Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. The Value of Implicit Measures for Internet Advertising: Evidence of Mere Exposure Effects and their Impact on Consumer Preference for Pop-Ups and Banner Ads A thesis presented in partial fulfilment of the requirements for the degree of Master of Arts . In Psychology at Massey University, Albany, New Zealand Kuen Fong (Jenny) Cheung 2009 II Acknowledgements I would like to express my deep and sincere gratitude to my supervisor, Dr. Heather Buttle, for her contributions, support and patience over the past years. Her understanding, encouraging and personal guidance have provided a good basis for the present thesis. I would like to express my appreciation to the participants who gave their precious time and effort for this study. Also, the support staff in the Psychology Department, especially Kathryn Lee and Janet Mak, for all the instances in which their assistance helped me along the way. Thanks also go to Darrell Ng for proof-reading my thesis, and to Massey University's library for assistance with literature search and recommendations of books. Finally, I would like to thank my family for all the support they provided me through my entire life, and in particular I must acknowledge my partner and best friend, Chris, without whose encouragement and understanding, I would not have finished this thesis. Also, thanks to my fellow students and friends, particularly Tiffany Chew, Elizabeth Howells, and Namrata Gulani, for our debates, exchanges of knowledge, skills, and venting of frustration during my graduate years, which helped enrich the experience. Thank you. 111 Contents Abstract 1 Introduction 2 Internet Advertising 2 The Effectiveness of current Internet advertising campaigns 3 Literature Review 7 Seeing Leads to Liking: Evidence from "mere exposure" effect for visual stimuli 7 Subliminal Presentation Paradigm 8 Affective Priming Hypothesis: Affect without Cognition 10 Mere Exposure Effect and Advertising 11 Use of Explicit memory measures by Advertisers: 'Information only available if remembered' 12 Implicit memory measures: Evidence from cognitive psychology literature 15 Jacoby's Attribution Theory of Memory: Fluency of Processing 16 Bridging the gap between implicit memory measures and advertising research 18 Pe1fect & Askew (1994): Print Adverts are not remembered, but memorable 21 The Scope of the Present Study 22 Affective Influences of Selective Attention: Investigating the Internet setting 23 Investigation for Internet adverts: Pop-Ups and Banner Ads Pop-Ups Banner ads Exposure Durations and Mere Exposure Effects Hypotheses 24 25 26 27 28 Method Results Participants Apparatus Materials and Stimuli Design Procedure Familiarity Effects Figure 1: Mean familiarity ratings for each group for pop-ups and banner ads, previously seen and not-seen Recognition Distinctive Eye-catching Memorable Appealing Figure 2: mean ratings for previously-seen and not-seen adverts items for each of the 5 affective dimensions Discussion Recognition and Pop-ups vs. Banner ads Mere Exposure Effects and Pop-ups vs. Banner ads Exposure durations and pop-ups vs. banner ads Mere Exposure Effects vs . Attentional Inhibition Familiarity effects Limitations Bias of Familiarity Effects Familiarity effects: Physical features of the adverts Familiarity effects: Prior experience with adverts IV 29 29 29 29 31 32 33 33 35 35 36 36 37 37 38 39 39 39 40 42 43 45 45 45 46 Practical Implications and Suggestions for Future Research Quality vs. Quantity of advertising exposure Investigation. for Attention.al Inhibition: Relaxed-viewing vs. Goal-directed context Investigation of long-term effects underlying the mere exposure effect Future Investigations for new on line ad formats Measurement and Design issues Contributions to Research Conclusion References Appendices Appendix 1: lnfornwtion sheet Appendix 2: Reading task instructions Appendix 3: Samples of the advertising stimuli used in the experiment Appendix 4: An alphabeticised list of the adverts used in the experiment 47 47 49 50 51 52 53 54 55 62 62 64 65 66 V Abstract Cognitive and neuropsychological research points to evidence that "repeated, unreinforced exposure to a stimulus is sufficient to enhance one's attitude toward that stimulus", referred to as the mere exposure effect. The main purpose of this study is to bring to attention sources of mere exposure effects data that, while rarely cited within the advertising literature, are highly relevant to questions of online advertising effectiveness. This study further sought to 1) investigate whether if pop-ups would be better recognised than banner ads, and 2) whether if pop-ups would be rated more favourably (due to the mere exposure effect) than banner ads (less liked due to devaluation-by-inhibition). The experiment was a 2 (Group A, Group B) x 2 (Advert Type: Banner Ads, Pop-ups) x 2 (Whether Seen: Seen, Not Seen) design. Twenty participants were instructed to read twelve articles on the screen, where each reading trial was presented with a banner ad located in the peripheral display of the full-screen article, and a pop-up that was sequentially displayed at the centre of the screen which would be instructed to click away. A total of 48 adverts (i.e., previously seen and not-seen pop-ups and banner ads) were then rated on each of the five affective dimensions (i.e., familiarity, eye-catching, appealing, memorable, and distinctive); followed by a recognition question as to indicate whether participants had seen the advert whilst reading the articles. As predicted, mere exposure effects occurred for all previously exposed adverts, but no difference in recognition was found between the advert types or any significant differences in the magnitude of favourability for the previously seen pop-ups and banner ads. These findings suggest that the magnitude to which favourability would be related to varying lengths of exposure durations presented with pop-ups and banner ads (i.e., brief vs. longer presentation times), and/or the level of attention guided towards pop-ups and banner ads (i.e., central, had to be clicked away vs. peripheral, could be ignored), was not supported. Practical implications and suggestions for future research regarding potential applicability of mere exposure theory in an Internet setting will be discussed. Kuen Fong (Jenny) Cheung 2 Introduction Intemet Advertising Internet advertising has undoubtedly attracted an enormous amow1t of investment from the commercial advertising industry. According to the Interactive Advertising Bureau (IAB) (2007), Internet advertising revenue increased by 26% from $US16.9 billion in 2006 to $US21.2 billion in 2007. In recent years, the Internet has evolved into an important medium for advertisers and marketers for both branding and direct-selling purposes (Hoffman & Novak 1996; Shen, 2002). During 2007, total spending on Internet advertising exceeded that of radio and cable television (TV) in the United States (IAB, 2007). It has been predicted that Internet advertising spending could surpass that of magazine advertising by 2010 (ZenithOptimedia, 2007). The commercial domain is by far the largest and fastest growing segment on the World Wide Web, with consumer-related advertisers continuing to represent the largest category of Internet advertising spending, accow1ting for 55 % of total advertising revenues in 2007. In particular, display-related advertising (e.g., display ads, rich media) accounted for 34 % of full-year revenues in 2007. Driving this rapid growth is the fact that the Internet has been drawing sizable audiences from other types of media (e.g. TV, magazines), which has turned the Web into a far more attractive venue for mass market advertisers (IAB, 2007; Shen, 2002). Nielsen/ etRatings (2008) estimates there are more than 1.4 billion Internet users worldwide with this number expected to grow dramatically every year. Technological innovations have made the Internet an attractive medium for advertisers. Server-based technologies enable advertisers to display banner ads according to user profiles and interests. As an advertising medium, the Internet offers all the elements of other media and much more. Banner advertisements (or more commonly known as banner ads) can include everything from simple text and graphics, streaming audio and/or video, to highly dynamic plus interactive banner ads delivered with rich-media animations (Shen, 2002). In addition to conventional banner-ads, interstitials (or pop-up Kuen Fong (Jenny) Cheung ads) represent another popular techniques used to deliver rich media ads that can contain more sophisticated messages on the Web (Huang & Lin, 2006). The Effectiveness of current Internet advertising campaigns 3 The growth in Internet advertising comes with an increasing demand for effective advertising and media planning strategies. Despite the Internet's phenomenal growth, measurement and pricing practices on the Web are far from being standardised. Added to this is the need for accurate information and tracking of site traffic, advertising delivery, and user response for both Web publishers and advertisers (Shen, 2002). Although numerous efforts have been made to facilitate the process of developing viable measurement and pricing standards, such as guidelines proposed by the IAB in 1997 and 2002, attempts at standardisation thus far have been w1successful (Bhat, Bevans, & Sengupta, 2002). Ongoing discussions of measurement standardisation have been based on whether Internet advertising effectiveness should be assessed by direct response (i.e., click-through) or through mere exposure (i.e., impressions). Because the Internet enables advertisers to track responses to Internet ads, some reason that advertisers should pay for their ads on the basis of responses or performances, as evident in most online advertising campaigns that click-through to websites; while others argued that Internet ads have a role in creating brand awareness and image, and that such pricing and measurement methods would dismiss online advertising's brand-building value (see Danaher & Mullarkey, 2003; Shen, 2002 for reviews). Despite the importance of this issue for both Web publishers and advertisers, there is little research available on the methodology used by advertising agencies to price, measure, and pretest Internet ads. A recent review by Shen (2002) has addressed this gap and sought to determine which measures and methods are being used to gauge the effectiveness of banner advertising. The majority of advertising agencies have indicated the use of click-through rates (i.e., the percentage of users who click on ad baimers to connect to an Kuen Fong (Jenny) Cheung 4 advertisers' site) and outcomes (i.e., the number of products purchased) to measure campaign effectiveness, rather than banner ad exposures or impressions (i.e., a broad indicator of the overall exposure of an ad that measures success in terms of visibility or brand recognition). This use of click-through has highlighted the interactivity of the Internet, indicating an active response toward advertising, rather than probable exposure to it which is believed to be essential in gauging the performances of online campaigns. Conversely, advertisers remained sceptical about the brand-building capabilities of on.line advertising, due to concerns that they have limited control, if any, on the exposure of advertised content. It is believed that the exposure-based approach may not be a good indicator of user involvement with an ad, as it does not track whether users interacted with the ad (Bhat et al., 2002; Hoffman & Novak, 1996). Regardless of all this, there have been concerns that the accepted click-through approach may still not be appropriate to evaluate the success of online advertising campaigns (e.g., Shen, 2002) . It is argued that relying on click-through as a metric of effectiveness could underestimate the brand-building power of online advertising. This argument points to evidence that online advertising can create a brand- building effect in the absence of click-through, based on the notion that measuring exposure to an ad can lead consumers to go directly to the site reference or look up the Website or product later, regardless of whether or not they click on the ad (Bhat et al., 2002; Briggs & Hollis, 1997). These concerns are also consistent with the evidence that online commerce has experienced a downturn in the recent past, as determined by declining click-through rates registered for most Internet advertising (Danaher & Mullarkey, 2003; Hanson 2000). As a result, more recent studies have made attempts to use traditional survey research to provide insights into viewers' attitudinal motivations (e.g., Chandon, Mohamed, & Fortin, 2003; Danaher & Mullarkey, 2003; Hollis, 2005). These studies have relied on explicit memory retrieval using conventional measures, such as ad memory, brand awareness, and purchase intention, to anticipate attitudinal behaviours (Shen, 2002). For example, Dynamic Logic's Adindex has become the industry standard for measurement of the Kuen Fong (Jenny) Cheung 5 attitudinal effects of online advertising, which in one study has found that, a single exposure to an on.line advertisement can have a positive impact on online advertising awareness, total brand awareness, and purchase consideration (Hollis, 2005), as summarised by the observation that "the least 20 memorable advertisements negatively impacted brand favourability and purchase intent and had little to no impact on brand awareness and online ad awareness" (Mallon, 2005, cited in Hollis, 2005, p. 262). Similarly, Danaher and Mullarkey (2003) concluded that the longer a person is exposed to a web page containing a banner ad, the more likely one will remember the ad. In contrast, cognitive and neuropsychological research points to evidence that affect is generated for a stimulus by merely exposing the subject to that stimulus repeatedly, referred to as the mere exposure effect, rather than being based on the cognitive processing of information (e.g., Zajonc, 1980, 2000). This strongly suggests that explicitly remembering an advert is rnmecessary, and perhaps even irrelevant, to the change in attitudes that result from a single exposure to Internet advertising (Perfect & Askew, 1994). Therefore, open questions can be raised regarding the reliability of existing design and measurement methods used for online advertising - n-1ost importantly, can the traditional practices of advertising campaigns reliably determine the true impact of online advertising effectiveness? In summary, it still remains unclear as to whether or not Internet ads are effective as a medium for marketing communication. Although there is greater interest in understanding which methods or measures work better in gauging Internet advertising effectiveness within the advertising literature (e.g., Shen, 2002), there is little available comparative and empirical research to resolve the debate. Given the similarities and differences between the Internet and traditional media, it is expected that mere exposure effects on the Web may differ or be similar in certain ways to traditional media (e.g., television commercials, print advertising), and this deserves further investigation (Hoffman & Novak 1996; Wong & Fortin, 1999). Specifically, it has been suggested that the majority of Internet advertisements do not receive any active processing due to the attention-demanding nature of tasks involved; this may imply that Kuen Fong (Jenny) Cheung Internet ads receiving limited attention compete for control over resources that may influence affective evaluation (Edwards, Li, & Lee, 2002). 6 However, the context of Internet advertising may somewhat differ from traditional media due to the interactive nature of the Internet. The online medium allows advertisers to present Internet ads in a form (e.g., pop-up, banner ads) that requires sufficient interest and motivation on the part of the consumers to interact with the advertisement and access appropriate information, rather than being passive recipients of the message (Hanson, 2000). In addition, given the empirical evidence from cognitive and neuropsychological research, it is striking that those who have recently drawn attention to the attitudinal outcomes (e.g., brand awareness) have largely dismissed the underlying mechanisms that would influence consumers' responses resulting from exposure to a banner ad, namely the mere exposure effect. Although empirical studies of mere exposure effects in an advertising context are still rare, growing research attention has shown the importance of non-conscious processes on consumer choice, suggesting that advertising effectiveness can be shown in the absence of explicit recollection. In particular, Perfect and Askew (1994) have shown the existence of mere exposure in print advertising by measuring unconscious effects of prior experience through implicit tests of memory. In contrast to the widespread notion that the issue of on.line advertising effectiveness has been settled in favour of the direct response approach due to the universal measurement of click-through (e.g., Hollis, 2005), the main purpose of this study is to bring to attention sources of mere exposure effects data that, while rarely cited within the advertising literature, are highly relevant to questions of online advertising effectiveness. Kuen Fong (Jenny) Cheung