented to the participant in
an environment matching the subject of the ad. Within the experiment the congruent scenario represented Puma ads displayed alongside an article about the advancements in football shoes technology by both Nike and Adidas. For the incongruence setting – the football ads were presented to the participant in an environment that did not match the subject of the ad. More specifically, in the experiment, that scenario represented Puma ads appearing along with an article informing about the properties of 3D technologies.
Dependent Variable
screen as quickly as they can, based on its valence. The second was coded to appear immediately after the prime and stay there until it has been categorized as either positive (by clicking the ‘’Z’’ button on the keyboard) or negative (by clicking the ‘’/’’ button on the keyboard). Each categorization was followed by an interval of one second and then the next prime would appear, followed by a word in need of categorization. In total, participants viewed 40 prime words, 30 of which were neutral (e.g. chair, plate, belt, etc.) and 10 were simple brand name appearances – “Puma”. Each prime appeared twice – once followed by a randomly chosen positive word from a list of 10 (e.g. sun, love, happiness, freedom, etc.) and once followed by a randomly chosen negative word from a list of 10 (e.g. storm, death, mud, slavery, etc.). Thus the complete affective priming task consisted of 80 trials in total. Each trial was characterized by a specific time latency based on the time taken by a participant to categorize each word. Once the response latencies were collected from the participants of the experiment, they were recomputed for the purpose of analyzing them easier. More specifically, the response latencies for brand prime appearances were singled out from the rest and were split into two groups. The first represented a total of the time taken by participants to classify the band primes followed by a random positive word, while the second
represented a total of the time needed for classification of the brand primes followed by a random negative word. With that a final calculation was made – time taken per person to classify brand primes followed by positive words was subtracted from the times of brand primes followed by a negative words. Resulting values above zero represented positive implicit attitudes toward the brand, while values lower than zero indicated negative implicit attitudes toward the brand. Thus, the higher the resulting value, the more positive the participant’s implicit evaluation of the product is. Logically, lower and negative values would indicate negative implicit evaluation.
In order to assess the explicit evaluations, each respondent was asked three questions – “Do you judge Puma favorably?”, “Do you think Puma is a product of high quality?” and “And do you think Puma makes you look cool?” Respondents had to give evaluations on a rating scale from 1 (not at all) to 7 (definitely). These items formed the overall index of the respondents’ explicit product evaluations (Chronbach’s alpha = .799).
Results
implicit and explicit attitudes toward the brand. Both dependent variables were z-transformed for the purposes of the analyses. Analyses were performed twice – one using the transformed values and once using the untransformed values and results were identical. Therefore, for ease of presentation the raw untransformed values will be reported, while z-scores will only be used for direct comparison (Appendices A and B). Initially the study had eighty participants in total, but two consequent residual checks led to the exclusion of two of them, thus ending with the seventy eight participants from which the following results were derived.
As argued in H1, the level of online ads’ intrusiveness is positively related to the implicit attitudes toward the brand depicted in the ad. Thus, it is expected that the implicit attitudes toward the brand will be significantly higher when high intrusiveness ads are used in comparison to low intrusiveness. Contrary to the predictions made, the results did not show such an effect – F(1,74) = .28, p = .60. In fact, implicit attitudes formed as a result of low intrusiveness ads were more positive (M = .77, SD = 2.81) than those formed as a result of high intrusiveness ads (M = .43 , SD = 2.72). Therefore, no support is found for H1. For H2 I argued that the level of online ads’ intrusiveness is positively related to the explicit attitudes toward the brand. In other words, increasing the intrusiveness of ads from low to high should
lead to more positive explicit attitudes toward the brand by the participants. In line with my reasoning, such a relationship did exist – F(1,74) = 4.71, p = .03. Explicit attitudes toward the brand formed after exposure to low intrusiveness ads (M = 4.67, SD = .87) were significantly lower in comparison to those formed after exposure to high intrusiveness ads (M = 5.14, SD = .99). The third and fourth hypotheses argued for respectively - the existence of moderating effect of congruence on the relationship between intrusiveness and implicit attitudes toward the brand (H3) and the existence of moderating effect of incongruence on the relationship between intrusiveness and explicit attitudes toward the brand (H4). For H3 the implicit attitudes toward the brand would be more positive for ads that are congruent with the environment they are placed in. For H4 the explicit attitudes toward the brand would be more positive for ads that are incongruent with the environment they are placed in. Contrary to the predictions made, such moderating effects were not present for any of the hypotheses –
F(1,74) = .05, p = .81 for H3 and F(1,74) = .65, p
3.12 for high intrusiveness). However, as simple comparison between the means can show, the low intrusiveness ads were judged more favorably than high intrusiveness ones, which contradicts our hypothesis. Therefore, H3 is rejected. Explicit attitudes towards the brand in the incongruent setting were more positive (M = 4.80, SD = .95 for low intrusiveness and M = 5.19, SD = .94 for high intrusiveness) in comparison to the explicit attitudes toward the brand for the congruent scenario (M = 4.54, SD = .80 for low intrusiveness and M = 5.18, SD = 1.06 for high intrusiveness). Although the results are in the predicted the direction, their significance was not strong enough. Therefore, H4 is also rejected.
Figure 1 – Mean product evaluations (z-transformed) for all four scenarios (cf. Hafner and Trampe, 2009)
Discussion and Conclusions
The central foundational question of this paper, aimed at unfolding whether congruity holds moderating power over the relationship between web-site ad intrusiveness and both implicit and explicit attitudes toward the brand, remains unanswered. The results section provided enough evidence to conclude that there is no significant support for the proposed moderating powers of congruence. On the other hand the effects of intrusiveness over the explicit attitudes toward the brand were proven. Thus, although failing to achieve its core goal, the study still provides both managers and researchers alike with important conclusions.
The two initial hypotheses I started this paper with argued that the independent variable of intrusiveness is positively related to both implicit (H1) and explicit attitudes toward the brand (H2) – the two dependent variables. Furthermore, based on several well renowned theories, two more hypotheses were stated, the first of which arguing for existing moderating powers of congruity over the relationship between intrusiveness and implicit attitudes toward the brand (H3), while the second explored the same argument but for the other dependent variable – explicit attitudes toward the brand (H4). No support was found for the first hypothesis, meaning there is no positive
relationship between intrusiveness and implicit attitudes toward the brand. In line with my predictions for the second hypothesis, very strong support was found for an existing positive relationship between intrusiveness and explicit attitudes toward the brand. In other words, with the increase of ad intrusiveness – switching from simple banner ads to pop-up no-skip ads – explicit attitudes towards the brand depicted in the ad grew more positive. Lack of support was found for the two hypotheses arguing for congruence’s moderating powers over the relationships between intrusiveness and implicit attitudes toward the brand and intrusiveness and explicit attitudes toward the brand. Thus, although these relationships are well grounded in theory with regard to traditional media, there is no proof that congruence has the same moderating power within the online environment that is the World Wide Web.
From a researcher’s point of view the paper poses more questions than the ones it managed to answer and in that way reveals the complicated nature of the issues it had to deal with. My work failed at proving existing theories for other media as holding fully for the internet space, meaning simpler relationships should be explored first before engaging into a model such as the one I utilized and which turned out to be a bit over complicated. However, this paper did manage to prove one existing relationship – that
between intrusiveness and explicit attitudes toward the brand. With that, researchers in the fields of both marketing and consumer research can use this article as a reliable first step towards further exploration. Having found proof for online ads intrusiveness affecting explicit evaluations, it is clear that the greater the interaction between the ad and the person viewing it, the more positive the explicit attitudes toward the brands are. Thus, the proper path towards advertising effectiveness by the form of explicit attitudes toward the brand is now known, meaning researchers can now search for the most efficient way towards advertising effectiveness. When the experiment for this study was devised, only one type of high intrusiveness ads was utilized. In other words, researchers should perform this experiment once more, but instead of using different levels of intrusiveness, concentrate on high intrusiveness only and use more than one type of high intrusiveness ads – e.g. audio, video and animation pop-ups. Thus, the research implications of this article, although limited, represent a reliable starting point for the exploration of an intriguing topic that has yet to receive increased attention.
concerning the online environment in order to provide the reader with a new unified guideline for properly engaging into marketing activities within the web. On the one hand, evidence was provided for an existing relationship between intrusiveness and explicit attitudes toward the brand, while on the other - no support was found for any of the other three hypotheses. This is a clear indication of how complicated understanding the online space is and the effort needed to set up reliable ground rules of tackling the challenges presented to a marketer upon entering the battle arena that is the World Wide Web. This is one of the reasons why this article is to be considered of high importance for any marketer. The evidence of an existing positive relationship between intrusiveness and explicit attitudes toward the brand has direct implications for business considering advertising within the web-space. First of all, it is now clear that the main question a marketer should ask himself is not whether to blend in or stand out. Based on the lack of support for congruence’s moderating power over the relationship between intrusiveness and both explicit and implicit attitudes toward the brand – it has never been the right question. Instead, a marketer should ask himself how to get noticed. Based on the findings of this article, the more intrusive the ad, the higher and more positive the explicit attitudes toward the brand depicted in the ad. Therefore, interrupting people while
browsing and exposing them to an ad is the key to success. Moreover, highly intrusive ads negate the effects of “browsing experience” – the ability to identify advertisements and completely ignore them (Dahlen, 2001). Of course, being hardly acquainted with the properties of the online space, other approaches will become available with time, some of which may be even more effective and able to affect consumers’ implicit attitudes toward the brand. Yet, for now, based on what this paper managed to prove – making sure prospective customers are successfully exposed to an advertisement and not being given the chance to switch their attention away from it is the single most reliable strategy a marketer can use in order to positively enhance explicit attitudes toward the brand.
and the ad itself are blocked before even being processed in the mind of the customer. This effect becomes automated and its execution becomes faster and faster with highly experienced internet users. With the constantly changing internet environment, people using the internet learn constantly – there is no optimum level of knowledge. The participants of the experiment were all university students around the age of twenty, which are in general highly experienced when it comes to the internet. With that, the effect of the banners used for the purposes of this study might not have had effect on them, since, unlike the pop-up no-skip ads, they did not require a given amount of attention to be paid to them before being able to continue reading the information on the site. That might explain the fact that both hypotheses dealing with implicit attitudes toward the brand were rejected. Moreover, this paper tried to explore a topic that has yet to receive great research attention, by utilizing a complex contextual model of diverse causal relationships. That called for a highly complex testing procedure, the results of which supported only one of the four hypotheses this paper intended to prove. Therefore, in order to tackle the limitation of this paper in the future, researchers are advised to inspect more basic causal relationships before moving forward to more complicated ones. However, with setting the bar higher than expected as a common
practice within the scientific fields, researchers are welcome to accept the challenge by adopting the same conceptual model and testing procedure, but incorporate heterogeneous groups of participants instead of the homogenous that this one had available, in an attempt obtain results that would prove the relationships that I failed to provide evidence for.
References
Aaker, David A., Douglas M. Stayman, and Michael R. Hagerty , (1986) "Warmth in Advertising: Measurement, Impact, and Sequence Effects," Journal of Consumer
Research, 12 (4)
Bellman, Steven, Anika Schweda, and Duane Varan, (2009) "A Comparison of Three Interactive Television Ad Formats," Journal of
Interactive Advertising, 10 (1)
Berlyne, D. E. (1970) "Novelty, Complexity, and Hedonic Value.", Perception and Psychophysics 8 Brennan, Ian, Khalid M. Dubas, and Laurie A. Babin, (1999) "The Influence of
Placement Type and Exposure Time on Product-Placement Recognition," International Journal of
Advertising, 18 (3)
The Role of Message Repetition." In
Psychological Processes and Advertising Effects,
L. F. Alwitt and A. A. Mitchell, eds. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.
Cho, Chang-Hoan. (1999) “How Advertising Works on the WWW: Modified Elaboration Likelihood Model.” Journal of Current Issues &
Research in Advertising
Cho, Chang-Hoan; Lee, Jung-Gyo; Tharp, Marye. (2001) “Different Forced-Exposure Levels to Banner Advertisements.” Journal of Advertising
Research
Cobb-Walgren, Cathy J., Cynthia A. Ruble, and Naveen Donthu, (1995) "Brand Equity, Brand Preference, and Purchase Intent," Journal of
Advertising, 24 (3)
Coulter, Keith S. (1998) “The Effects of Advertising Responses to Media Context on Advertising Evaluations.” Journal of Advertising, 27 (4)
Dahlen, Micael. (2001)“Banner Advertisements through a New Lens.” Journal of Advertising
Research
Gong, Wen; Maddox, Lynda M.; (2003)”Measuring Web Advertising
Effectiveness in China” Journal of Advertising Research, Vol 43(1)
Gotlieb, Jerry B. and Dan Sarel,
(1991)"Comparative Advertising Effectiveness: The Role of Involvement and Source Credibility,"
Journal of Advertising, 20 (1)
Fang, X., S. Singh, and R. Ahluwalia. (2007) "An Examination of Different Explanations for the Mere Exposure Effect." Journal of Consumer
Research, 34 (1)
Grunert, K. G . "Automatic and Strategic
Processes in Advertising Effects. (1996)" Journal
of Marketing, 60 (4)
Hansen, Flemming, Jorgen Kai Olsen, and Steen Lundsteen ,(2006) "The Effects of Print vs. TV Advertising, Documented Using Short-Term Advertising Strength (STAS) Measures,"
International Journal of Advertising, 25 (4)
Häfner, Michael; Trampe, Debra (2009). “When Thinking is Beneficial and When It Is Not: The Effects of Thin and Round Advertising Models.”
Journal of Consumer Psychology, 19 (4)
Hofacker, Charles F.; Murphy Jamie. (1998) "World Wide Web Banner Advertisement Copy Testing." European journal of Marketing, 32 (7) Hovland, Carl I.; Janis, Irving L.; Kelley, Harold H. (1953). Communication and persuasion; New Haven: Yale University Press
for Internet Advertising Planning.” Journal of
Advertising, Vol 35(2)
Interactive Advertising Bureau (IAB), (2009) "IAB Internet Advertising Revenue Report," from IAB Website: http://www.iab.net/media/file/IAB-Ad-Revenue-Full-Year-2009.pdf
Kamins, Michael A., (1990) "An Investigation into the 'Match-Up' Hypothesis in Celebrity Advertising: When Beauty May be Only Skin Deep," Journal of Advertising, 19 (1)
MacKenzie, Scott B. and Richard J. Lutz (1989) "An Empirical Examination of the Structural Antecedents of Attitude toward the Ad in an Advertising Pretesting Context," Journal of
Marketing, 53 (2)
Mitchell, Andrew A.; Olson, Jerry C. (1981) ”Are Product Attribute Beliefs the Only Mediator of Advertising Effects on Brand Attitude?” Journal
of Marketing Research (JMR), 18 (3)
Moore, Robert S.; Stammerjohan, Claire Allison; Coulter, Robin A. (2005) “Banner Advertiser – Web Site Context Congruity and Color Effects on Attention and Attitudes.” Journal of Advertising, 2 (2)
Moorman, Marjolein; Neijens, Peter C.; Smit, Edith 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, Vol. 31 (4)
Olney, Thomas J., Morris B. Holbrook, and Rajeev Batra ,(1991) "Consumer Responses to Advertising: The Effects of Ad Content, Emotions, and Attitude toward the Ad on Viewing Time," Journal of Consumer Research, 17(4)
Russell, Cristel A.,(1998) "Toward a Framework of Product Placement: Theoretical
Propositions," Advances in Consumer Research, 25 (1)
Stang, David J. (1975) ”Effects of “Mere Exposure” on learning and affect”. Journal of Personality and Social Psychology, Vol 31(1) Yi, Youjae (1990b), “The Effects of Contextual Priming in Print Advertisements," Journal of
Consumer Research, 17 (9).
Yoo, C. Y. (2005) "Preattentive Processing of
Weh Adxwrtising,"Dissertation from the
University of Texas at Austin.
Zajonc, Robert B. (1968) "The Attitudinal Effects of Mere Exposure." Journal of Personality and
Social Psychology, 9
Zeff, Robbbin; Brad, Aronson. (1999) ”Advertising on the internet”. New York, NY: John Wiley & Sons, Inc.
APPENDIX A – ANOVA output with Z-scores Between-Subjects Factors Value Label N CV_congruity 1.000 Congruent Setting 39 2.000 Incongruent Setting 39
IV_intrusiveness 1.000 Banner ad - Low intrusiveness 38 2.000 Pop-up ad - High intrusiveness 40
Tests of Between-Subjects Effects
Source Dependent Variable
Type III Sum of
Squares df Mean Square F Sig. Corrected Model Zscore(DV_implicit) .543a 3 .181 .173 .914
Zscore(DV_explicit) 4.907b 3 1.636 1.842 .147 Intercept Zscore(DV_implicit) .011 1 .011 .011 .919 Zscore(DV_explicit) .260 1 .260 .293 .590 CV_congruity Zscore(DV_implicit) .199 1 .199 .190 .664 Zscore(DV_explicit) .156 1 .156 .175 .677 IV_intrusiveness Zscore(DV_implicit) .293 1 .293 .280 .598 Zscore(DV_explicit) 4.188 1 4.188 4.716 .033 CV_congruity * IV_intrusiveness Zscore(DV_implicit) .056 1 .056 .053 .818 Zscore(DV_explicit) .578 1 .578 .651 .422 Error Zscore(DV_implicit) 77.419 74 1.046 Zscore(DV_explicit) 65.709 74 .888 Total Zscore(DV_implicit) 77.970 78 Zscore(DV_explicit) 70.825 78 Corrected Total Zscore(DV_implicit) 77.961 77 Zscore(DV_explicit) 70.616 77 a. R Squared = .007 (Adjusted R Squared = -.033)
Estimated Marginal Means
1. CV_congruity
Dependent Variable CV_congruity Mean Std. Error
95% Confidence Interval Lower Bound Upper Bound Zscore(DV_implicit) Congruent Setting .062 .164 -.264 .389
Incongruent Setting -.039 .164 -.365 .288 Zscore(DV_explicit) Congruent Setting -.102 .151 -.403 .198 Incongruent Setting -.013 .151 -.314 .288
2. IV_intrusiveness
Dependent Variable IV_intrusiveness Mean Std. Error
95% Confidence Interval Lower Bound Upper Bound Zscore(DV_implicit) Banner ad - Low
intrusiveness
.073 .166 -.257 .404
Pop-up ad - High intrusiveness
-.049 .162 -.372 .273
Zscore(DV_explicit) Banner ad - Low intrusiveness
-.290 .153 -.594 .015
Pop-up ad - High intrusiveness
3. CV_congruity * IV_intrusiveness
Dependent Variable CV_congruity IV_intrusiveness Mean Std. Error
95% Confidence Interval Lower Bound Upper Bound Zscore(DV_implicit) Congruent Setting Banner ad - Low
intrusiveness
.150 .235 -.317 .618
Pop-up ad - High intrusiveness
-.026 .229 -.481 .430
Incongruent Setting Banner ad - Low intrusiveness
-.004 .235 -.472 .464
Pop-up ad - High intrusiveness
-.073 .229 -.529 .382
Zscore(DV_explicit) Congruent Setting Banner ad - Low intrusiveness
-.420 .216 -.851 .010
Pop-up ad - High intrusiveness
.215 .211 -.204 .635
Incongruent Setting Banner ad - Low intrusiveness
-.159 .216 -.589 .272
Pop-up ad - High intrusiveness
APPENDIX B – Raw Data Means
Descriptive Statistics
CV_congruity IV_intrusiveness Mean Std. Deviation N DV_implicit Congruent Setting Banner ad - Low
intrusiveness .98505 3.154799 19 Pop-up ad - High intrusiveness .50235 2.347063 20 Total .73751 2.743822 39
Incongruent Setting Banner ad - Low intrusiveness .56189 2.491251 19 Pop-up ad - High intrusiveness .37215 3.126921 20 Total .46459 2.799624 39
Total Banner ad - Low intrusiveness .77347 2.811966 38 Pop-up ad - High intrusiveness .43725 2.729751 40 Total .60105 2.757228 78
DV_explicit Congruent Setting Banner ad - Low intrusiveness 4.54386 .802854 19 Pop-up ad - High intrusiveness 5.18333 1.062244 20 Total 4.87179 .987096 39
Incongruent Setting Banner ad - Low intrusiveness 4.80702 .951385 19 Pop-up ad - High intrusiveness 5.10000 .949607 20 Total 4.95726 .949544 39
APPENDIX C – Screenshots of the ads as they appeared on the web sites.
Banner ad – example of low intrusiveness ad
