Journal of Interactive Advertising, Volume 3, Number 1, Fall 2002
Rapid growth in online advertising revenues indicates the viability of World Wide Web advertising as an alternative to that of traditional media. Practitioners and academicians recognize that building credibility is important in this relatively new environment. To date, no academic research has explored the interplay of vehicle and advertiser source credibility in determining advertising effects on the web. The present study explores antecedents and consequences of online advertising credibility and examines the effects of (1) website credibility, (2) ad relevance (the advertised product’s relevance to website content), and (3) advertiser credibility on ad credibility, ad and brand attitudes, and product purchase intentions. Structural equation modeling systematically tested and refined a model representing interrelated relationships among the relevant variables. The results suggest that source credibility is vital to understanding web advertising effectiveness. Managerial implications and directions for future research are provided.
The Internet has become a part of everyday life. Over 160 million people are estimated to use the Internet (Nielsen NetRatings 2002) for a variety of tasks including information search and entertainment. Since 1994 when the first Internet banner ads were sold, advertisers have been quick to develop the medium and reach their audiences. Internet advertising revenue in the United States marked $1.55 billion for the first quarter of 2002 (IAB 2002), and although its revenues declined 18% from the same period in 2001 due to the slowing economy and the subsequent downward slope in the overall advertising market, Internet advertising is expected to continue. In fact, advertising on the Internet is predicted to grow to $11.5 billion in 2003, exceeding advertising spending in some traditional media such as magazines and radio (IAB 2000). Considering Internet advertising’s growth, there is little doubt that the Internet is a powerful and viable alternative to traditional media advertising.
With abundant information available to consumers from ostensibly unlimited and often unfamiliar sources, building credibility and recognition is an essential challenge to Internet marketers (Elliott 1999; Moran 1999; Smith 2000). Conventional wisdom is that credibility or trust is vital to successful electronic businesses. However, the credibility of information presented on the Internet has been questioned (Rieh 2002; Wathen and Burkell 2002). Although many consumers use it, the Internet is still a relatively new and sometimes disorientating place for them. Consumers rarely find the face-to-face or voice-to-voice interpersonal reassurance provided by sales representatives. Nor are they likely to find firsthand, physical or tactile experiences that are available through brick and mortar stores. For some, the concern for the risks of doing business online may be assuaged by trust for an established, credible brand or company. Thus, online consumers are likely to seek out and purchase brands they trust (Moran 1999). Although marketers are still not sure how to develop and maintain their credibility or trust (Moran 1999; Smith 2000), dot-com companies spent a considerable amount of money on advertising in traditional media (Diaz 2000; Fitzgerald 2000; Webster 2000) to achieve initial brand recognition and direct consumers to their websites.
Practitioners and academics both have acknowledged that the website credibility, and the Internet at-large, should be put into an equation of its message effectiveness (Johnson and Kaye 1998; Moran 1999; Rieh 2002; Smith 2000). However, the sources and effects of website and banner advertising credibility have not been fully examined; indeed, it is not yet known if present theories of advertising effects can accommodate the medium.
The study presented in this article explores the role of source credibility in determining Web advertising effectiveness. Using traditional advertising effects theories, the study goes beyond recent investigations of vehicle effects on consumer response to banner ads. Specifically, the study examined the effects of website credibility and advertising credibility on ad credibility, ad and brand attitude, and product purchase intention. The effects of banner ad relevance (i.e., product category advertised to website editorial content) on consumer perceptions of the ad and brand are conceptually developed based on studies of the matching of celebrity sources to products (Kamins and Gupta 1994; Misra and Beatty 1990; Till and Busler 2000). A structural equation model of the variables supports the use of traditional effects models; source credibility and website content/advertised product relevance significantly influence banner ad outcomes. The results contribute to the development of a theory of source credibility and advertising effectiveness on the Web and provide practical insights into advertising media strategies.
Web advertising has diverse formats. Sponsorships, classifieds, and interstitials are among various formats of advertising on the web. Yet, banner ads are the most prevalent form of web advertising (Zeff and Aronson 1999) accounting for one-third of total revenues during the second and third quarters of 2001 (IAB 2001). Some view banners to be the "ad for the ad" (Harvey 1997, p. 12) because they direct audiences to their associated target sites (Rossiter and Bellman 1999), often viewed as the "actual ad" (Harvey 1997). Target sites, which include corporate websites, campaign sites, and electronic commerce sites, provide a great amount of information that cannot be offered in banners. Regardless of their physical space limitations, banner ads have proven to successfully generate advertising effects such as brand awareness, attitudes, and purchase intention (Briggs and Hollis 1997; IAB 2001) as well as build web traffic (Li 1998). For this reason, they have attracted the most attention by practitioners and researchers to date.
Incipient research on banner ad effectiveness essentially ignored credibility and focused on ad or message-related characteristics (Ju-Pak 1999; Li 1998). For example, one experimental study found that animated (versus static) banner ads generated better recall, and large (versus small) banner ads enhanced viewer comprehension (Li and Bukovac 1999). More recent work has attempted to develop a theoretical framework for advertising effects on the web (e.g., Cho 1999; Rodgers and Thorson 2000; Rossiter and Bellman 1999) with a shared assumption that audiences might process web advertising information differently from that in traditional media due to the distinctive characteristics of the medium such as interactivity. In response to consumer and practitioner concerns (Barilla 2000; Quinn 2001) academic attention has turned to credibility (Flanagin and Metzger 2000; Johnson and Kaye 1998, 2000).
Credibility on the Web
The Internet has become an important source for information. Users encounter abundant and diverse information online, but often will intentionally seek information thought to be useful for achieving their search objectives (Wathern and Burkell 2002); perhaps not an easy task in the new environment where central control over information is absent. With no overarching quality control or editing process, anyone can author or provide information on the Internet. Accordingly, understanding consumer judgments of web information credibility is an essential issue for the development of ad effects models.
Consumers have been found to use cues available at the site to filter information on the Web. For example, the website provider’s reputation and URL domain type (e.g., edu, gov, org, or com) are among the criteria consumers utilize to judge credibility of information presented in the site (Rieh 2002). Consumer judgments of website information credibility are more a function of the website provider’s credibility (viewed as the source of information), than by the perceptions of the actual author or creator of the content (Rieh 2002). Likewise, credibility of advertising on the Web would be influenced by the reputation of the website provider. If traditional theoretical approaches to advertising effects apply to the Internet, then the perceived credibility of the website may affect the credibility perception of an ad placed in the site and the subsequent attitudinal outcomes of the ad. A useful approach to understanding these relationships is source credibility.
"Source credibility" is a term commonly used to refer to a message sender’s positive characteristics that influence the receiver’s acceptance of the message communicated (Ohanian 1991), and source credibility has been found, in part, to determine the persuasiveness of a message (Petty and Cacioppo 1986). Source credibility perceptions affect message evaluation, attitudes, and behavioral intentions; more credible sources (often endorsers in advertising) have been found to produce more favorable attitudes and stronger behavioral intentions than less credible sources (Atkin and Block 1983; Freiden 1982; Friedman and Friedman 1979; Kamins et al. 1989; Ohanian 1991; Petty, Cacioppo, and Schumann 1983; Sternthal, Phillips, and Dholakia 1978).
Expertise and trustworthiness have been identified as potentially important and enduring dimensions of source credibility (Hovland, Jannis, and Kelley 1953; Ohanian 1990; Tripp, Jensen, and Carlson 1994). Expertise is the extent to which a communicator is perceived to be able to provide valid, accurate information (Hovland, Jannis, and Kelley 1953). In advertising, it is the knowledge that an endorser or spokesperson seems to possess to support the claims made in the advertisements. Trustworthiness refers to an audience’s belief that the communicator provides information in an honest manner, without a motivation for manipulation or deception (Ohanian 1991). Friends and family, for example, are perceived to be more trustworthy than salespeople (DeSarbo and Harshman 1985).
Perhaps the most commonly studied source in advertising is the message presenter or product endorser, often a celebrity. Well-known people or celebrities are viewed as more credible and influential than average people in most cases (Atkin and Block 1983; Freiden 1982; Friedman and Friedman 1979; Kamins et al. 1989; Petty, Cacioppo, and Schumann 1983; Sternthal, Phillips, and Dholakia 1978). However, the process of advertising creation and placement involves more communicators or sources. Stern (1994) extended the concept of source credibility by recognizing that research that focuses only on the persona (i.e., a visible presenter in an ad) does not explain the broader picture of source effects in advertising; a full model of source effects must include the sponsor (or advertiser) and the author of the ad. Creators of ads (e.g., designers or producers) in general are not known to the audience. Of particular note is that advertisers or sponsors of ads appear to be an important source for ad effects (Goldsmith, Lafferty, and Newell 2000; MacKenzie and Lutz 1989).
Recently, advertiser or corporate credibility has received close attention (Goldberg and Hartwick 1990; Goldsmith, Lafferty. and Newell 2000; Lafferty and Goldsmith 1999). Corporate credibility is defined as "the extent to which consumers believe that a firm can design and deliver products and services that satisfy customer needs and wants" (Keller 1998, p. 426), and has been found to have direct, positive effects on attitude toward the ad, attitude toward the brand, and purchase intent (Goldsmith, Lafferty, and Newell 2000). In some Attitude-toward-the-Ad research, advertiser credibility, defined as "the perceived truthfulness or honesty of the sponsor of the ad," appeared to exert a strong effect on attitude toward the ad by enhancing ad credibility sponsored by the company or advertiser (MacKenzie and Lutz 1989, p. 51). Ad credibility refers to the extent to which consumers perceive the message in the ad to be believable, and is based largely on "the trust a consumer places in the source of that particular ad" (the credibility of the advertiser) (MacKenzie and Lutz 1989, p. 61). Accordingly, advertiser credibility has been identified as one of the important factors determining advertising effectiveness. However, the effects of advertiser credibility in the context of web advertising are not examined yet. Given the little information available in banner ads, well-known advertiser names might lead to credible perceptions of the ad.
Based on the findings of MacKenzie and Lutz (1989) and Goldsmith et al. (2000) concerning the significant influence that advertiser credibility may have on a number of ad-related constructs, the following hypotheses were tested:
H1a: Perceived advertiser credibility is positively and directly related to ad credibility.
H1b: Perceived advertiser credibility is positively and directly related to attitude toward the brand.
H1c: Perceived advertiser credibility is positively and directly related to purchase intent.
Although Goldsmith et al. (2000) found direct effects of corporate credibility on attitude toward the ad, the Attitude-toward-the-Ad model suggests the influences of advertiser credibility on ad attitudes be mediated by ad credibility. Thus, no direct relationship of advertiser credibility to attitude toward the ad was proposed.
Another source of influence, which has been studied less than other types of sources, is the medium of advertising placement. Consumer perceptions of a medium in general, and the particular vehicles (in this case websites) of ad placement would be expected to have effects on consumer responses to an ad.
Vehicle Source Effects: Website Credibility
Advertisements do not appear alone. Mass media advertisements typically appear embedded in a non-advertising (i.e., editorial or programming) context within a medium and vehicle. Vehicle source effect is "a measure of the relative value of an ad exposure as a function of the exposed vehicle" (Aaker and Brown 1972). That is, it represents the differential effect that an ad exposure will have on an audience exposed in one vehicle as compared to that exposed in another. Several empirical studies have shown that different vehicles can influence the persuasiveness of the same ad (Aaker and Brown 1972; Freiden 1982; Winick 1962). High-status vehicles are in general more effective than low-status vehicles in producing positive advertising outcomes.
These vehicle effects evidenced in traditional media, most notably the creation of credibility perceptions, reasonably could be expected for banner ads on the web as well (Shamdasani, Stanaland, and Tan 2001). Consumers are presented with thousands of websites when searching for information and they can easily move from one site to another sometimes even accidentally by clicking on links. Unlike in conventional media such as newspapers and magazines, consumers have instant access to a huge number of vehicle sources on the web. With the large number of websites available, consumers tend to seek credible information provided by well-established websites. In addition, since banner ads are inherently short on information, it is more likely that people might use other cues available at the time of processing (e.g., vehicle source or site provider) to make judgments of the ads. Indeed, website reputation has been found to serve as a cue for consumer inference-making about the content carried in the site including the ads appearing in the website (Hermes 1996; Shamdasani, Stanaland, and Tan 2001).
Shamdasani, Stanaland, and Tan (2001) assessed three-way interaction effects between website reputation, advertised product/website content relevance, and product involvement. The results suggest that consumer evaluations of banner ads for low-involvement products are mainly determined by website reputation, whereas the effectiveness of the banner ads for high-involvement products is enhanced by relevance between the advertised product category and content of the host site. Of note is that both website reputation and relevance were found to have positive effects on the dependent variables (i.e., attitude toward the ad, attitude toward the brand, intention to click on banner ads, and purchase intention). Accordingly, website credibility was expected to lead to favorable advertising outcomes and the following hypotheses were tested to assess the effects of the website credibility on ad and brand evaluations. The credibility perceptions of the website were hypothesized to have a direct effect on ad credibility that in turn influences attitude toward the ad (MacKenzie and Lutz 1989).
H2a: Perceived site credibility is positively and directly related to ad credibility.
H2b: Perceived site credibility is positively and directly related to attitude toward the brand.
H2c: Perceived site credibility is positively and directly related to purchase intent.
As evidenced in Shamdasani, Stanaland, and Tan’s study (2001), content of the website in which banner ads appear influences consumer responses to the ads and the advertised product. Along with vehicle source effects, effects of the context or environment of the vehicle in which ads are embedded have been suggested in conventional media.
Context Effects: Relevance Between Website Content and Advertised Product
Research has shown that an advertiser or endorser delivering the same advertising message to the same audience can generate different effects depending on the ad environment – the context in which the ad is embedded (Aaker and Brown 1972; Chaiken and Stangor 1987). Context can influence audience perceptions of the ad and the advertised product. When the advertised product fits well with the context, the content presented in the environment, the audience is likely to find the ads more useful and interesting. Individuals are more likely to use vehicles that meet their special interests or informational needs within a medium. Correspondingly, ads that promote products that are relevant to the content of the vehicle in which the ads appear would induce more positive audience responses.
Relevancy between content of a vehicle (i.e., website) and the product category of banner ads placed in the site has been found to enhance consumer responses to banner ads (Cho 1999; Shamdasani, Stanaland, and Tan 2001). Research on celebrity endorsers in advertising also provides evidence that a good match or congruence between source and product leads to more positive ad and brand evaluations (Kamins and Gupta 1994; Kamins et al. 1989; Till and Busler 2000). Perhaps the expertise dimension of source credibility can explain relevance or congruence effects. That is, if the site is considered as reputable or prestigious, the relevance or compatibility between the content of the site and the product category extends the credible image of the source, which is perceived as an expert by the consumer. People might be more likely to perceive the placement of the ad as the vehicle’s approval of or endorsement for the product if they notice a congruent fit between the expertise of the source (i.e., website provider) and the product category. Relevance or congruence between the content of the site and the advertised product was predicted to influence consumer credibility perceptions of the ad as well as brand attitudes and purchase intent, and the following hypotheses were proposed to test the expected relationships.
H3a: Relevance between the content of the site and the advertised product on the site is positively and directly related to ad credibility.
H3b: Relevance is positively and directly related to attitude toward the brand.
H3c: Relevance is positively and directly related to purchase intent.
Attitudes toward the ad and the brand and purchase intention have long been examined as advertising outcome variables. The aforementioned roles of website credibility and website content/advertised product category relevance have been discussed in enhancing credibility perceptions of the ad, brand attitudes, and purchase intention. However, they are also expected to exert indirect influences on attitude toward the ad through ad credibility, as evidenced in the Attitude-toward-the-Ad study (MacKenzie and Lutz 1989). In addition, the interrelationships among ad attitudes, brand attitudes, and purchase intent have been suggested and tested in prior research. Many studies have assessed the effects of attitude toward the ad on attitude toward the brand that subsequently influences purchase intent (Goldsmith, Lafferty, and Newell 2000; MacKenzie and Lutz 1989; MacKenzie, Lutz, and Belch 1986). Thus, the following hypotheses are put forth to create a more comprehensive picture of the relationships among the variables, and explain ad effectiveness more completely.
H4a: Ad credibility positively and directly relates to attitude toward the ad.
H4b: Attitude toward the ad positively and directly relates to attitude toward the brand.
H4c: Attitude toward the brand positively and directly relates to purchase intentions.
A study was undertaken to examine source effects in the web advertising process. Although the issue of credibility on the web has received increased attention due to consumer uncertainty and security concerns, research on what determines perceived web ad credibility is scant. The purpose of the present study was to systematically assess the effects of advertiser credibility, website credibility, and website content/advertised product category relevance on the subsequent ad outcome variables: ad credibility perceptions, attitudes toward the ad/the brand, and purchase intent. Toward this end, a model was constructed with the hypothesized causal relationships among the variables (Figure 1). Structural equation modeling systematically and simultaneously tested the interrelated, structural linkages and provided a comprehensive understanding of the process.
Websites and Banner Ads
The current study measured online consumer responses to a realistic banner ad appearing in a created website. For this purpose, eight versions of websites and banner ads were professionally developed to reflect several cases of banner ad placement. A concern of this study was to generate variations in consumers’ credibility perceptions to detect the structural relationships of the relevant components in the proposed model. Correspondingly, two online photojournalism magazine sites with an editorial feature and two banner ads were developed. "LIFE" was chosen as a well-established online magazine and a real but barely known magazine site "foto8" was selected from a search on the Internet as a less known vehicle. To create a more "realistic" exposure setting, the content of the website, a feature story about digital photography was chosen for its easiness to read and expected personal relevance to the participants (i.e., college students) in the study.
The banner ads featured a digital camera or an MP3 player. These product categories were selected for their relevance and availability to the study participants. Digital cameras were expected to more closely relate to the content of the website (high relevance) than MP3 players (low relevance). Kodak and Sony were selected as reputable advertisers for the digital camera and MP3 player ads respectively. A fictitious advertiser, Keica was used to represent the case of an unknown or unfamiliar advertiser promoting its products online. A fictitious, neutral, brand name, MX5200 was used for both products in all of the banners, as a newly introduced brand to prevent any possible effects of prior experience or knowledge with real brands. The rectangular-type banner ads were placed on the top of the front pages of the websites. Participants saw one of the eight versions of the materials. Four banners used for the study are shown in Appendix A.
A total of 294 male (45%) and female (55%) students participated in the study. Subjects were recruited from undergraduate courses at a Midwestern university. The courses were campus-wide electives so that the participants represented a variety of majors and backgrounds. Their ages ranged from 18 to 28 years, with an average of 21 years. Over 90 percent of the participants were single and Caucasians. Missing data were treated with listwise deletion of cases and a final sample size of 267 was used for analysis.
The study was administered in groups of 10-20 in a computer lab on campus. Participants were asked to browse one of the eight varieties of the websites as they normally would. After 5 to 10 minutes passed, they turned off the monitor and filled out a paper-pencil questionnaire. Upon completing each session, they were debriefed and dismissed.
Most of the measures employed here were adapted from relevant research and have been commonly used in the domain of this topic. Attitudes toward the ad and the brand were measured on a three-item, seven-point semantic differential scale. The items were anchored by "good/bad," "pleasant/unpleasant," and "favorable/unfavorable" (MacKenzie and Lutz 1989). The scales were found to be reliable (ad attitudes: = .90; brand attitudes: = .88). Credibility perceptions of the website were assessed on a three-item, seven-point semantic differential scale by asking how "believable/ unbelievable," "convincing/unconvincing," and "credible/not credible" participants thought the website was (MacKenzie and Lutz 1989). This website credibility scale appeared to be less reliable ( = .73). It may be because website credibility has never been measured on the scale although it was employed as a relevant measure. Advertiser credibility ( = .92) was assessed on a six-item, seven-point semantic differential scale anchored by "believable/unbelievable," "credible/not credible," "trustworthy/not trustworthy," "dependable/not dependable," "reliable/unreliable," and "reputable/unreputable" (Lafferty and Goldsmith 1999; MacKenzie and Lutz 1989). Purchase intention was measured on a two-item, five-point Likert scale ( = .78). Relevance between advertised product and website content was measured on a four-item, seven-point, semantic differential scale with endpoints of "compatible/not compatible," "good fit/bad fit," "relevant/irrelevant," and "congruent/not congruent" (Shamdasani, Stanaland, and Tan 2001; Till and Busler 2001). The scale was found to be reliable ( = .94).
Before testing the proposed hypotheses, initial analyses examined differences that resulted from the different versions of the materials shown to the participants. No information on the websites and advertisers were provided and thus the subjects’ evaluations of them were expected to rely on their own judgment. Since the websites and advertisers were either known or unknown, familiarity scores were obtained to assess differences between the two types of websites or advertisers respectively. Next, the effects of advertiser/website familiarity and relevance were tested using ANOVAs.
Subsequently, to better understand interrelationships between the constructs of interest, a proposed structural equation model was tested and retested with AMOS 4, using the two-step model-building approach that tests the measurement model before examining the hypothesized structural linkages (Anderson and Gerbing 1988). A measurement model that included latent constructs was first analyzed. As the next step, the proposed model with hypothesized relationships was tested and refined through a series of tests in an attempt to better explain the data. A final, modified model is presented as a result.
Testing Differences across Conditions
T-test results on website familiarity and involvement with the website content showed no significant differences between the two websites (website familiarity t = .45, df = 265, p > .1; content involvement t = .57, df = 265, p > .1). However, another t-test indicated that known advertisers were significantly more familiar to the participants (M = 6.28) than the unknown advertiser (M = 2.9; t = 20.28, df = 265, p < .01). Significant differences in perceived relevance between the advertised product and the website content was also found between the two product categories by using a t-test (digital camera M = 5.86; MP3 player M = 4.28; t = 9.40, df = 265, p < .01), although the relevance scores for both products were above the mid-point on the scale.
A series of ANOVAs were performed to examine differences in scores on the variables of interest across the eight versions of the materials. The results showed no significant differences (p > .05) in the measures depending on the varieties of the materials the participants were exposed to. One exception was advertiser credibility; the known advertisers were viewed to be more credible (M = 5.77) than the unknown advertiser (M = 4.67; F(1, 259) = 85.80, p < .01), although the unknown advertiser was also deemed relatively credible as the rating was above the mid-point on the 7-point scale. Although advertiser credibility was found to vary with advertiser types, the ANOVA results did not indicate significant differences in the other relevant variables (i.e., ad credibility, ad attitudes, brand attitudes, and purchase intent) based on the two advertiser conditions. Given the non-significant differences among the materials, structural equation modeling was pursued to test the proposed hypotheses and investigate the interrelationships among the relevant constructs. Differences in the advertiser credibility perceptions were expected to generate variations in the construct of advertiser credibility that was hypothesized to subsequently influence the other variables in the model.
Testing the Measurement Model
The measurement model was tested in the first step. The means, standard deviations and scale reliabilities of each latent construct are reported in Table 1.
Table 1. Means, Standard Deviations, and Scale Reliability Coefficients
The measurement model was analyzed and improved with AMOS 4. All the items or indicators significantly loaded on the corresponding factors (p < .01). Some error terms were allowed to covary as modification indices available in AMOS 4 suggested when the error terms were considered to have reasonable, correlated relationships conceptually as well. Table 2 presents correlations of all latent constructs in the measurement model used in the development and refinement of the structural model and Table 3 reports the factor loadings for each latent variable. The goodness-of-fit indices indicated a good fit of the measurement model with the data: 2 = 263.62, df = 217, p < .05; GFI = .93; AGFI = .90; NFI = .94; CFI = .99; RMSEA = .03.
Table 2. Correlations of Latent Constructs
Table 3. Factor Loadings of Indicators of Latent Constructs
Analyzing Structural Relationships
The first step in testing the fit of the model was to estimate the paths in the proposed model in Figure 1. The parameter estimates for this model are reported in Table 4. Of the twelve proposed relationships, nine were statistically significant in the expected direction (p < .05). H1a (advertiser credibility to ad credibility), H1b (advertiser credibility to brand attitudes), H1c (advertiser credibility to purchase intent), H2a (site credibility to ad credibility), H3b (relevance to brand attitudes), H3c (relevance to purchase intent), H4a (ad credibility to ad attitudes), H4b (ad attitudes to brand attitudes), and H4c (brand attitudes to purchase intent) were supported. Contrary to expectations, however, three relationships were not significant (p > .05), disconfirming H2b (site credibility to brand attitudes), H2c (site credibility to purchase intent), and H3a (relevance to ad credibility).
Table 4. Parameter Estimates
The goodness-of-fit indices suggest that the proposed model fits the data well; 2 = 276.67, df = 223, p < .01; GFI = .92; AGFI = .89; NFI = .94; CFI = .99; RMSEA = .03. However, in an attempt to refine the model and achieve parsimony, the non-significant relationships were tested in the second phase of the analysis by systematically relaxing a restriction and examining the resultant change in chi-square. That is, chi-square indices of the models without each of the paths were reestimated and compared to the proposed model. The paths (1) from site credibility to brand attitudes, (2) from site credibility to purchase intent, and (3) from relevance to ad credibility did not yield significant changes in chi-square: (1) 2difference = 1.27, df = 1, p > .05; (2) 2difference = .61, df = 1, p > .05; (3) 2difference = .73, df = 1, p > .05). Thus, these three paths, which did not significantly contribute to improvement of the model fit, were removed.
Parameter estimates for the revised model are presented in Table 4. Overall this final model accounts for the data well; 2 = 278.71, df = 226, p = .01; GFI = .92; AGFI = .90; NFI = .94; CFI = .99; RMSEA = .03. To compare the original model and the modified model, the chi-square difference between the two models was estimated. The resultant chi-square change did not indicate significant improvement of the final model in the model fit; 2difference = 2.04, df = 3, p > .05. However, the modified model was chosen as the final model because it is more parsimonious without any non-significant causal linkages and the goodness-of-fit statistics indicate a slightly better fit of the revised model with the data (i.e., AGFI = .89 for the proposed model; .90 for the final model). See the final model in Figure 2.
With its brief history, the World Wide Web is still struggling to build credibility or trust on the consumer side; in fact the Wall Street Journal committed an entire section of its September 16, 2002 issue to trust in e-commerce (Wall Street Journal 2002). Trust and the believability or credibility of website providers, their content, and advertising is still at issue. This study proposed and examined a structural equation model for understanding antecedents and consequences of online advertising credibility in the context of banner ads. Overall this study expands the literature on source credibility and web advertising effectiveness by examining relevant variables in a systematic way. The results of the study revealed many significant relationships among the variables of interest. Of particular note are the effects of advertiser credibility and website credibility on consumer responses to banner ads. They may deserve future research attention because they appear to drive ad credibility perceptions that ultimately lead to ad and brand attitudes and purchase intention. Most importantly, the final "best fit" model shows that source credibility perceptions and website/product relevance cannot be ignored in studying ad and brand-related outcomes and have merit to further our understanding of web advertising effectiveness.
Results suggest that credibility perceptions play an important role in the development of ad/brand attitudes and purchase intent. Both advertiser and website credibility perceptions appear to influence banner ad credibility that subsequently affects attitudes toward the ad/brand and purchase intention. The credibility perceptions of these two ad sources (i.e., vehicle and advertiser) determine consumer judgment of how believable the ad is which in turn strongly influences ad attitudes. Although website credibility was not found to exert direct effects on attitude toward the brand and purchase intent, advertiser credibility appears to directly influence brand attitudes and purchase intention. Consistent with the findings in traditional media (Goldsmith, Lafferty, and Newell 2000), advertiser credibility enhances ad/brand attitudes and purchase intention. Perhaps in this still relatively new environment (i.e., the Internet) the credibility of the company who manufactures and sells the advertised product plays a more important role in consumer responses to banner ads, than the credibility of its site of placement. Consumers want to rely on companies they have known and thus they can trust. Website credibility may be more important when the site actually serves as retailer.
Vehicle source effects and context effects in traditional advertising may hold true for online advertising, and therefore should be taken into account to develop effective web advertising strategies. As for media placement strategies on the web, relatively unknown advertisers, who are not perceived to be credible by consumers but want to produce positive consumer perceptions of the brand quickly through online advertising, may need to find places that are viewed as credible sources for their ads and product information in them. When their brand does not have an established image, it would be a good idea to place its ads on websites perceived as credible and relevant to the product category by target audiences. This would facilitate the process of building a credible, favorable image of the brand and greater intention to purchase the product by making the ad of the brand more credible. On the other hand, well-known advertisers with established reputations can rely on their own credible image to generate positive ad outcomes. Although credible websites still help to make their ads more credible, reputable advertisers have more options for vehicles selections. They can place banner ads in somewhat less credible websites because creating ad credibility is not a big concern for them. Rather, they may need to consider relevance as an important factor in selecting websites for banner ad placements because when the website content is closely related to the advertised product category, brand attitudes and purchase intent are enhanced.
Unexpectedly, website content/advertised product relevance was not found to exert a direct effect on ad credibility although it directly influenced attitude toward the brand and purchase intent. The absence of relevance effects on ad credibility may be a function of the ads and conditions created for this study. This study appeared to be a test of the role of relevance when relevance exists (digital camera M = 5.86 / MP3 player M = 4.28), and advertiser credibility is moderate to high (digital camera M = 5.77 / MP3 player M = 4.67). It is possible that the effect of relevance, or the lack thereof, on credibility may be triggered or only salient when it does not exist. Internet users may have developed an expectation of customization of website content. The savvy user may desire and expect advertising to match his or her needs as defined by the website choice, hence a match between the advertiser and information content of the website. Indeed, ad relevance maximizes the utility of both ad and website content to the site visitor. Users may not notice when ads are relevant since a match meets expectations or schema for desirable websites, and so would not trigger processing interrupts (Bettman 1979) or additional elaborations as predicted by schema theory (Hastie 1984; Srull 1981) for unexpected pairings. However, when ad content does not match site content, the user may notice, stop processing, and discount the credibility of the advertiser and ad content on the basis of its low utility. Experienced users may wonder why irrelevant ads appear on the website, and perhaps infer that the advertiser is not as savvy as the user. Schema theory may hold promise to direct future research. Methodologically, future studies might try to create and compare poor matches of product category and website content to relatively good matches for credibility effects.
Although the present study offers a more comprehensive picture of the issue of web advertising credibility by examining the interrelationships between its antecedents and consequences, other variables can be added to the picture. Consumer involvement with the product was not included in the model presented in this study. In addition, involvement with the website content might have differential impacts on these effects. For instance, when consumers have strong interest in the content, they might find the information on the product relevant to the content more useful and in turn have more favorable responses to the ad. Similar to the concept of credibility of advertising in general in the Attitude-toward-the-Ad research in conventional media (McKenzie and Lutz 1989), credibility of the web advertising in general might also determine consumer credibility perceptions of a particular banner ad. Also, future research could test these relationships in the context of other, more traditional media, so that media driven similarities and differences in consumer response patterns can emerge. Source credibility is likely to be important in all media, but consumer perceptions of credibility in web advertising may be less certain than in more familiar, traditional media and vehicles. Audiences might view conventional media as more credible than the Internet (consider for example, how newspapers are perceived as a more credible medium than other traditional media) ("NAA Report" 2001; Flanagin and Metzger 2000), and this difference in media credibility might also create differences in ad and brand credibility, ad and brand attitudes, and purchase intentions.
Pragmatically, the results of the study suggest that popular websites with high prices that reach many Internet users may not be good choices for banner ad placements if the website content is too general to generate relevance perceptions and the subsequently desirable brand attitudes and purchase intent. In addition, although they are related, website ad rates and website perceived credibility may not exactly correspond. Ad rates are mostly determined by direct-response measures such as the number of visitors to the sites or click-through rates. Consumer judgment of website credibility as well as relevance between website content and advertised product should be put into the media placement equation.
Figure 1. Proposed Model
Figure 2. Modified Model
Aaker, David A. and Phillip K. Brown (1972), "Evaluating Vehicle Source Effects," Journal of Advertising Research, 12 (August), 11-16.
Anderson, James C. and David W. Gerbing (1988), "Structural Equation Modeling in Practice: A Review and Recommended Two-step Approach," Psychological Bulletin, 103 (3), 411-423.
Atkin, Charles and Martin Block (1983), "Effectiveness of Celebrity Endorsers," Journal of Advertising Research, 23 (February/March), 57-61.
Barilla, Jean (2000), "Untangle the Web," Better Nutrition, December, 30-31.
Bettman, James R. (1979), An Information Processing Theory of Consumer Choice, Reading, MA: Addison-Wesley.
Briggs, Rex and Nigel Hollis (1997), "Advertising on the Web: Is There Response before Click-Through?" Journal of Advertising Research, 37 (2), 33-45.
Chaiken, Shelly and Charles Stangor (1987), "Attitudes and Attitude Change," Annual Review of Psychology, 38, 575-630.
Cho, Chang-Hoan (1999), "How Advertising Works on the WWW: Modified Elaboration Likelihood Model," Journal of Current Issues and Research in Advertising, 21 (1), 34-50.
DeSarbo, Wayne S. and Richard A. Harshman (1985), "Celebrity-brand Congruence Analysis," Current Issues and Research in Advertising, 17-52.
Diaz, Ann-Christine (2000), "Dot-com Ads Help Power 5.2% Magazine Page Rise," Advertising Age, January 17, 48.
Elliott, Stuart (1999), "When Dot-coms Want to Build Up Their Images, They Hitch Their Web Sites to a Star," New York Times, November 16, C14.
Fitzgerald, Kate (2000), "Dot-coms the Critical Upfront Unknown," Advertising Age, May 15, S60.
Flanagin, Andrew J. and Miriam J. Metzger (2000), "Perceptions of Internet Information Credibility," Journalism and Mass Communication Quarterly, 77 (3), 515-540.
Freiden, Jon B. (1982), "An Evaluation of Spokesperson and Vehicle Source Effects in Advertising," Current Issues and Research in Advertising, 77-87.
Friedman, Hershey H. and Linda Friedman (1979), "Endorser Effectiveness by Product Type," Journal of Advertising Research, 19 (October), 63-71.
Goldberg, Marvin E. and Jon Hartwick (1990), "The Effects of Advertiser Reputation and Extremity of Advertising Claim on Advertising Effectiveness," Journal of Consumer Research, 17, 172-179.
Goldsmith, Ronald E., Barbara A. Lafferty, and Stephen J. Newell (2000), "The Impact of Corporate Credibility and Celebrity Credibility on Consumer Reaction to Advertisements and Brands," Journal of Advertising, 29 (3), 43-54.
Harvey, Bill (1997), "The Expanded ARF Model: Bridge to the Accountable Advertising Future," Journal of Advertising Research, 37 (2), 11-20.
Hastie, Reid (1984), "Causes and Effects of Causal Attribution," Journal of Personality and Social Psychology, 46 (January), 44-56.
Hermes (1996), "Fourth Hermes Internet Survey," University of Michigan, <http://www.personal.umich.edu/~sgupta/hermes/survey3/>.
Hovland, Carl, I. L. Jannis and H. H. Kelley (1953), Communication and Persuasion, New Haven, CT: Yale University Press.
IAB (2000), "Internet Advertising Bureau (IAB) Reports Online Ad Revenue Tops $1.2 Billion for 3Q 1999," <http://www.iab.net/news/contents/revenues.html> (accessed on 1/5/2002).
— (2001), "Internet Advertising Revenue Holds Steady As All Ad Sectors Decline," <http://www.iab.netnews/contents/revenues.html> (accessed on 2/7/2002).
— (2002), "Internet Ad Revenue Totals $1.55 Billion in Q1 2002," <http://www.iab.net/news/pr_2002_08_02.asp> (accessed on 9/2/2002).
Johnson, Thomas J. and Barbara K. Kaye (1998), "Cruising is Believing?: Comparing Internet and Traditional Sources on Media Credibility Measures," Journalism & Mass Communication Quarterly, 75 (2), 325-340.
— and — (2000), "Using is Believing: The Influence of Reliance on the Credibility of Online Political Information among Politically Interested Internet Users," Journalism & Mass Communication Quarterly, 77 (4), 865-879.
Ju-Pak, Kuen-Hee (1999), "Content Dimensions of Web Advertising: A Cross-national Comparison," International Journal of Advertising, 18 (2), 207-231.
Kamins, Michael A., Meribeth J. Brand, Stuart A. Hoeke, and John C. Moe (1989), "Two-sided Versus One-sided Celebrity Endorsements: The Impact on Advertising Effectiveness and Credibility," Journal of Advertising, 18 (2), 4-10.
Kamins, Michael A. and Kamal Gupta (1994), "Congruence Between Spokesperson and Product Types: A Matchup Hypothesis Perspective," Psychology and Marketing, 11 (6), 569-586.
Keller, Kevin Lane (1998), Strategic Brand Management, Upper Saddle River, NJ: Prentice Hall.
Lafferty, Barbara A. and Ronald E. Goldsmith (1999), "Corporate Credibility’s Role in Consumers’ Attitudes and Purchase Intentions When a High versus a Low Credibility Endorser Is Used in the Ad," Journal of Business Research, 44, 109-116.
Li, Hairong (1998), "What Makes Users Click on a Banner Ad: Two Field Experimental Studies of Banner Ad Size, Type and Incentive," Proceedings of the 1998 Conference of the American Academy of Advertising, Darrel D. Muehling, ed., Lexington: Kentucky.
— and Janice Bukovac (1999), "Cognitive Impact of Banner Ad Characteristics: An Experimental Study," Journalism & Mass Communication Quarterly, 76 (Summer), 341-353.
MacKenzie, Scott B. and Richard J. Lutz (1989), "An Empirical Examination of the Structural Antecedents of Attitude toward the ad in an Advertising Pretesting Context," Journal of Marketing, 53 (2), 48-65.
—, —, and George E. Belch (1986), "The Role of Attitude Toward the Ad as a Mediator of Advertising Effectiveness: A Test of Competing Explanations," Journal of Marketing Research, 23 (2), 130-143.
Misra, Shekhar and Sharon E. Beatty (1990), "Celebrity Spokesperson and Brand Congruence: An Assessment of Recall and Affect," Journal of Business Research, 21 (2), 159-173.
Moran, Nuala (1999), "Brand and Trust on the Internet: How to Build Amazon.com," Chemical Week, August 18, S14-S15.
"NAA Report: Newspapers Most Trusted Medium," (2001), The Quill, 89 (6), 6.
Nielsen NetRatings (2002), "Average Web Usage: Month of January, 2002," <http://pm.netratings.com/nnpm/owa/NRpublicreports.usagemonthly> (accessed on 2/11/2002).
Ohanian, Roobina (1990), "Construction and Validation of a Scale to Measure Celebrity Endorsers’ Perceived Expertise, Trustworthiness, and Attractiveness," Journal of Advertising, 19 (3), 39-52.
— (1991), "The Impact of Celebrity Spokespersons’ Perceived Image on Consumers’ Intention to Purchase," Journal of Advertising Research, 31 (February/March), 46-54.
Petty, Richard E. and John T. Cacioppo (1986), Communication and Persuasion: Central and Peripheral Routes to Attitude Change, New York: Springer/Verlag
—, —, and David Schumann (1983), "Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement," Journal of Consumer Research, 10 (2), 135-146.
Quinn, Michelle L. (2001), "Projecting Credibility," Marketing News, 35 (18), 4.
Rieh, Soo Young (2002), "Judgment of Information Quality and Cognitive Authority in the Web," Journal of the American Society for Information Science and Technology, 53 (2), 145-161.
Rodgers, Shelly and Esther Thorson (2000), "The Interactive Advertising Model: How Users Perceive and Process Online Ads," Journal of Interactive Advertising, 1 (1), < http://www.jiad.org/vol1/no1/rodgers/index.html> (accessed on 1/5/2002).
Rossiter, John R. and Steven Bellman (1999), "A Proposed Model for Explaining and Measuring Web Ad Effectiveness," Journal of Current Issues and Research in Advertising, 21 (1), 13-31.
Shamdasani, Prem N., Andrea J. S. Stanaland, and Juliana Tan (2001), "Location, Location, Location: Insights for Advertising Placement on the Web," Journal of Advertising Research, 41 (4), 7-21.
Smith, Craig (2000), "Virgin’s Reputation Gives Its E-activity Instant Credibility," Marketing, April 13, 19.
Srull, Thomas K. (1981), "Personal Memory: Some Tests of Associative Storage and Retrieval Models," Journal of Experimental Psychology: Human Learning and Memory, 7 (6), 440-63.
Stern, Babara (1994), "A Revised Communication Model for Advertising: Multiple Dimensions of the Source, the Message, and the Recipient," Journal of Advertising, 23 (2), 5-15.
Sternthal, Brian, Lynn W. Phillips, and Ruby Dholakia (1978), "The Persuasive Effect of Source Credibility: A Situational Analysis," Public Opinion Quarterly, 42 (Fall), 285-314.
Till, Brian D. and Michael Busler (2000), "The Match-Up Hypothesis: Physical Attractiveness, Expertise, and the Role of Fit on Brand Attitude, Purchase Intent and Brand Beliefs," Journal of Advertising, 29 (3), 1-13.
Tripp, Carolyn, Thomas D. Jensen and Les Carlson (1994), "The Effects of Multiple Product Endorsements by Celebrities on Consumers’ Attitudes and Intentions," Journal of Consumer Research, 20 (4), 535-547.
Wall Street Journal (2002), "Whom Do You Trust," special section on E-Commerce, Section R, September 16.
Wathen, Nadine C. and Jacquelyn Burkell (2002), "Believe It or Not: Factors Influencing Credibility on the Web," Journal of the American Society for Information Science and Technology, 53 (2), 134-144.
Webster, Nancy Coltun (2000), "Ad Spending in Dailies Rises 9.4% in 1999," Advertising Age, May 8, S10.
Winick, Charles (1962), "Three Measures of the Advertising Value of Media Context," Journal of Advertising Research, 2 (June), 28-33.
Zeff, Robbin and Brad Aronson (1999), Advertising on the Internet, New York, NY: John Wiley & Sons, Inc.
1. Known Advertiser / Digital Camera
2. Unknown Advertiser / Digital Camera
3. Known Advertiser / MP3 Player
4. Unknown Advertiser / MP3 Player
Sejung Marina Choi (Ph.D., Michigan State University) is an Assistant Professor of Advertising at The University of Texas at Austin. Her research interests are in the areas of source credibility, cause-related marketing, advertising on the Internet, and cross-cultural consumer behavior; Email: [email protected]. Nora J. Rifon (Ph.D., City University of New York) is an Associate Professsor in the Department of Advertising at Michigan State University. Her research interests include consumer privacy, cause-related marketing, sponsorships, and source credibility; Email: [email protected].
Copyright © 2002 Journal of Interactive Advertising