Effectiveness and consumer preference of online advertising

EXECUTIONAL CUES, INTERACTIVITY, AND LEVEL OF INVOLVEMENT IN BANNER ADVERTISEMENT: AN EXPERIMENTAL APPROACH TO UNDERSTAND ONLINE ADVERTISING EFFECTIVENESS HUANG SHANSI (B.A., FUDAN UNIVERSITY) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF ARTS COMMUNICATIONS AND NEW MEDIA PROGRAMME NATIONAL UNIVERSITY OF SINGAPORE 2007 ACKNOWLEDGEMENTS I would like to gratefully express my sincere appreciation to all the people who in some way contributed to the completion of this dissertation either with support and encouragement, or with discussion and advice. First and foremost, I owe many thanks to my supervisors Dr. Hichang Cho and Dr. Byungho Park, whose invaluable insight, stimulating suggestions, and precious guidance helped me finally complete the long journey. I am also deeply indebted to Dr. Siyoung Chung and Mr. Raaj Chandran for their constant encouragement and crucial support during my experiment. I would like to express my gratitude to the Communication and New Media Programme Head, Prof. Millie Rivera for supporting me with a graduate scholarship for 2 years. I also sincerely thank all the faculty members of CNM for their inspiring lectures and seminars. I am thankful to all the fellow graduate students of CNM for valuable sharing and help when I was in need. My gratitude also extends to the registered students of NM3215 “Advertising Strategy” and NM1101 “New Media and Society” in Semester I, 2006-2007 Academic Year for giving up their precious time in participating my experiment. Special appreciation goes to my parents for their continued support all my life, instilling in me their morals and values, and steering me in a lifetime quest for knowledge. Last but not least, I am most thankful to my dear husband, Jinghao, who besides tolerating my many moods and swings, had confidence and faith in me and assisted me throughout all the difficulties. i Table of Contents ACKNOWLEDGMENTS………………….……….…………………………….…i TABLE OF CONTENTS…………………………………………………………….ii ABSTRACT…………………………………………………………….....................iv LISTS OF TABLES……………………………………………….............................v LISTS OF FIGURES……………………………………………………………….vi CHAPTER 1 INTRODCUTION……………. …………………………………….1 CHAPTER 2 LITERATURE REVIEW……………………………………………6 2.1 Online Advertising………………….…………………………….......................6 2.2 Conceptualization of Online Advertising Effectiveness…………………...........8 2.2.1 Criteria of Effectiveness…………………………………………………...8 2.2.2 Measures of Effectiveness of Online Advertising………………………..13 2.3 Executional Cues & Effectiveness……………………………………………..19 2.4 Involvement & Effectiveness…………………………………………………..26 2.5 Interactivity…………………………………………………………………….32 CHAPTER 3 METHODOLOGY…………………………………………….……39 3.1 Research Design………………………………………………………………..39 3.2 Participants……………………………………………………………………..40 3.3 Stimuli………………………………………………………………………….40 3.4 Pretest…………………………………………………………………………..45 3.5 Procedure……………………………………………………………………….46 3.6 Measures………………………………………………………………………..47 3.6.1 Involvement………………………………………………………………47 3.6.2 Advertising Effectiveness………………………………………………...49 3.6.3 Other Covariates………………………………………………………….50 CHAPTER 4 DATA ANALYSIS AND RESULTS..……………………….……... 51 4.1 Sample Size and Composition…………………………………………………51 4.2 Descriptive Statistics & Scale Reliability……………………………………...52 ii 4.3 Executional Cues & Effectiveness……………………………………………..53 4.4 Level of Involvement & Effectiveness…………………………………………56 4.5 Interaction between Involvement & Executional Cues………………………...58 4.6 Interactivity & Effectiveness…………………………………………………...60 4.7 Interaction between Involvement & Interactivity.……………………………...61 4.8 Summaries of Findings…………………………………………………………63 CHAPTER 5 DISSUSTION, CONCLUSIONS, AND IMPLICATIONS……….66 5.1 The Impact of Executional Cues.………………………………………………66 5.2 The Impact of Interactivity……………………………………………………..68 5.3 The Impact of Involvement…………………………………………………….71 5.4 The Moderating Effects of Involvement……………………………………….73 5.5 Theoretical Implications……………………………………………………......77 5.6 Managerial Implications………………………………………………………..79 5.7 Limitations & Suggests of Future Study……………………………………….81 REFERENCES……………………………………………………………………...84 APPENDIX I TABLE OF MEASURES & QUESTIONS………………………..96 APPENDIX II EXPERIMENTAL BANNERS……………………………………98 APPENDIX III INDEX WEB PAGE……………………………………………..104 APPENDIX IV SAMPLE OF MAIN QUESTIONNAIRE……………………...105 iii ABSTRACT This study attempts to discover the source of effectiveness in banner advertisements by exploring the effects of several potential factors such as executional cues, level of involvement, and interactivity on advertising effectiveness. This study employed the theoretical premise developed on the role of consumer’s involvement on advertising by Richard E. Petty and John T. Cacioppo’s Elaboration Likelihood Model (ELM). An experiment was conducted using custom-tailored banner advertisements to facilitate either peripheral or central route of advertising message processing. It was found that the level of advertising involvement had a significantly positive relationship with the effectiveness of banner advertisement, which supports the ELM in the context of online advertising generally. Unexpectedly, it was also found that the banner advertisements presented in the format of peripheral cues and higher interactivity did not yield the most desirable effects on advertising effectiveness in generally; while only under low involved situation, banners that employed peripheral cues and non-interactive features were more effective. The findings suggest that advertisers need to take full advantage of the consumer involvement to produce effective Web advertising. However, the enhanced capabilities of the new medium have little effects on advertising effectiveness. iv List of Tables Table 4.1 The Comparison of Four Experimental Groups............................................52 Table 4.2 Descriptive statistics for scales used in the experiment……………………53 Table 4.3 Results of Advertising Exposure for Three Product Categories……………55 Table 4.4 T-test Results for Effects of Execustional Cues…………………………….56 Table 4.5 Correlations between Involvement & Effectiveness……………………….57 Table 4.6 Low Involvement: T-test Results of Effects of Executional Cues………….59 Table 4.7 High Involvement: T-test Results of Effects of Executional Cues…………60 Table 4.8 T-test Results for Effects of Interactive Features…………………………...61 Table 4.9 High Involvement: T-test Results of Effects of Interactive Features……….62 Table 4.10 Low Involvement: T-test Results of Effects of Interactive Features………63 Table 4.11 Summaries of Findings…………………………………………………….65 Table 5.1 Means Comparison Between Type Peripheral & Central…………………..75 Table 5.2 Means Comparison Between Type Interactive & Non-interactive…………75 v List of Figures Figure 2.1 The Traditional “Hierarchy of Effects” CAB Model………………………9 Figure 2.2 The FCB Grid Model……………………………………………………...12 Figure 2.3 The Venn Diagram of Contemporary CAB Criteria……………………...13 Figure 3.1 Research Design for Banner Type Manipulation…………………………42 Figure 3.2 Experimental Stimuli for Each Experimental Group……………………..43 Figure 3.3 Sample Web Page and Banner Ad………………………………………..44 Figure 3.4 Presentation Orders of Experimental Banners……………………………45 Figure 3.5 Experiment Slots and Participants Details………………………………..46 Figure 4.1 Illustration of Banner Regrouping………………………………………..54 vi Chapter 1 Introduction Within less than twenty years, the Internet has fundamentally transformed the landscape of traditional communication and business. It has not only facilitated global sharing of information and resources, but also provided potential efficient channels for advertising, marketing, and even direct distribution of certain goods and information services (Hoffman, Novak & Chatterjee, 1995). These unique capabilities of the Internet have contributed to the spectacular diffusion of the Web as a new commercial media in the last several years. The world also witnessed an astonishing expansion in the popularity of Internet users, which has reached 1.08 billion in 2005 (Nielsen//NetRatings, 2006). Internet users have been revealed to own higher incomes and have better education than the general population (PEW Internet, 2006). The trend of significant migration from TV and print consumption to Internet usage has also been confirmed, which was up to fifty percent, in the number of users as well as the amount of time spent (Gluck, 2000). In order to cater this promising new market, advertisers show great enthusiasm for expanding their horizons to encompass online advertising. Since the first banner advertisement appeared on the online magazine Hotwired Website in 1994 (Adams, 1995), the growth of online advertising has been exponential. The revenue of online advertising was only $267 million in 1996; it ascended to a historical peak of $8087 million in 2000 (IAB & PricewaterhouseCoopers, 2003). Recently a new record of $3.9 billion for the first quarter of 2006 was announced, marking the highest quarter ever reported (IAB & PricewaterhouseCoopers, 2006). 1 The rapid increment of annual revenues shows the vitality of the burgeoning advertising, as well as the confidence of committing huge funds. This gravity was driven by the continual optimistic reports of the industry. Therefore many companies have been agitated for putting up their online advertisements by indiscriminately applying some online advertising approaches, such as large sizes, animations and 3-D, to each single product/brand, without clear goals for their online advertising presence, or even without the means of measuring whether they are getting a return on their investment. Differences between the new medium and the traditional media were also disregarded. The traditional advertising approaches have been adopted without any reformation, even though researchers have already raised the pressing necessity to refine both of the general principles and measures of traditional advertising to meet the need of new environment. The risks of advertising in the new medium are increased as research into online advertising is still in its infancy. A lot of debates have not been settled in respect of the effectiveness of online advertising. Notwithstanding compelling evidence to prove that online advertising is an effective tool for promoting products (Briggs, 2001; Dreze & Hussherr, 2003; Wakeling & Murphy, 2002), it is not yet clear, however, how to successfully gauge and achieve advertising effectiveness online. The question of online advertising effectiveness becomes especially pertinent in light of the deteriorated click-through rates, less than 2% in most cases (Tuten, Bosnjak & Bandilla, 2000), and the inflated advertising expenses. In 2000, the average CPM (Cost-Per-Thousand impressions), which is the dominant approach to online advertising pricing, is around $34 (Adknowledge, 2000). By contrast, CPMs range from $6-$14 for national television advertising, $8-$20 for magazines advertising, and $18-$20 for 2 newspaper advertising (The Economist, 2006). Obviously, advertisers are struggling to justify online advertising expenditures for one of the largest line items in their marketing budget. In order to be more strategic and precise in planning and optimizing online campaigns, they starve for detailed guidelines of effectively adopting the relevant characteristics of online advertising. This paper attempts to address this unmet need in online advertising effectiveness by exploring the potential influential factors and their respective impacts on online advertising. As one of the outstanding characteristics of online advertising, effects of executional cues have received considerable attention. Executional cues include central cues and peripheral cues. It has been suggested that some executional cues play important role in enhancing awareness, eliciting positive attitudes and engendering purchase intention. As an equally consumer-controlled media, the Internet presents a vast array of possibilities when it comes to the creative execution of these advertising cues, which, nevertheless, might differ from traditional media. In order to further branch the research of executional cues into the new media, this paper focuses on investigating the application of different executional cues for banner advertisement and the impact on advertising effectiveness in the context of online advertising. The literature also suggested the impact of advertising on consumer be predicted and moderated by product involvement, which is perfectly interpreted by Elaboration Likelihood Model (ELM) (Petty & Cacioppo, 1981). Elaboration Likelihood Model postulates two divergent paths of information processing, central route vs. peripheral route, contingently upon consumers' level of involvement and information-processing ability. According to ELM, when people are highly involved, they are motivated to have 3 a deep examination and a diligent consideration of the product-related messages based on past experience or knowledge. The peripheral route portrays the situation when people are low involved with the product. They will pay attention to non-product related stimuli such as advert model, background music or graphics. This study tries to examine how consumer involvement interacts with executional cues in the new environment of online advertising. Interactivity is a unique feature only endowed by the new medium Internet. Owing to its new emergence and complexity, very few researches have touched upon this issue and its impact on advertising effectiveness was far from comprehension. However, there is a conventional belief that interactivity positively boosts advertising effectiveness. This study is designed to determine how interactivity affects effectiveness of online advertising. The interaction between involvement and levels of interaction will also be examined. In sum, the purpose of this study is to examine whether traditional principles of mass media advertising apply in this new commercial environment; whether several acknowledged important characteristics of online advertising can produce short-term effects on each dimension of effectiveness among proficient Internet users. This research aims to answer the following questions: RQ1. Do executional cues, interactivity and involvement have an impact on online advertising effectiveness? RQ2. How does the interaction of involvement with executional cues and interactivity affect online advertising effectiveness? RQ3. How can we manipulate different executional cues or different level of 4 interactivity to enhance the effectiveness of online advertising? This dissertation contributes in several ways to the growing body of advertising knowledge about the Internet medium. First, this study employed Petty and Cacioppo’s Elaboration Likelihood Model as the conceptual framework and empirically examined the model in the context of online advertising. Second, this paper explored the role of customer’s involvement on the effectiveness of online advertising, which has important implications for advertisers to optimize their online campaigns. Third, this study investigated the function and impact of executional cues of banner advertisement. The findings help to provide insights on the value of different executional features and reveal more practical routes of achieving online effectiveness. As such, this study answers the recall by advertising practitioners for furnishing a set of guidelines for online advertising. Finally, this study enriches the literature in interactivity of online advertising, especially indicating potential factors affecting the impacts of interactivity, which may serve as a foundation or a springboard for continuing research in effective interactivity on online advertising persuasiveness. 5 Chapter 2 Theoretical Background 2.1 Online Advertising Online advertising generally follows the same principles as traditional advertising. Nevertheless, the Internet as a unique medium has led to huge differences between online and traditional advertising. Firstly, owing to the digital Web, online advertising is interaction-orientated. Online ads can be directly activated, which is a form of interaction, a kind of user response that provides evidence for the novel role of addressees. Beyond that, different types of online ads also allow different degrees of interactivity (Janoschka, 2004). For example, interactive online game has higher level of interactivity compared to non- synchronous e-mail feedback. Moreover, online advertising is generally considered to be less intrusive since the Web is a pull, not a push medium. That is, advertising message is available to consumers who are willing to reach for and pull it out (Sterne, 1999), while traditional advertising is usually embedded within the program content. By displaying online, it can also be accessed on demand 24 hours a day, seven day a week regardless of geographic location. Furthermore, the multimedia nature of the Web lends full support for multimedia applications of online advertising, which allows it to optionally package the capabilities similar to those of newspaper (text, graphics), radio (audio) and TV (video) (Ainscough & Luckett, 1996; Breitenach & Van Doren, 1998). In addition, online advertising offers advertisers the opportunity of precisely targeting an audience and measuring responses instantly (Zeff & Aronson, 1997) with the help of technologies such as cookies—programs that unobtrusively keep track of a visitor's previous activities on a site. 6 In general, online advertisements can be classified as passive ads and active ads based on the amount of control exercised by the consumer over their exposure (Hoffman & Novak, 1998). Usually, active ads consist of one or more Web pages that are totally dedicated to advertising contents. They customarily locate in the advertisers’ or publishers’ servers and can only be reached by clicking the hyperlinks on related passive ads. Thus, most of the online advertisements now under discussion are considered to be passive ads, which are forcefully exposed to consumers and act as a “gateway” by providing hyperlinks to those active ads, typically in form of banners. Banners are small text and graphic-based billboards that spread across the Web page, primarily aiming at informing users about the existence of certain Web sites or Web pages and persuading them to visit them via clicking. In a sense, the banners are hyperlinks that enable activation through their users. According to annual reports of Interactive Advertising Bureau, banners have been one of the most common formats of online advertising all along since 1999, accounting for over half of the online advertising dollars spent. Thus, this paper focuses on measuring the effectiveness of online advertising in form of banners since the consensus supports the banner advertisements as the dominant and most prevalent form of online advertising. 7 2.2. Conceptualization of Online Advertising Effectiveness 2.2.1 Criteria of Effectiveness Despite its expanded functions, the issue of criteria of online advertising effectiveness is still part of the broader question of advertising effectiveness in general. Thus, the effectiveness of online advertising should be examined in a similar way as that of traditional advertising (Li & Leckenby, 2004; Pavlou & Stewart, 2000). Researchers and advertising practitioners have long sought to understand how advertising works. A number of measures have been proposed to empirically evaluate advertising effectiveness. Some contend that advertising is effective only when it sells (Little, 1979). Others argue that there is a series of stages between the point of unawareness of a product and/or brand and the ultimate purchase/sale of a particular brand, that is, the hierarchy of effects (Colley, 1961; Schultz, 1990). As literature indicates that the direct shot-term sales effect of advertising is, in general, quite low (Aaker & Carman, 1982; Assmus, Farley & Lehmann, 1984; Tellis, 1988), this study thus adopts the latter view of accessing the impact of advertising. The model of hierarchy of effects was initially developed by attitude researchers in acknowledgement that individuals engender a series of responses in a certain order as effectively exposed to the advertising. This model started as early as in 1898, from “Attention- Interest- Desire- Action” (AIDA) model advanced by Elmo St. Lewis (Barry, 1987). Through AIDA model, Elmo St. Lewis proposed a systematic way of discussing criteria of effectiveness for the first time. It hypothesizes that attention, interest, desire, and action are the most important responses consumers might make to advertising with attention being the initial response and action being the last. 8 Many works had been endeavored to develop AIDA model. But most of them involved modifying the conceptual outlines or frameworks in a small way, based on intuition and logic. Not until 1961 had significant progress been made by Lavidge and Steiner to furnish the main body of modern hierarchy of effects literature (Barry, 1987). Lavidge and Steiner (1961) made the first attempt to identify advertising impact in terms of cognitive, affective, and conative categories of responses. Enclosed within those categories was a six-stage hierarchy that includes awareness, knowledge, liking, preference, conviction, intention, and purchase. As illustrated in Figure 2.1, Lavidge and Steiner (1961) postulated that cognition generally leads to affection which, in turn, leads to conation; consumers would inevitably go through a series of steps to that threshold of purchase. It was the first time that the criteria of effectiveness were linked to the areas of interest in the field of social psychology, thereby leading the literature in that large field to the issue of criteria in advertising and related field (Li & Leckenby, 2004; Weilbacher, 2001). Figure 2.1 The Traditional “Hierarchy of Effects” CAB Model Behavior Affection Cognition 9 In their work, “Cognition” represents the intellectual, mental or rational state, concerning the process or faculty of becoming specifically aware of a solution to fit one’s need, relating to the process of encoding, storing, processing, and retrieving information. It includes steps of perception, memory, thinking and understanding. More precisely, perceives stimuli (signals) is sensed by human beings and transformed into data in working memory, which will be compared with long term memory and manipulated by reasoning processes such as problem solving, planning, judgments until reaching a decision and executing a response (Toth, 2004). “Affection” refers to feeling or emotional states, for example, liking or disliking. “Conation” refers to the striving or behavioral states, meaning the aspect of mental processes or behavior directed toward action or change. It is closely associated with the concept of volition, defined as the use of will, or the freedom to make choices about what to do. The basic principles of the hierarchy went unchallenged until Palda (1966) posed his concerns over the lack of support of experimental evidences, which stirred a new developmental phase in the theory. This challenge was strongly reinforced by Ray's insightful suggestion that perhaps there were alternate orders to the hierarchy of effects (Ray, Sawyer, Rothschild, Heeler, Strong & Reed, 1973). Ray et al.’s research indicated that advertising did not always lead or cause people to change their attitudes and behavior towards a product; there is no specified sequence of stages which must occur as in Lavidge and Steiner’s (1961) view. Consumers may make decisions in a “non-rational” manner and could possibly “skip” stages. The process of consumer decision will not necessarily be linear or one-dimensional. As a result, the model should be provided feedback loops. Further, he proposed three orders of hierarchy of effects as a refinement 10 of the traditional model: the traditional or learning hierarchy (cognition-affect-conation), the dissonance-attribution hierarchy (conation-affect- cognition), and the low involvement hierarchy (conation-cognition-affect). He suggested that consumers might respond differently to advertising messages under certain circumstances. Audiences may follow the Learning Hierarchy model that they think and perceive, then feel or develop attitudes and then behave as the traditional CAB model describes. Consumers could also first behave, then develop attitudes and feelings as a result of that behavior, and then learn or process information that supports the earlier behavior, which is the Dissonance Hierarchy, the total reverse of the Learning Hierarchy. At the same time, the Low Involvement Hierarchy maintains that consumers behave, then learn as a result of that behavior and then develop attitudes as a result of the behavior and the learning. Ray et al. (1973) emphasized that all three orders of hierarchies are feasible and can be correct. Subsequent research on advertising effects led to the inception of integrative models. Vaughn’s (1980) FCB grid illustrated the adaptive nature of advertising effects. As shown in Figure 2.2, the grid features four effects sequences that vary according to level of involvement (high/low) and type of inclination (rational/emotional) of the consumer towards a product category. The FCB grid’s sequences are: cognition-affectconation (informative), affect-cognition-conation (affective), conation-cognition-affect (habitual), and conation-affect-cognition (satisfaction). Vaughn (1980) suggests that advertisers should use the grid as a guide in shaping advertising messages to ensure that a product’s advertising is based on the right tactic and appeal to an audience. 11 Figure 2.2 The FCB Grid Model Smith and Swinyard’s (1982) developed Information Integration Response Model (IIRM), with its major contribution to introduce belief type (higher and lower order beliefs) to the literature of advertising effects. Their findings suggest that in lower order belief product purchases (where trial is cheap and easy) advertising works by increasing consumer awareness and reinforcing previous consumer experiences with a product, whereas in higher belief order situations (where trial is costly and risky) advertising works simply as a source of information for the consumer. Similarly, in Deighton’s twostage model (1984) advertising effects adapt to different consumer and product contexts. Kreshel further (1984) contended that an emotional response to a stimulus such as an advertisement might consist of a physiological, affective and cognitive response occurring simultaneously. 12 Based on the earlier works, Li and leckenby (2004) suggest a Venn diagram, shown in Figure 2.3, to illustrate the non-linear and overlapping nature of the three criteria of the effectiveness. The new diagram expresses the mutually non-exclusive nature of the three criteria, in which there is no pre-determined starting point or ending stage. In addition, it is possible to have more than one criterion developing at the same time. Thus, a full view of the effectiveness of the advertising message requires being measured on all three dimensions. Figure 2.3 The Venn Diagram of Contemporary CAB Criteria Behavior Cognition Affection 2.2.2 Measures of Effectiveness of Online Advertising At present, the measures of online advertising effectiveness have not reached consensus. The question of what constitute the appropriate measures of effectiveness remains highly debatable (Wright-Isak, Fable & Horner, 1996). Quite a number of measures have been proposed to empirically evaluate advertising effectiveness, which generally fall into two categories: actual response and impressions (Danaher & Mullarkey, 2003). 13 Click-through rate was once the most widely and exclusively used measure for online advertising effectiveness (Forrester, 2002 & 2001), which is the percentage of the total number of ad exposures that induce a surfer to actually click on a banner in response to an advertised message (Novak & Hoffman, 1997). As a measure of actual response, Click-through rate has the advantage of a behavioral response that is easy to observe, and indicates an immediate interest in the brand being advertised (Berthon, Pitt &Watson, 1996; Briggs & Hollis, 1997). However, merely click-through fails to quantify the impact of advertising exposure on a consumer’s cognition and affection (Briggs & Hollis, 1997). Drèze and Hussherr (2000) also argued click-through rates not capture the full extent of an advertisement's effectiveness, as pre-attentive processing does not lead to immediate action. Furthermore, response measures are not objective as the outcome can depend on the advertising copy strategy (Danaher & Mullarkey, 2003). In sum, emphasizing clickthrough rates only will predicatively lead to the overlook of effects that occur before or after the clicking action (Chtourou & Chandon, 2000). Meanwhile, it is acknowledged that click-through rate has declined steadily from 5% in 1998 and seems to have stabilized at 0.5% (Doubleclick, 2003a). In response to these situations, other measures of online advertising effectiveness should be adopted. In academic research, the currently favored measures to assess changes of impressions, include awareness, Attitude towards the ad (Aad), Attitude towards the brand (Abrand), and purchase intention (PI) (Li & Leckenby, 2004; Wu, 1999). Methodologically, the vast majority of communication effects research in the past few decades has relied on these measures to examine the persuasive impact of advertisement exposure. 14 Awareness Awareness is one of the traditional measures on advertising effectiveness, which means consumers recognize the existence of a brand, service or idea. Awareness has been claimed to be the first step to assess the impact of advertising in many hierarchy of effects models. For example, as discussed in the previous session, AIDA model, suggests that awareness is affirmatively the initial response when consumers react to advertising. Colley (1961) also claimed that awareness represents the minimal level advertisers seek for advertising goals. Awareness includes advertising awareness and brand awareness. Advertising awareness refers to the recall or recognition of a specific advertisement. Research has provided evidence that there is relationship between advertising awareness and short-term sales (Hollis, 1994). Brand awareness is defined as a pyramid level of knowledge involving recognition, recall and top-of-mind awareness of the brand (Hoyer & Brown, 1990). Brand awareness has long been recognized as essential prerequisite for establishing brand images (Engel, Blackwell & Miniard, 1995) and the gateway leading to consumers’ purchase intention. Attitude towards ad (Aad) & Attitude towards brand (Abrand) Attitude is the most common measure in the hierarchy of effect. Engel et al. (1995) justified it to be simply an overall evaluation. As proposed by traditional hierarchy views, such as Lavidge and Steiner’s CAB model in 1961, attitude is deemed as linear sequences from cognition to affection and conation. A more contemporary perspective argued that attitude can be formed by either cognitive, and/or affective and/or cognitive dimensions, with conation being either a determinant or an outcome of attitude (Eagly & Chaiken, 1993; Li & Lekenby, 2004; Ray, 1973) 15 A great number of empirical studies have documented attitudinal shifts resulting from advertising exposure (e.g., Batra & Ray, 1986; Cacioppo & Petty, 1985; Ha, 1998; Kim, Jang & Lim, 2000; Lutz et al., 1983). Thus, the attitude towards the ad (Aad) is conceptualized as “a predisposition to respond in a favorable or unfavorable manner to a particular advertising stimulus during a particular exposure occasion” (MacKenzie, Lutz, & Belch, 1986, p. 130). That is, Aad represents an individual’s evaluation of the overall advertising stimulus. Some researchers (Batra & Ahtola, 1991; Olney, Holbrook & Batra, 1991) argued that the measure of attitude toward the ad should be considered multidimensionally with hedonism, utilitarianism and interestingness as Aad’s attitudinal components. Here, hedonism refers to the evaluation of entertainment value of the ad; utilitarianism refers to the evaluation of usefulness of the ad, and interestingness is viewed as an evaluation of curiosity. The attitude towards the ad (Abrand) measures the extent to which respondents have a positive or favorable opinion of the brand. Researchers and practitioners have been using multi-attribute models to study consumer attitude towards brand for two decades (Trout, 2005), which propose that individuals perceive brands as possessing a number of attributes that provide the basis on which consumers form their attitudes. Using this approach, an individual's overall attitude toward the brand is determined by a consumer's evaluative response or attitude toward brand attributes and a subjective estimation of the probability that the brand actually has the attribute (Pechmann & Stewart, 1989). Belch and Belch (2006) represented the attitude towards the brand as being influenced by consumers’ beliefs about specific brand attributes and different levels of importance attached to these attributes. 16 Purchase Intention Purchase intention (PI) is the most immediate step preceding actual behavior. It tends to measure the likelihood of respondents to taking a purchase action in the future (indicate will buy, or test-drive, for example). Many studies have reported Aad as a mediator of advertising effects on brand attitude and purchase intention (Aaker, Stayman & Hagerty, 1986; Lutz, 1985; MacKenzie, Lutz & Belch, 1986). According to MacKenzie et al. (1986), there can be a one-way or two-way flow between Aad and Abrand owing to different situations; both Aad and Abrand are independent determinants of purchase intention. As discussed above, the effectiveness of online advertising should be addressed in general context of the criteria of cognition, affection and behavior by measuring changes in awareness, brand perceptions, attitudes, and purchase intention. However, online advertising differs from traditional advertising with expanded capabilities. One significant aspect is that online advertising is endowed with the capability of interactive communication, which attributes more power to the users over controlling the communication processes that the users can not only be actively involved in, but also have a wide range of freedom and opportunities. In this sense, user’s control should be considered as an outstanding issue with respect to online advertising effectiveness. In response to that, some researchers suggested online advertising employ new measures with respect to user’s control. According to the Interactive Advertising Model (IAM) proposed by Rodger and Thorson (2000), items such as Internet uses, information processes, and the responses that result from the encounter of this processing were clarified as the new measures for consumer’s control. Li and Leckenby (2004) further modified the traditional measures by suggesting and specifying a list of measures that 17 might be under the control of either the user or the advertiser. The measures under control of advertiser are largely the standard measures, such as attention, attitude and intention; while the newly suggested measures are grouped under user’s control, include consumer goals, consumer expertise, consumer ideals, personalization effects, quality of decision, trust, internet motivation, and active participation. Notwithstanding the necessity of refinement of the general measures to meet the need of contemporary environment, the new measures with respect to user’s control will not be adopted in this study. Firstly, as for the current status, it is presumable that the new measures are put forward on the basis of taking into consideration of active online ads. As discussed above, active ads can only be reached by clicking the passive ads, which requires the actual responses from consumers. However, under the circumstance of the low 0.5% click-through rate, most banner ads on the Web are acknowledged as the passive form of non-interactive advertising. Thus, it is not appropriate to apply new measures of active ads to the current passive ads. Secondly, very few studies have suggested or applied new schemes for accessing online advertising effectiveness till now. As pioneer, the new measures are still in the immature stage and need more empirical validation. As such, these new measures will not be employed as criteria of online advertising effectiveness in this study. However, some items from new measures, such as consumer expertise and active participant, were practically measured in this study as controvertible factors for their potential effects on advertising effectiveness. To summarize, there is no one best way to measure online advertising effectiveness. Multiple measures should be employed to obtain more insight. Advertising awareness, attitude towards ad, attitude toward brand, and purchase intention are the most commonly 18 used measures for assessing online advertising effectiveness, which will be adopted in this study. 2.3 Executional Cues & Effectiveness Online advertising has been proved to be an effective tool for promoting products and brands by compelling evidences (Briggs, 2001; Dreze & Hussherr, 2003; Wakeling & Murphy, 2002). However, the effectiveness varies, owing to wide-ranged factors related to design, execution and context of online campaigns. A fairly large body of research in the online advertising suggests the considerable effect of the executional cues within the banner ad, such as textual content, image, banner size, animation, and incentives. Based on specific characteristics and functions, these executional cues of advertising can generally be subsumed and categorized into central and peripheral (Maclnnis, Rao & Weiss, 2002). The potential central cues include those that clearly demonstrate product superiority by providing a large number of arguments (Petty & Cacioppo, 1984a; Sewall & Sarel, 1986), focus on the product (Maclnnis & Stayman, 1992) and its appealing attributes and benefits to consumers (Maheswaran & Sternthal, 1990; Stewart & Furse, 1986), and those that indicate product difference in light of competitive offerings (Stewart & Furse, 1986). In a word, central cues refer to the rational executional cues that credibly demonstrate the benefits of the product or speak to the product’s quality as such allow consumer to engage in issue-relevant thinking and evaluation of the true product merits. The role of central cues is to encourage deeper message processing, that is, to move 19 individuals along the continuum of depth processing toward elaboration. It has been found that lasting persuasion would be produced if the receiver thinks, or rehearses, favorable thoughts about the message of a central cue; whereas, a boomerang effect (moving away from the advocated position) is likely to occur if the subject rehearses unfavorable thoughts about the product-related argument. The peripheral cues have been conceptualized in terms of the extent to which they are affectively based or operate as heuristic cues (Maclnnis, Rao & Weiss, 2002). Affective-peripheral cues refer to cues that can induce positive feelings in views and thus positively impact ads and brand attitudes, comprising likable sources (Petty & Cacioppo, 1981; Petty, Cacioppo & Schumann, 1983), dramas (Deighton et al., 1989), warm appeals (Asker, Stayman & Hagerty, 1986), visually appealing pictures (Mitchell & Olson, 1981; Stewart & Furse, 1986), and likable music (Bierly, McSweeney & Vannieuwkerk, 1985). Heuristic-peripheral cues are shortcuts that enable inferences about brand benefits or quality, including credible source (Craig & McCann, 1978; Sternthal et al., 1978), a source like themselves (Brock, 1965), an expertise source (Yalch & Yalch, 1984), relevant pictures (Mitchell & Olson, 1981), or relevant music. As discussed above, peripheral cues employ communication strategies such as trying to associate the advocated position with things the consumer already thinks positively towards, using an expert appeal, and attempting an inference effect where the advocated position could be achieved after several easily-processed persuasion cues, which the receiver could make reasoning about, have been presented. If the peripheral cue association or inference is accepted then there may be a temporary attitude change and possibly future elaboration. However, if the peripheral cue association is not accepted, 20 or if it is not present, then the person retains the attitude initially held. With the multimedia nature that the Web endows, online advertising expands the character of peripheral cues. Online advertisement can be externally paced and integrated with visual effects and movements, sounds and texts. Thus, it is believed that some popular executional cues of online ads, such as animation and image all fall into the categories of peripheral cues (Kalyanaraman & Oliver, 2001). Amination especially has been deemed as one of the most important peripheral cues of online advertising as only 91 insertions over the 1,258 studied were not animated (Chandon et al., 2003). In traditional advertising, in the last few decades, a shift has been made to presenting advertisements that are more focused on peripheral cues such as imagery, and less on clear reasons to action or benefits of buying the product. Many studies on consumer behavior also claim the superiority of peripheral cues over central cues, from two major perspectives. One approach focused on distinguishing the effects of peripheral cues in terms of enhancing consumers' awareness and brand recall. One of the most predominant findings is that information conveyed by peripheral cues, such as image and animation is more easily recalled and recognized than textual information. For example, Childers and Houston (1984) reported picture-superiority effect on memory. Finn (1988) found significant and positive relationships between image presence and comprehension in 3 cases out of 5 and memorization in 8 cases out of 12. Unnava and Burnkrant (1991a) also argued that image-based information facilitates the use of mental imagery, which is recreated in the mind of consumers after stimulus. He claimed graphic information can generate more mental codes than verbal information and as a result memorization can be improved. Leong, Huang and Stanners (1998) compared the effectiveness of the Website 21 with traditional media and found that information conveyed by image is more easily recalled and recognized than textual information. Yoo (2003) found Internet banner advertisings to be better suited to the pictorial type than the textual one. He reported that pictorial banners were rated as significantly superior to textual banners in recall, favorableness, credibility, purchase intention, and click-through rate. Animation, as one of the most important peripheral cues in online advertising, also has been studied extensively. A growing body of experimental research documents the psychological superiority of animated Web ads over static ones. Commercial studies usually reveal that animated banners catch the eyes better and thus generate more attention (IAB, 2001). In the same vein, academic researchers suggested that animation is a powerful tool for online ads to generate desirable advertising effects. For example, animated ads are able to elicit stronger orienting responses (Lang et al., 2002), better click-through rate or the intention to click (Cho, 1999), faster click-through (Li & Bukovac, 1999), higher arousal (Heo & Sundar, 2000a), and better memory for ad content (Heo & Sundar, 2000b; Lang et al., 2002; Li & Bukovac, 1999). For another approach, studies probe into the effects of peripheral cues on consumers' evaluation of product. Kisielius and Sternthal (1984) examined the effects of vividness on consumer’s attitudinal judgment. They detected that vividness of message inhibits positive judgment as it stimulates the cognitive elaboration of information and the favorableness of available information determines the persuasive effect of vividness. It was found that animated ads are also able to elicit more positive attitudes toward the ads. Kalyanaraman and Oliver (2001) devoted their attention to examining the effect of multiple peripheral cues on attitude change and found that the presence of animation leads to more positive evaluations of online ads. 22 Taken together, it has been argued that peripheral cues are effective to produce a desired result on effectiveness of online advertising. However, the peripheral cues may not always produce positive reaction to online advertisement. Yoo and Kim (2005) explored the effects of animation on memory and attitudes towards ads. The results showed inverted U-shaped relationships between the level of animation and both recognition rates and Aad, suggesting the existence of unintended negative effects of highly animated online banners. Under high-animation conditions, subjects experienced negatively valenced thoughts and unpleasant feelings, which negatively influenced Aad. Also, subjects were highly aroused, as indicated by the increased level of emotional intensity; this arousal inhibited subjects' ad recognition performance. Also, Gao et al. (2004), who studied online advertising effectiveness in terms of attitude forming, found that perceived irritation is positively related to the use of continuous animation and popups by e-commerce consumers, which demonstrates a significant negative correlation with attitude. This finding was validated in observations of several industry surveys (Johnson et al., 1999) and practitioners' advice (Nielsen, 2000). As to images used as peripheral cues, Donthu et al. (1993) showed that there is no significant relation between the proportion of space occupied by the image and the memorization for advertising. Chtourou and Chandon (2000) found that the effects of picture presence on the memorization and the intention to act were not linear. They argued that the use of picture in banner is not predicatively associated with better recall and higher purchase intention. Although not many studies focus on the central cues of online ad, literature has demonstrated the influential impact of argument contents on advertising effectiveness. According to the Elaboration Likelihood Model (Petty et al., 1983) and other similar 23 dual-processing models (i.e., Chaiken & Maheswaran, 1994), when messages are deemed to be personally relevant, systematic (central route) processing occurs, which leads to a greater cognitive elaboration of message arguments or advertising content (Petty & Wegener, 1998). In this situation, a well-written, informative ad can be very persuasive. More specifically, the information of the true merits of product presented in the ad will effectively induce highly involved people to have diligent consideration; the consumers tend to learn cognitively and comprehend the ad-delivered information at deeper, more elaborate level and therefore result in attitude’s change, such as product beliefs, positive brand attitudes, and purchase intention (Peter & Olson, 1996). It was also found by Ducoffe (1995) that consumers are more likely to positively evaluate banner ads with relevant information versus those with less relevant information, since they help to reduce search time and costs. Kim and Leckenby (2000) examined the effectiveness of creative factors in interactive advertising in respect of click-through rate among total of 243 actual banners. They categorized all the creative factors in two aspects: message factors and other executional factors. Product benefit approach was considered as one of the important message factors and 61.7% of the tested banners actually adopt this approach; while size, shape and animation of banner were included in other executional factors. Results showed that message factors are more highly related to click-through rates than are other executional factors. This is consistent with previous findings of traditional media. As proved by Stewart and Furse (1984), the message variables perform better than other executional variables at recall, comprehension and persuasion in the TV commercials situation. Holbrook and Lehmann’s study (1980) shows a similar finding that message factor can be more resultful than the other executional factors in terms of 24 advertising effectiveness. To sum up, the effects of both central and peripheral cues of banner have been distributed in considerable prior literature. Although many studies confirmed the superiority of peripheral cues over central cues on the effectiveness of online advertising in one way or another, results also proved the efficiency of central cues on eliciting positive attitudes towards the online ads. Using different executional cues in advertising is certainly not a new concept, however applying those to banner ads in the online environment might present a dissimilar way to connect with consumers by combining visual creativity with the new features of online advertisng, for example, relatively less intrusive nature and smaller sizes. Thus, it is necessary to empirically investigate the application and impact of different executional cues on advertising effectiveness in the context of banner advertising. Following the prediction about the superiority of peripheral cues by many literatures, the first hypothesis is phrased as: H1: In general, peripheral cues are more capable than central cues in producing a positive effect on the effectiveness of online advertising. 25 2.4. Involvement & Effectiveness Involvement has long been a topic of interest in the advertising literature. Its effects on consumers’ responses to advertising have been shown to be numerous and significant. Following Barki and Hartwick (1989) and Krugman (1967), involvement refers to a subjective psychological state of a consumer and defines the importance and personal relevance that consumers attach to an advertisement or product. However, an examination of potential sources of product involvement should serve to point out that the construct might exist in two different forms. One form is situational involvement, which refers to a temporary concern that consumers may experience with a product during its purchase when there are high stakes associated with the purchase outcome (Rothschild, 1979). The second form of involvement exists on a long-term basis and occurs on the strength of the product’s relationship to individual needs, values, or the self-concept. Houston and Rothschild (1978) termed this involvement variant enduring involvement. It is conceivable that during a purchase, it may be difficult to practically separate the portion of consumer arousal that constitutes situational involvement and that could be labeled enduring involvement. Considerable progress has been made by researchers in understanding the nature of product involvement and in identifying its attitudinal and behavioral correlates. It is generally assumed that, up to moderate levels, involvement facilitates advertising message processing. Greenwarld and Learvitt (1984) suggested a principle of higherlevel dominance. That is, the effects associated with the highest level of involvement on advertising message should be dominant. They argued that this principle of higher-level of dominance is plausible; as advertising message is analyzed in-depth at higher levels of 26 involvment and the effect associated with the higher levels tend to be stronger and longer-lasting. Furthermore, researchers have found positive correlation between involvement and specific aspects of advertising effectiveness. Various empirical studies have claimed the effect of involvement on recall of advertising message (Gardner, 1983; Gardner et al., 1985; Leigh & Menon, 1987; Nelson, Duncan, & Fronczak, 1985). One of the earliest studies in support of the positive recall effects hypothesis was that of Clancy and Kweskin (1971), who showed that involvement of the program had a significant positive impact on memory for advertising. Petty et al. (1983) indicated that highly involved individuals had higher levels of recall of brand name and product category. One recent study conducted by Galpin and Gullen (2000) also demonstrated the positive effects of involvement on memory for commercial content within a media vehicle. Meanwhile, several other studies examined the relationship between involvement and attitudes, brand perception. For example, it was found that consumer’s involvement positively influenced attitude toward the advertisement and attitudes toward the advertised brand (Murry, Lastovicka & Singh, 1992). Shamdasani, Stanaland and Tan (1997) also concluded that attitudes formed under high involvement conditions tend to decay less than attitudes formed under low involvement conditions. In addition, past research has shown that the more involvement a person has with a product, the more likely he/she will be to express higher purchase intent. This effect has been demonstrated in both off-line (Singh, Balasubramanian & Chakraborty, 2000) and on-line (Raney, Arpan, Pashupati & Brill, 2003; Yoo & Stout, 2000) media where the effects of infomercials and on-line mini-movies, respectively, were studied. However, 27 these studies have looked at involvement and purchase intent based upon a person watching a long message. In sum, dozens of literature convinced that consumer involvement effectively influences the outcome of advertisements effectiveness. The more involved a consumer is, the better will be the quality and the amount of responses. This leads to the second hypothesis: H2: The higher involvement level of the consumer will significantly enhance the effectiveness of online advertising. In prior studies, researchers have also found much merit in the potential of involvement as an explanatory variable for the hierarchy of effects and its moderating effect on advertising effectiveness. Krugman (1965) has been credited with initially proposing that consumers’ levels of involvement will dictate the manner in which they will process and respond to advertising information; different levels of involvement are associate with different sequences of impacts on the attitude components of affect, behavior and cognition. Krugman thus suggested that, under low involvement situations, consumers passively and thus randomly gather information, as opposed to being active information processors and seekers when highly involved. Krugman also argued that, with high involvement, a communication should act most directly to modify beliefs (that is, verbalizable propositions); by contrast, with low involvement the impact should be more on perceptions (that is, sensory organizations, such as brand logos or package configurations) and should occur more gradually, being effective only with repeated exposures. Krugman’s suggestion was further developed by Ray et al. (1973) and 28 reinforced by Vaughn’s (1980). Ray et al. (1973) argued under high involvement, a communication is likely to affect cognitions, then attitudes, and then behaviors; whereas, under low involvement, a communication is more likely to affect cognitions, then behaviors, then attitudes. Vaughn (1980) also stated that consumers acting in their own self-interest would calculate consumption decisions. They would process information, develop attitudes, and then behave accordingly. In essence, involvement has been deemed as an import mediating variable in the effectiveness of adverting executions and the sequential nature of the three main phases of the hierarchy of effects: cognition, affection and conation. This view is later thoroughly described by Petty and Cacioppo (1983) through the Elaboration Likelihood Model (ELM). The ELM is based on the idea that attitudes are crucial as attitudes guide individuals’ decisions and other behaviors. While attitude changes may result from a number of factors, persuasion is a primary source. The ELM thus tries to address the process by which persuasive communications (such as ads) lead to persuasion by influencing attitudes and the differences in the ways consumers process and respond to persuasive message. According to this model, the attitude formation or change process depends on the amount and nature of elaboration, or processing, of relevant information that occurs in response to a persuasive message. High elaboration means the receiver engages in careful consideration, thinking, and evaluation of the information or arguments contained in the message; while low elaboration occurs when the receiver rather makes inferences about the position being advocated in the message on the basis of simple positive or negative cues. The ELM shows that elaboration likelihood is a function of two elements, motivation and ability to process the message. Motivation depends on 29 such factors as involvement, personal relevance, and individual needs and arousal; thereinto, involvement is the key variable. Ability relies on the individual’s knowledge, intellectual capacity and opportunity to process the message. The ELM postulates two basic routes to persuasion or attitude change. Under the central route, the receiver is viewed as an active, involved participant in the communication process whose motivation and ability to attend are high. When central route occurs, the consumer pays close attention to message content and scrutinizes the message arguments. This diligence is likely to result in a high level of cognitive response activity, and predominantly favorable cognitive responses may lead to favorable attitude change in product beliefs, brand attitudes, and purchase intention. This attitude is relatively enduring and should resist subsequent efforts to change it. The peripheral route portrays quite a different explanation of the persuasion process. Under the peripheral route to persuasion, the receiver is viewed as lacking the motivation or ability to process information. Rather than engaging in detailed cognitive processing and evaluating the message, the receiver relies on peripheral cues that may be incidental to the main argument. Favorable attitudes may be formed if certain executional aspects of peripheral cues are viewed as likeable, or have positive associations or heuristics. Attitude resulting form peripheral processing is temporary. In application of the ELM with respect to involvement, it is presumable that when consumers have a high level of involvement they are motivated to pay attention to the central, product-related information, such as product attributes and benefits or demonstrations of positive functional or psychological consequences; they are willing or able to exert a lot of communication processing effort and tend to learn cognitively and 30 comprehend the advertising message at deeper, more elaborate level, which may lead to enduring positive attitude change. Conversely, if consumers are low involved, they are willing to put effort on communication processing. They may also change attitudes, not as a result of committed cognitive activity, but rather because of the positive or negative cues associated with the issue or product in the advertisement, such as a picture or color of ad, or a celebrity endorser, for the entertainment value of the advertisement. As such, involvement is viewed as the hinge if the message is processed under central or peripheral route. In applying executional cues or other features to banner advertisement, it is important to take into account different levels of involvement. As explained by ELM, when consumer involvement decreases, the importance of peripheral cues to persuasion increases; higher the involvement, central cues more effective. Empirical findings support these assumptions. For instance, prior research has shown that peripheral cues such as source reputation exert an important influence on consumers' responses to low-involvement products (Greenwald & Leavitt, 1984; Petty, Cacioppo & Schumann, 1983). Wen (2000) examined the impact of textual-based banner and visualbased banner on advertising effectiveness in condition of different level of involvement. She found that textual messages are more likely to be persuasive under high involvement, whereas visual messages affect consumer responses under low involvement situation. Yoon (2003) also tried to determine the sources of effectiveness of banner advertising in terms of ad type and content by incorporating the theoretical premises developed on the role of the consumer's level of involvement. It was found that level of involvement has a significant impact on the effectiveness of banners, which shows higher advertising preference for advert type and advert content when consumers were highly involved. 31 To summarize, involvement works as a moderator between executional cues and the effectiveness of online advertising. In application of executional cues to banner advertising in the context of existing different levels of product involvement, it is suggested that central cues should be the primary influence of advertising effectiveness for highly involved consumers. In other words, in the case of a high-involvement product, banner ad effectiveness is expected to be central-cue-driven. Following the predictions of the ELM, peripheral cues will have a favorable impact for low-involved consumers; under this situation, consumer evaluations will hinge on the peripheral cues, such as image and animation. This rationale generates the following hypotheses: H3a: A peripheral-cue-based banner is more effective than a central-cuebased banner when consumers are less involved. H3b: A central-cue-based banner is more effective than a peripheral-cuebased banner when consumers are highly involved. 2.5 Interactivity Interactivity is considered as the key feature of the medium Internet (Ha & James, 1998; Hoffman & Novak, 1996; Steuer, 1992); thus, it accordingly becomes an important characteristic of online advertising. The concept of interactivity, had actually appeared early in traditional media research. As the Internet evolves dramatically, it has been revivified and endowed a more complex nature, which makes it remain “an overused, under-defined concept” (Heeter, 2000). Researchers have viewed interactivity from different angels. Thus, multiple definitions exist, but almost all definitions emphasize the importance of interaction 32 between user and system. For example, Rafaeli (1988) defined interactivity in terms of the responsiveness of participants and the degree to which a communication process resembles human discourse. He also referred interactivity to the ability to select the timing, content, and sequence of a communication act within a mediated environment. Steuer (1992), in his work on virtual reality, also addresses the notion of interactivity. He claimed that interactivity and vividness contribute to the so-called "telepresence" experience by which one feels present in the mediated environment. His definition of interactivity identifies three factors: speed, range, and mapping. Among the three factors, range refers to the number of possible actions at a given time; mapping refers to the ability of a system to map its controls to changes in the computer-mediated environment in a natural and predictable manner. Thus he defined interactivity as “the extent to which users can participate in modifying the form and content of a mediated environment in real time” (P84). Liu and Shrum (2000) specified three dimensions of interactivity: active control, which is characterized by voluntary and instrumental action; two-way communication, referring to the ability for reciprocal communication between two parties and enabling transactions directly online; and synchronicity, that is, the degree to which users' input into and the received response from the communication are simultaneous. In conclusion, although definitions of interactivity differ across previous studies, interactivity should be based on consumers’ actual control over information and communication flow. That is, interactivity should allow consumers to seek and gain access to the information on demand by directly controlling the content and sequence of communication (Chung & Zhao, 2004). Even though interactivity are largely studied with Web sites and computers, 33 interactivity of online advertising have not yet been clearly defined. This is partly because online ads mostly are not really interactive. As a matter of fact, most online advertising has still followed the traditional media of placing ads in strategic locations in order to catch user attention. The most that banners do by way of interactivity is to attract users to click on them so that the users can be transported to the advertiser’s site. With the click-through rates as low as 0.5%, it is perceivable that the concept of interactivity of online advertising is more referred to the Web medium but not advertisements themselves (McMillan & Hwang, 2002). As suggested by Sundar and Kim (2005), an online advertisement that captures the essence of interactivity nature should have multiple layers of navigable content, accessible via clickable tabs, rollovers, and hyperlinks embedded in the ads themselves. Most importantly, clicking on these ads does not simply lead the user to another site but instead refreshes content within the ad space, keeping the rest of the Web page constantly on the screen. In this study, the experimental banners were manipulated in a similar way as suggested above, focusing on not only the effects of vividness, but increasing levels of dynamism, and successful attempts to induce experiential encounters, so that consumers could have different degrees of gaining or accessing banner information as they want. Although literature search yielded little empirical investigations of levels of interactivity in online advertisements for experimentally determining their effects, people generally presume interactivity as a beneficial attribute to advertising effectiveness, in reference of the findings of other online facilities. Some studies confirmed the positive impact of interactivity on consumer response such as attitude toward the advertisement, as these studies found higher levels of vividness and dynamism to produce potent 34 experiential encounters between consumers and the online environment. For example, Hoffman and Novak (1996) argued, from a consumer behavior perspective, that interactivity can help create a "flow" experience that would make consumers more involved and more alert to website contents. Huang (2003) also determined that increasing levels of interactivity make for higher levels of experiential flow among visitors. Flow has been proposed by Hoffman and Novak (1996) as essential to understanding consumer navigation behavior in online environments. It is defined as the state occurring during network navigation, which is: 1) characterized by a seamless sequence of responses facilitated by machine interactivity, 2) intrinsically enjoyable, 3) accompanied by a loss of self-consciousness, and 4) self-reinforcing. In a similar vein, banner advertisements can construct “flow” experience by adopting interactive features, such as multiple layers of navigable content accessible in the banner ads themselves. According to Hoffman and Novak, the experience of flow facilitated by interactivity in the process of network navigation has a number of positive consequences, including increased consumer learning, exploratory behavior, and positive affect. There were other empirical findings to reinforce the effect of this flow experience of interactivity on consumers’ attitude changes in one way or the other. A content analysis conducted by Ghose and Dou (1998) also showed that the levels of interactivity are directly related to a website's attractiveness and ranking. Parsons, Zeisser and Waitman (1998) also pointed out that interactivity is vital in engaging users' interest and participation, as well as retaining them. Likewise, Wu (1999) measured participants' attitudes toward and their perceived interactivity of particular websites. Results showed a strong correlation between the two. Similarly, Cho and Leckenby (1997) measured 35 participant’s intention to interact with a target (banner) ad and found positive relations between intention to interact with the ad and attitudes toward the ad, attitudes toward the brand, and purchase intention. Yoo and Stout (2001) also observed this general pattern of results. Taking the above into consideration, it can be posited that interactive features of banner advertising will induce experiential flow within consumers online and thus result in positive attitude changes towards the advertising message (Coyle & Thorson, 2001). As such, another hypothesis is advanced: H4: In general, banner advertisements with interactive features will be more effective than non-interactive banner advertisements. In a similar way, the moderating effect of involvement between interactivity and advertising effectiveness has been suggested in prior studies. It is found that people in high involvement situations with high need for cognition are more likely to engage in active cognitive processing by interacting with communication messages; the higher the involvement, the higher the need for cognition, and the more desirable the interactivity. When Cho and Leckenby (1999) tried to position interactivity as a measure of advertising effectiveness, they found that high personal and product involvement were more likely to have a higher intention for interactivity. That is, when people are highly involved, interactivity can be a positive predictor for advertising effectiveness. Cho (1999) also pointed out the level of personal and product involvement be positively related to clicking of banner ads, i.e., the higher involvement is, the more clicking of banner is. Moreover, McMillan (2000) recorded a strong, positive correlation between the user's 36 involvement with the subject of the site and perceptions of interactivity. Yoo and Stout (2001) also considered involvement as one of the important factors affecting users' Interactivity with the Web Site. They reported that higher product involvement is leading to a higher level of user interactivity. Thus, generally speaking, in this situation of high involvement, consumers are willing or able to exert more cognitive processing effort and thus are more likely to demand greater information to satisfy their intrinsic need for information and cognition; that is, they are more likely to request and search for more information they are involved with by clicking related tabs, buttons, rollovers, and hyperlinks given in the online banner ad, which might positively result in attitudes change or purchase intention. In contrast to a high-involvement situation, consumers in low-involvement situations have lower motivation to process the advertising message. Therefore, they are less likely to request further information provided by the interactive features of online ad. That is, they are less likely to click related buttons or hyperlinks in the Web ad than highly involved consumers, which suggests the interactive features result in no impact on banner effectiveness. Furthermore, interactive aspects of online advertising may inhibit consumers' information processing under the condition of low involvement. By incorporating interactive features, the application of some other effective characteristics for low involved consumers, such as image or animation, would inevitably be hindered to a certain degree as the banner size is restricted. At the same time, the complexity of interactive features would make low involved consumers feel a sort of "lostness" and could not focus on messages that banners want to convey; thus the effectiveness of advertising could also be swayed. For example, Bezjian-Avery, Calder, and Iacobucci (1999) argued that interactivity actually had a detrimental effect as users in 37 the interactive condition viewed the ads less and indicated less purchase intention than did those in the linear condition. In addition, Anderson (1996) found that inadequate levels of interactivity might instill boredom and make consumers terminate their navigation sessions prematurely. As low involved consumers would not have the motivation to spend a great deal of time trying to figure out how the interactive features work, interactive features may result in consumers’ negative attitude change towards the ad. As such, it is predictable that the interactive features of banner would positively affect advertising effectiveness for highly involved consumers; in contrast, the banner without interactive features would be more likely to result in advertising effectiveness under a low-involved situation. Thus, the following hypotheses are set forth: H5a: A non-interactive banner is more effective than an interactive banner when consumers are less involved. H5b: An interactive banner is more effective than a non-interactive banner without interactive features when consumers are highly involved. 38 Chapter 3 Methodology 3.1 Research Design This study focuses on banner advertising on the Web and seeks to examine if consumers’ attitudes and behaviors are effectively influenced by different features of banners under diverse conditions. This study is designed as an experiment since experimental design is seen as one of the most rigorous research designs to illuminate causal inference for social science research (Shadish, Cook & Campbell, 2002). When applied to advertising research, experiments are claimed to be able to efficiently isolate the impact of the specific advertising campaign on people’s impression of the brand while controlling any effects from other advertising (Romeo & Nyhan, 2002). A research design of 2 (Interactive / Non-interactive) × 2 (Central / Peripheral) × 5 (Presentation Order) × 3 (Repetition) was used, where repetition is a within-subject factor and the others are between-subjects factor. Although presentation order is a betweensubjects factor, it was used to eliminate order effect and not used for analysis. In this study, subjects were randomly assigned to one of four experimental groups, making all the conditions (e.g., brand loyalty, past experience with computer and the Internet, possible exposure of ads, and effects of offline ads…etc) exactly the same except for different treatments. By executing a study in a carefully controlled setting, reliable data could be provided. 39 3.2 Participants One hundred and ninety one undergraduate students taking modules from the Communications and New Media Programme at the National University of Singapore participated in this study, for extra credit awarded to their final grades and lucky draw opportunities to win shopping vouchers, movie tickets, and electronic calculators. Participants’ age ranged between 18 and 26(M = 20.5, SD = 1.43). They had an overwhelming majority majoring in social sciences (98%) while the rest was either majoring in computer science or information systems. There were more female (68%) than male (32%) participants in this study. 3.3 Stimuli Based on the between-subjects factors, Interactive / Non-interactive and Central / Peripheral, subjects were randomly assigned to one of four conditions. Each condition had subjects exposed to three different banner advertisements. Thus, a total of twelve banner advertisements were created to represent two between-subjects factors and one within-subject factor. These banner ads were custom-tailored for this study by a specialist. Although previous experimental studies suggest that actual ads are better than mock ads in deriving more natural responses from the subjects (Mitchell, 1986), it is almost not possible to have three different brands whose banner advertisements have all four different versions that could represent the two between-subjects factors specified in this study. However, these banner ads were embedded into real Web pages that did exist on the World Wide Web at the time of study, in order to make subjects’ experiences closer to real-life 40 settings during the experiment. For repetition, three ads were created to represent three different categories of products/services. These categories were selected from those that frequently appear online. One was a sliming capsule (diet pill) named Shape-fast, another was an automobile named Vela, and the last one was an online dating service DreamDate. All three brands were made up by the designer, shunning the possibility of preconceived preferences towards any particular brands that affect the results by using real brands. For each brand, four different versions of banner ads were created. The first version was equipped with peripheral cues and interactive features; second version had peripheral cues without interactive features; the third version had no peripheral cues but employed interactive features; and the fourth version did not have peripheral cues or interactive features. As discussed above, this study adopted large images with animation that can serve as peripheral cues. The banner ads without peripheral cues were designed to emphasize information about product/service features and/or specifications to encourage information to be processed through the central route. The criterion of “interactive features” was adopted from McMillan (2000). McMillan (2000) introduces thirteen features of Web sites indicative of high interactivity, which includes search engines, survey/comment forms, registration forms, e-mail links, chat rooms and games. A search engine was chosen to be incorporated into the custom-designed banners for this study as this feature can be easily found in real-life banner ads, while other features introduced in McMillan (2000) are extremely rare or do not exist. Therefore, the banner ads with interactive features in this study had text boxes and buttons, which are the distinctive features of a typical search engine, while the banners without interactive 41 features did not have any such elements. Thus, four different types of banners were developed for each of three brands. As shown in Figure 3.1, type 1 ads are banner ads largely featuring on peripheral-cues and buttons for a search engine, type 2 ads are banner ads with peripheral cues without any buttons for a search engine, type 3 ads are banner ads that focuses on providing product/service-related information with buttons for a search engine, and type 4 ads are banner ads that focuses on providing product/service-related information and does not have any buttons for a search engine. A detailed description of experimental stimuli for each experimental group is shown in Figure 3.2. Figure 3.1 Research Design for Banner Type Manipulation PeripheralCue-Based Central-CueBased With Interactive Features Without Interactive Features Type 1 Type 2 Type 3 Type 4 42 Figure 3.2 Experimental Stimuli for Each Experimental Group Group 1 Group 2 DreamDate with peripheral cues & interactive feature DreamDate with peripheral cues & without interactive feature Vela with peripheral cues & interactive feature Vela with peripheral cues & without interactive feature Shapefast with peripheral cues & interactive feature Shapefast with peripheral cues & without interactive feature Group 3 Group 4 DreamDate with central cues & interactive feature DreamDate with interactive feature Vela with central cues & interactive feature Vela with central cues & without interactive feature Shapefast with central cues & interactive feature Shapefast with central cues & without interactive feature central cues & without All banners, regardless of the brands or features, were created to have only one size, which is called Skyscraper (120 × 600 pixels) by the online advertising industry. This specific size was picked because it is one of the most frequently used online advertising formats with 87 percent preference from the top advertisers (AdRelevance, 2000). Manipulation check was subsequently performed to validate the context congruency of the ads designed. Twenty subjects (undergraduate students from the National University of Singapore) were recruited through convenient sampling and requested to evaluate those ads on a 7-point scale for the following three criteria: peripheral, informative, interactivity. If a certain ad had not been sorted into the one and only corresponding category as it should, it would be excluded. Web pages where the banner ads were embedded used the layout and content of the 43 scientific news Web site Nature.com (http://www.nature.com), as the layout and articles were rather neutral and less likely to vary in the amount of interests a typical university student, like the subjects participating in this study, might have. The Web pages are consisted of three parts: a heading bar of the Web site, a news article of around 1,000words in length, and a section for other information (such as the highlighted news, list of the ‘most reviewed’ news, and the news-searching engine). Industry reports were selected for the content, since they tend to be least arousing for college students who are the subjects and unlikely to interact much with the processing of the banner ads presented with them. Figure 3.3 gives a full view of the sample Web page. As shown, experimental banner ads were positioned vertically on the left side of Web page. The Web page was designed to maintain one size that users do not need to scroll up and down to read the whole article. This was done to keep the whole banner ad to stay on screen while subjects were reading the Web pages. Figure 3.3 Sample Web Page and Banner Ad 44 To sum up, participants were randomized in four experimental groups, being exposed to three different ads of the same type. The presentation order of those three ads was also carefully randomized, which is shown in Figure 3.4. Figure 3.4 Presentation Orders of Experimental Banners Presentation Order Presentation Content Order 1 Order 2 Order 3 Order 4 VelaN1/ShapefastN2/DreamDateN3 ShapefastN1/VelaN2/DreamDateN3 DreamDateN1/ShapefastN2/VelaN3 ShapefastNI/DreamDateN2/VelaN3 Order 5 DreamDateN1/VelaN2/Shapefast/N3 * N1, N2, and N3 are referred to the three different news stories. 3.4 Pretest Using a convenient sample (undergraduate students from the National University of Singapore), 10 people participated in the pilot survey. The purpose of the pretest was two-fold: first and the most important, to verify if the experimental procedure was well understood and smoothly performed, and questionnaires were clearly written. Since the participants need to view three different Web pages in a certain order, the pretest tested the whole process and evaluated it by the criteria of perspicuity and clarity. Open-ended questions were included in the pilot test, the answers of which later helped to modify the final questionnaire with some new items added; secondly, to seek comments on the wording and arrangement of the questionnaire for the sake of further refinement. Face-toface interviews were made to collect their comments about the survey after the pretest. Based on the comments, a few questions were reworded to improve comprehension and eliminate confusion. 45 3.5 Procedure The experiment was an off-line experiment with forced exposure manipulation. Participants were told that they would be taking part in a study investigating user’s perception of Web sites design, which was to rouse moderate online activities related to Web surfing. All twenty experimental sessions were conducted in a computer lab of the National University of Singapore, equipped with twenty-four computers that have the same 15-inch LCD monitors. The total number of participants for each session was controlled not to exceed fifteen, in order to ensure sufficient interspaces and to avoid interference and interaction among subjects. Over one week of time, twenty experiments were accomplished, with details listed in Figure 3.5. Figure 3.5 Experiment Slots and Participants Details Group 1 Group 3 Session by Date Sep 15th, 2006 Sep 19th, 2006 Sep 20th, 2006 Sep 21st, 2006 Pretn Order 1 2 3 4 Subject No. 11 12 11 6 Sep 22nd, 2006 5 Session by Date Sep 18th, 2006 Sep 19th, 2006 Sep 20th, 2006 Sep 21st, 2006 Sep 22nd, 2006 Pretn Order 1 2 3 4 5 5 Total: 45 Subject No. 12 11 11 12 4 Total: 50 Group 2 Group 4 Session by Date Sep 18th, 2006 Sep 19th, 2006 Sep 20th, 2006 Sep 21st 2006 Pretn Order 1 2 3 4 Subject No. 11 14 9 9 Sep 22nd, 2006 5 Session by Date Sep 18th, 2006 Sep 19th, 2006 Sep 20th, 2006 Sep 21st, 2006 Sep 22nd, 2006 Pretn Order 1 2 3 4 5 4 Total: 47 Subject No. 11 10 9 14 5 Total: 49 46 During each session, participants were first shown an index Web page with procedural instructions and three hyperlinks to the three experimental Web pages (See Appendix III). Participants were supervised to browse those Web pages strictly by the certain orders indicated. They were also instructed to spend enough time looking at each Web page before going to the next one. However, the actions of skipping and backtracking Web pages were not allowed; clicking any hyperlinks on those experimental Web pages were also prohibited, which was to eliminate interferences that may influence the original recognition and attitude of the participants. Participants were asked to fill up a set of corresponding questionnaires right after exposure. To avoid ceiling effects during memory test, the questionnaire began with a distractor task, by asking for participants’ evaluation of the Web page design, Net-surfing habits and demographic questions. It was followed by memory test, which was conducted in the order of free recall, cued recall, and recognition tests. After finishing the memory test, subjects were asked to answer questions about their attitude and behavioral responses toward the banner ads. The questionnaire for this part was provided only after the memory test was completed, since it included images of banner ads that, if exposed prior to, will influence the memory test. The whole questionnaire consists of nearly 200 questions; only the questionnaires that have been completed and submitted on the spot were accepted as valid. 47 3.6 Measures 3.6.1 Involvement Involvement has been conceptualized as being two types: situational and enduring. Situational involvement is transient and occurs only in context of a situation such as purchase time (Richins & Bloch, 1986), or a temporary price reduction (Celsi & Olson, 1988). In contrast, enduring product involvement is considered to be a stable phenomenon as it represents the consumer’s personal interest in the product over long periods of time (Laurent & Kapfere, 1985). In this study, the dimension of situational involvement was not taken into consideration since the experimental ads did not give the incentive of price reduction and a purchase was unlikely to happen as the tested brands do not exist. Thus only the dimension of enduring product involvement is indexed by employing Zaichowsky Personal Involvement Inventory (PII) (Zaichkowsky, 1994), which consists of a 10-item bipolar scale with the following anchor points: important/unimportant, relevant/irrelevant, means a lot to me/means nothing, valuable/worthless, interesting/boring, exciting/unexciting, appealing/unappealing, fascinating/mundane, needed/unneeded, involving/uninvolving, importance, means a lot, and valuable are the significance attached to the ad. In PII, Relevance and needed is something related to what someone is seeking. Interest is the intellectual or emotional involvement with the advertisement. Exciting, appealing, and fascinating are the pleasing or enjoyable emotion that is evoked by advertisement. 48 3.6.2 Advertising Effectiveness As mentioned above, advertising awareness, attitude towards ad (Aad), attitude towards brand (Abr) and purchase intention (PI) would be used for assessing online advertising effectiveness. Advertising Awareness Advertising awareness was first measured by asking participants to recall the brand names of as many ads as they could without any hints. This is called free recall or unaided recall, which is a substantially more difficult task than recognition, and is associated with strong brand position (Unnava & Burnkrant, 1991b). Cued recall was proposed whereafter to measure advertising awareness, as a reminder may be necessary for recall for some consumers (Wen, 2000). Thus, subjects were asked to name the brands with cues of product categories. Subjects were given the choice of “Unaware” if they could not recall anything about the advertisements. Recognition was then used to further measure awareness. Subjects were asked to choose the right ad recognition items (i.e., "Have you even read exactly the following sentences in the advertisements: . . .") from two multiple-choice, closed-ended response options. Recognition items were designed to include both textual recognition and image recognition questions. Attitude towards the ad Olney, Holbrook, and Batra’s (1991) three-item indices of Hedonism, Utilitarianism, and Interestingness were used to measure consumer’s attitudes toward the ad. The 12-point semantic differential items are listed as follows: fun to see/not fun to see, pleasant/unpleasant, entertaining/not entertaining, enjoyable/not enjoyable, important/not important, helpful/not helpful, informative/not informative, 49 useful/not useful, make me curious/does not make me curious, boring/not boring, interesting/not interesting, and keep my attention/does not keep my attention. Attitude towards the brand Adapted from Mackenzie and Lutz (1989), a 4-item scale to measure consumer’s attitudes toward the brand was devised, which includes the following anchors: favorable/unfavorable, like/dislike, poor quality/high quality, and appealing/unappealing. Purchase Intention Purchase intention toward products featured in each of the four ads was measured via a 2-item seven-point scale (7 = very likely, 1 = not at all likely) adapted from Raman (1996). Another item adapted from “the 1997 IAB Study” proceeded to measure consumer’s likelihood to buy. 3.6.3 Other Covariates To account for the potential effects of uncontrolled factors that might impact on advertising recall, a number of covariates were also measured. These included the average amount of time spent on the web each day (measured on a 6-point interval scale, ranging from “five hours and more” to “less than one hour”), and their proficiency on the Web (measured on a 5-point "Not at all" to "Very" scale). 5-point and 9-point scales were also used to evaluate the layout of experimental Web pages and the news stories. Full details are given in the Appendix I. 50 Chapter 4 Results 4.1 Sample Size and Composition The experiment was conducted from 15 September 2006 to 22 September 2006, with a total of 200 subjects participating. Of the 200 questionnaires submitted, 191 cases were usable, while the rest of 9 cases were discarded due to invalid answers. The sample showed a male/female split of 62/129. Though the female subjects are significantly more than male subjects, the gender proportion is not a rare one in social sciences modules at the National University of Singapore. The sample was in the18-26 age group, with a mean age of 20.5. More than 80 percent accessed the Web for over 3 hour per day. Nearly 94 percent had Web experience more than 7 years and only 1.6 percent had used computer for less than 5 years. 68.6 percent have reported themselves familiar with the Web and almost 60 percent have claimed their proficiency on Web. The number of participants for each experimental group is shown in Table 4.1. A comparison in gender, age and time online composition of the resulted experimental groups is also made and described. A series of chi-square tests showed that there was no significant difference among all four experimental groups for their demographic data and Web surfing characteristics. 51 Table 4.1 The Comparison of Four Experimental Groups 4.2 Descriptive Statistics & Scale Reliability In this study, involvement is the most important construct. Zaichowsky Personal Involvement Inventory (PII) (Zaichkowsky, 1994) was borrowed as measurement. PII has seldom been applied to advertisements in the context of Web advertising; therefore, item internal consistency for the involvement scales was evaluated using the Cronbach's alpha reliability procedure (Cronbach, 1951). The scales of dependent variables used in this study, namely Attitude towards ad (Aad), Attitude towards brand (Abr), Purchase Intention (PI), were also tested and optimized for internal consistency, in order to ensure their validation in the online environment. Table 4.2 displays descriptive statistics and the results of the Cronbach's alpha tests. 52 All variables, dependent or covariates, have Cronbach's alphas satisfactorily ranging from .764 to .793. Table 4.2 Descriptive statistics for scales used in the experiment 4.3 Executional Cues & Effectiveness In order to further investigate the differences between the experimental banners in terms of two advertising characteristics (i.e. central cues vs. peripheral cues and interactive feature v s. without interactive feature, the four types of experimental banners) were rearranged. As shown in Figure 4.1, all peripheral-cue-based banners (types 1 and 2) were regrouped as Type Peripheral and central-cue-based banners (types 3 and 4) were regrouped as Type Central. In the same manner, all the banners with interactive features (types 1 and 3) were regrouped as Type Interactive, while banners without interactive features (types 2 and 4) were regrouped as Type Non-interactive. 53 Figure 4.1 Illustration of Banner Regrouping Type Peripheral With Interactive Features Without Interactive Features Peripheral Cue - Based Type 1 Type 2 Central Cue - Based Type 3 With Interactive Features Without Interactive Features Peripheral Cue -Based Type 1 Type 2 Central Cue -Based Type 3 Type Interactive Type Non interactive With Interactive Features Without Interactive Features Peripheral Cue -Based Type 1 Type 2 Central Cue -Based Type 3 Type 4 Type 4 With Interactive Features Without Interactive Features Peripheral Cue -Based Type 1 Type 2 Central Cue -Based Type 3 Type 4 Type 4 Type Central H1 hypothesizes that in general, peripheral cues are more capable than central cues on producing a positive effect on the effectiveness of online advertising. To verify whether the two experimental groups, Type Central and Type Peripheral, had any significant differences in advertisement recall or attitude change, a series of t-test was performed. Three different analyses were conducted based on three different product categories used in the experiment (automobile Vela, slim capsule Shape-fast, and online dating service DreamDate). Table 4.3 shows the mean of all the four evaluation scales 54 (advertising awareness, attitude towards ad, attitude towards brand and purchase intention) for Vela, Shape-fast and DreamDate. As the manipulation of non-existing brand names hindered subjects’ responses towards free recall and aid-recall and thus resulted in vein, the evaluation of banner awareness would only consider text and image recognition of experimental banners. On average, the mean score of peripheral banners is not significantly higher than that of central banners on all three products. Although subjects under Type Peripheral and Type Central indicate a significant different attitude change towards Vela, the results are not consistent. That is, for the automobile brand Vela, central cues showed a more positive effect than peripheral cues on attitude towards ad did (M=38.98 vs. M=33.17); however, peripheral cues were more effective than central cues in terms of attitude towards brand (M=10.63 vs. M= 9.62). Table 4.3 Results of Advertising Exposure for Three Product Categories 55 Table 4.4 sums the effects of advertising exposure on three product categories and presents a full view of t-test result. The mean categorized by two groups indicates a nonsignificant presentation format effect. That is, there is no statistically significant difference found to prove peripheral cues more effective than central cues on enhancing online advertising effectiveness. Thus, H1 was not supported. Table 4.4 T-test Results for Effects of Execustional Cues on Effectiveness * The results for awareness in term of free recall and cued recall are not shown, as they were not significant at all. 4.3 Level of Involvement & Effectiveness H2 predicts that the higher involvement will enhance the effectiveness of online advertising. Thus the impact of level of involvement on banner effectiveness was examined. In order to measure the level of involvement, 10-question items from Zaichowsky’s Personal Involvement Inventory (PII) (1994) were adopted and the responses to these questions were summed for each experimental banner. 56 Correlation analysis was then conducted to see if there is any significant relationship between level of involvement and banner effectiveness. It was found that advertising involvement had a consistently salient effect on advertising effectiveness in terms of Aad, Abr and PI, in all three different banners (all were p[...]... advertising 7 2.2 Conceptualization of Online Advertising Effectiveness 2.2.1 Criteria of Effectiveness Despite its expanded functions, the issue of criteria of online advertising effectiveness is still part of the broader question of advertising effectiveness in general Thus, the effectiveness of online advertising should be examined in a similar way as that of traditional advertising (Li & Leckenby, 2004;... reports of Interactive Advertising Bureau, banners have been one of the most common formats of online advertising all along since 1999, accounting for over half of the online advertising dollars spent Thus, this paper focuses on measuring the effectiveness of online advertising in form of banners since the consensus supports the banner advertisements as the dominant and most prevalent form of online advertising. .. full view of the effectiveness of the advertising message requires being measured on all three dimensions Figure 2.3 The Venn Diagram of Contemporary CAB Criteria Behavior Cognition Affection 2.2.2 Measures of Effectiveness of Online Advertising At present, the measures of online advertising effectiveness have not reached consensus The question of what constitute the appropriate measures of effectiveness. .. Researchers and advertising practitioners have long sought to understand how advertising works A number of measures have been proposed to empirically evaluate advertising effectiveness Some contend that advertising is effective only when it sells (Little, 1979) Others argue that there is a series of stages between the point of unawareness of a product and/ or brand and the ultimate purchase/sale of a particular... affects effectiveness of online advertising The interaction between involvement and levels of interaction will also be examined In sum, the purpose of this study is to examine whether traditional principles of mass media advertising apply in this new commercial environment; whether several acknowledged important characteristics of online advertising can produce short-term effects on each dimension of effectiveness. .. customer’s involvement on the effectiveness of online advertising, which has important implications for advertisers to optimize their online campaigns Third, this study investigated the function and impact of executional cues of banner advertisement The findings help to provide insights on the value of different executional features and reveal more practical routes of achieving online effectiveness As such,... Background 2.1 Online Advertising Online advertising generally follows the same principles as traditional advertising Nevertheless, the Internet as a unique medium has led to huge differences between online and traditional advertising Firstly, owing to the digital Web, online advertising is interaction-orientated Online ads can be directly activated, which is a form of interaction, a kind of user response... mediator of advertising effects on brand attitude and purchase intention (Aaker, Stayman & Hagerty, 1986; Lutz, 1985; MacKenzie, Lutz & Belch, 1986) According to MacKenzie et al (1986), there can be a one-way or two-way flow between Aad and Abrand owing to different situations; both Aad and Abrand are independent determinants of purchase intention As discussed above, the effectiveness of online advertising. .. endows, online advertising expands the character of peripheral cues Online advertisement can be externally paced and integrated with visual effects and movements, sounds and texts Thus, it is believed that some popular executional cues of online ads, such as animation and image all fall into the categories of peripheral cues (Kalyanaraman & Oliver, 2001) Amination especially has been deemed as one of the... evaluation of entertainment value of the ad; utilitarianism refers to the evaluation of usefulness of the ad, and interestingness is viewed as an evaluation of curiosity The attitude towards the ad (Abrand) measures the extent to which respondents have a positive or favorable opinion of the brand Researchers and practitioners have been using multi-attribute models to study consumer attitude towards brand ... Conceptualization of Online Advertising Effectiveness 2.2.1 Criteria of Effectiveness Despite its expanded functions, the issue of criteria of online advertising effectiveness is still part of the broader... Conceptualization of Online Advertising Effectiveness ……………… 2.2.1 Criteria of Effectiveness ……………………………………………… 2.2.2 Measures of Effectiveness of Online Advertising …………………… 13 2.3 Executional Cues & Effectiveness …………………………………………... Measures of Effectiveness of Online Advertising At present, the measures of online advertising effectiveness have not reached consensus The question of what constitute the appropriate measures of effectiveness