And now, a word from our sponsor: a look at the effects of sponsored content and banner...

According to a recent Jupiter Communications report (Peach, Blank, and Grahn 2002), on-line ad spending is expected to reach $14 billion by 2007, making the Internet the fourth largest advertising communication medium. This growth is fueled by increases in the on-line population, time spent

online, and the number of people adopting Internet commerce. In addition, the number of Internet ads a net surfer will be exposed to is expected to increase, so that by 2005, the average surfer could be exposed to 950 on-line marketing messages a day (Leuning 2001). Although this sounds like great news for on-line Web communities that adhere to an advertising-based business model, it will become increasingly challenging for advertisers to connect with on-line consumers.

Marketers are already struggling to find winning strategies to cut through on-line clutter and engage consumers to a point that they will process and respond to on-line advertising messages. To date, on-line advertising has not yielded the results some had originally expected, and there is increasing controversy over the effectiveness of banner advertising (Dietz 1998; Kennerdale 2001). Jupiter Research (Peach et al. 2002) predicts that on-line ad spending will double within the next three years; however, the Interactive Advertising Bureau reports that banner advertising revenue decreased significantly in 2000 (Bayan 2001). Although businesses seem committed to on-line advertising, they are not sure that banner advertising is their best option. One alternative to banner advertising is sponsored content. Sponsored content advertising is a text-based message typically embedded in the context of a Web site and related to the subject matter of the site ("Advertising on the Internet: Implications for Marketers and Advertisers" 1997). Sponsored content is especially popular among informational Web site communities, such as iVillage.com and salon.com. Sponsoring companies can develop relationships with community users by providing useful information and becoming part of the community. For example, both Ford Motor Company and L'Oreal Cosmetics are committed to sponsored content. Ford Motor Company sponsors a section of the iVillage.com Web site (a site that features women's issues) and recently provided site visitors with the results of a car-buying survey and information on automobile safety and maintenance (two key areas of interest for female car buyers). L'Oreal Cosmetics renewed its alliance with StarMedia Network to enhance its company brand among Latin American women by providing lifestyle, beauty, and fashion content for Cadamujer.com (a site that caters to Latin American women). This site has been designated the best women's site by iBest in Mexico and Brazil, and executives at Cadamujer note that traffic and su bscriptions have grown while attrition has diminished (Neufeld, Deeks, and Grahn 2002).

Anecdotal accounts and market prevalence indicate that sponsored content can offer several advantages to both the site and the sponsor. The site benefits on three levels: (1) It is able to secure expert, up-to-date content at no cost to the site; (2) it receives advertising revenue; and (3) it has an enhanced image, in that community members perceive the Web site as more responsive to consumer interests and thus have stronger affinity toward the site (Logan 2000; Mara 2000). The sponsor also benefits on several levels: (1) More favorable sponsor-related attitudes are developed; (2) specific beliefs, such as company expertise, trust, and credibility, are enhanced (Logan 2000), as is consumer responsiveness ("Advertising on the Internet: Implications for Marketers and Advertisers" 1997; Mara 2000); and (3) the company is able to develop more meaningful customer relationships, which will ultimately generate more sales (Pack 2001).

This research empirically tests several of the preceding claims through a series of two studies. Because of the high frequency of sponsored content on informational Web communities, the studies evaluate the effects of sponsored content on informational Web communities, rather than shopping portals, through a fictitious informational Web community and three different sponsors across three product categories. The experiments examine attitudinal effects and specific belief measures (e.g., responsiveness, quality, leadership, trust) related to the advertiser and the Web community. In addition, behavioral intentions are assessed through expressed purchase intention for the sponsoring brand/company and revisit intention for the Web site.

BACKGROUND AND THEORETICAL FOUNDATION

Web sites, similar to other publishing venues, cannot successfully operate on membership revenue and, as a result, use an advertising-based business model. On-line, this model typically has been built around the sale of banner advertisements displayed on the Web site. Through these banner ads, consumers are exposed to various advertising messages and are often able to click on the ad for additional information about the product and/or company. Although banner ads are still the cornerstone of Internet advertising-making up 60% of all Internet advertising in 1998 and an expected 40% in 2003 (Ianni et al. 1999)--corporate advertisers and industry analysts question the effectiveness of these ads, in that click through rates have declined substantially in the past few years (Snyder 2001). The decline in click-through rates may be due, at least in part, to the mass of banner ads that on-line consumers are exposed to, deficient creative tactics and placement, and poor strategic communication efforts by companies (Di etz 1998; Kennerdale 2001).

Despite the mixed reviews, many major corporations maintain their plans to keep an on-line banner advertising presence. For example, ESPN, Ford, IBM, and FirstUSA BankOne have used on-line advertising as a key element of their communication strategies. Procter and Gamble has announced that by 2003, 80% of its advertising budget may be spent on some form of Internet advertising (Krishnamurthy 2000). Forrester Research reports a $1 billion increase in on-line ad spending from 1998 to 1999 (Sinnereich 1999). This increase takes into account various forms of on-line advertising, including banner, content, and e-mail advertising. Jupiter Research suggests that although the proportion spent on banner ads has diminished as click-through rates have declined, other forms of Internet advertising have increased (2000). For example, in the second quarter of 1999, on-line sponsorships grew to account for 28% of all on-line advertising (Krishnamurthy 2000). Sponsored content advertising potentially provides an opportunity for advertisers to overcome many of the deficiencies associated with banner ads and helps companies build strong customer relationships, which may ultimately lead to increased brand equity.

Sponsorship

Both anecdotal and theoretical accounts confirm that offline sponsorships can be effective for developing consumer based brand equity (Becker-Olsen and Simmons 2002; Cornwell and Maignan 1998; Gwinner 1997; Gwinner and Eaton 1999; Madrigal 2000; Speed and Thompson 2000), so why not on-line?

Sponsorship is defined as an investment in an activity, cause, or event-in this case, a Web community--in return for access to exploitable commercial potential (Meenaghan 1983). Despite increasing corporate expenditures and interest in sponsorships (Cornwell and Maignan 1998), few studies have systematically examined sponsorship effects and the underlying theories that explain how sponsorship operates. Extant research on sponsorship has focused primarily on the objectives and motivations for investing in sponsorships, sponsor awareness, and identification. These studies show that companies use sponsorships for two primary goals: to increase brand awareness (Cornwell, Ray, and Steinhard 2001; Gwinner 1997; Johar and Pham 1999; Stipp 1998) and to enhance corporate or brand image (Cornwell, Ray, and Steinhard 2001; Gwinner 1997). Furthermore, they demonstrate that a consumer's ability to identify sponsors correctly is biased by market prominence and relatedness (Johar and Pham 1999). What these studies have fail ed to provide is an integrative framework for determining how sponsors can meet these objectives across various types of sponsorship programs (e.g., social versus sports sponsorships or off-line versus on-line) and how the sponsorship process affects consumer judgments of the sponsor as well as of the sponsored event.

Of the few studies that have systematically explored how sponsorship affects consumer judgments, three have explicitly examined consumers' reactions to sponsorship programs and identified factors that influence consumer processing of sponsorship information. Speed and Thompson (2000) use a classical conditioning framework in which they hypothesize that response to a sponsor is affected by attitude toward the event, prior attitude toward the sponsor, and perception of congruence between sponsor and event. Their results suggest that sponsor-event fit, perceived sincerity, perceived ubiquity of the sponsor, and prior attitude toward the sponsor are critical for generating favorable attitudinal and behavioral responses toward the sponsorship. Becker-Olsen and Simmons (2002) build on the congruence construct. They focus specifically on the role of fit in determining consumers' evaluations of sponsorship programs. They argue that fir is important, in that it alters consumers' amount and type of cognitive elaboratio n and ultimately influences consumer attitudes, beliefs, and behavioral intentions toward the sponsor. Their results show that fit affects behavioral intentions through both an affective (attitudinal) and cognitive (belief structure) route. Specifically, they demonstrate that low-fit programs engender more thoughts that are more likely to be negatively valenced than do high-fit programs. They also show that a moderately incongruent relationship, in which fit is created by supporting communications, will lead to increased cognitive effort, and more positively valenced thoughts, and that these effects are persistent over a one-year time period. In contrast, Madrigal (2000) uses a social identity framework to demonstrate the link between sponsorship and purchase intention. The focus is on the fit between consumers' interests and the sponsored event rather than congruence between sponsor and sponsored event. The results suggest that as consumers' interest in the sponsored event increases, so do purchase intentions.

In summary, these studies suggest that fir, cognitive effort, and social identification can significantly influence affective and behavioral responses to sponsorship programs.

Fit

Fit, or perceived congruity, is critical in that it is believed to be the underlying cause for the transfer of affect and the creation of synergy across various types of brand-building relationships, such as sponsorship, brand alliances, and brand extensions (Aaker and Keller 1990; Becker-Olsen and Simmons 2002; Broniarczyk and Alba 1994; Loken and John 1994; Simonin and Ruth 1998; Speed and Thompson 2000). In other words, without a high degree of fit, relationships experience little halo effect, and one of the brands or partners may experience dilution effects rather than brand building.

In addition, fit determines how much thought people give to a relationship and the nature of those thoughts. Because low-fit relationships are inconsistent with expectations, they are likely to engender additional, negatively valenced cognitive elaboration (Keller and Aaker 1992; Simonin and Ruth 1998; Weiner 1985). Conversely, high-fit sponsorship programs can generate more positive feelings toward a company and sponsored organization (Becker-Olsen and Simmons 2002; Harvey 2001). Thus, given what we know about fit, combined with the criticism that banner ads are ineffective because of poor placement (e.g., low-fit with audience and/or site content), this research examines only the case of high-fit relationships.

Cognitive Elaboration

Sponsored content advertising may, because of its informational context, force the user to engage in higher levels of processing and involvement than does banner advertising, and thus may influence consumer attitudes to a greater extent. However, more effortful processing is associated with both costs and benefits (Fiske and Taylor 1984; Garbarino and Edell 1997; Isen and Means 1983; Payne 1982). The costs are often linked to increased time and effort, which lead to the creation of negative affect, whereas the benefits are linked to the reduction of uncertainty and enhanced knowledge, which lead to positive affect.

It is hypothesized that the sponsored content advertiser will benefit from the cognitive effort expended, in that the negative affect induced by processing time will be offset by the greater positive affect generated from increased knowledge. The resulting positive affect should lead to positive company/brand evaluations. These hypotheses are based on the notion that consumers who visit informational Web sites are seeking knowledge; thus, learning may be the primary goal of visiting the Web site. Consumers expect that this increased knowledge will require additional cognitive effort, and thus, the negative affect associated with this effort will be minimal, in that prior expectations are not violated (Myers-Levy and Tybout 1989). However, as consumers engage in greater cognitive effort, they likely will evaluate the motivation of the advertiser (e.g., the advertiser may be motivated by altruism or the ability to influence potential consumers), as well as the quality of the specific content. If the consumer co ncludes that the motivation is more altruistic than commercial (e.g., to provide a better site or help community members) and that the content is useful, it is likely that additional positive affect will be generated and the sponsored content advertiser will benefit. On the other hand, if the consumer concludes that the motivation is more commercial and less sincere, it is likely that negative affect will be generated and the advertiser might not benefit (Speed and Thompson 2000).

The Web site may experience additional negative affect if the consumer holds it responsible for providing biased information or information that takes more than the expected amount of time to process. Thus, if the consumer perceives that he or she must expend extra effort to evaluate not only the message content, but also the message source and the motivation of the Web site and advertiser, negative affect is likely to increase. Furthermore, consumers may attribute less of the positive learning-induced affect to the Web site, because there is an identified sponsor or message source. The overall affect is likely to be negatively valenced.

It is furthermore expected that the Web site will experience the greatest positive affect when no advertising is present because in the absence of an identified content sponsor, the positive learning-induced affect likely will be attributed to the Web site; the consumer will not be expending cognitive effort to evaluate the motivation of the site; and the overall processing time is less than if sponsored content or banner advertising appeared, thus eliminating the negative affect associated with increased processing time.

In the case of a banner ad, the message is expected to require less effort to process but will not be associated with any of the positive affect induced by increased learning. Therefore, it is expected to lead to less positive affect for the advertiser.

In addition, cognitive effort aids in the retention of information (Kokkinaki and Lunt 2000), which increases the long-term benefits of the positive affect generated in the case of sponsored content. Although the long-term effects are beyond the scope of this study, they theoretically provide additional benefits to the sponsored content advertiser.

Social Identification and Reciprocity

Many consumers use on-line communities for social exchanges as well as for information, and develop a great sense of attachment to the community and its members. A person's connectedness to a site is suggestive of Belk's (1998) theory of the extended self, according to which people form emotional attachments to physical possessions, places, people, and groups (in this case, a Web community). This sense of extended, or social, self is important in that it strengthens consumers' feelings of reciprocity toward the site and the sponsoring company, which potentially increases behavioral intentions and actions.

Social exchange theory is predicated on the notion that all sides in an exchange are expected to both give and get. In the case of sponsored content, the sponsor is often viewed as a member of the group who participates in a social exchange while providing expert informational content. In other words, the sponsor, through the site, offers community members free information. Community members are expected to reciprocate with brand and purchase loyalty. The greater the feelings of connectedness or relatedness, the more likely it is that consumers will feel a shared responsibility for the welfare of other members (Lawler 2001). In the case of sponsored content, the stronger the feelings of membership affinity, the stronger are the resulting feelings of reciprocity. Thus, if a consumer has a strong positive relationship with a site, it is likely that he or she will feel good about the company that supplies information to the site and have a stronger sense of responsibility for the success or failure of the sponso r and site. This proposed connection is supported by research in sports sponsorship, which suggests that favorable purchase intentions are more likely to occur as identification with a particular team increases (Fisher and Wakefield 1998; Madrigal 2000). Although this appears to be a somewhat unbalanced agreement, previous research on reciprocity indicates that consumers often give much greater favors than they receive (Cialdini 1993). Therefore, the exchange of information for purchase intention or brand loyalty is a likely result.

HYPOTHESES

The hypotheses are derived from the previous discussion. It is important to note that the focus for these studies is on informational Web communities, rather than on shopping sites. Therefore, community members come to the site for editorial content, news items, and social exchanges, not for advertising messages or to make a purchase, though purchase intention may be triggered by reading some of the content (e.g., information about the hottest movies or concert tickets and when they will be available). The specific hypotheses are as follows:

H1: Sponsored content, when compared with banner advertising, will lead to (a) more positive company attitude, beliefs, and purchase intention; (b) more favorable Web community attitude, beliefs, and revisit intention; and (c) greater cognitive elaboration.

H2: Sponsored content, when compared with no advertising, will lead to (a) more favorable company attitude, beliefi, and purchase intention; (b) less favorable Web community attitude, beliefs, and revisit intention; and (c) greater cognitive elaboration.

H3: Banner advertising, when compared with no advertising, will lead to (a) more favorable company attitude, beliefs, and purchase intention and (b) less favorable Web community attitude, beliefs, and revisit intention.

STUDY 1

The primary purpose of this experiment is to test whether sponsored content is (1) more likely than banner advertising to increase positive affect and positively influence beliefs and behavioral intentions for sponsored content advertisers and (2) processed differently than banner advertising. Secondarily, this experiment evaluates consumer response toward a Web community when these two forms of advertising are present.

Method

Subjects were asked to read a printout of the homepage from a fictitious informational Web community, CollegeLife101.com (see Appendix for example). The printout contained "hot link" buttons down the left side of the page to information on studying, finances, technology, entertainment, careers, and other interesting news for college students. The center box on the page contained an article that was portrayed as written by someone from the Web community or an identified corporate sponsor. After reading the homepage, subjects were asked to answer a series of questions related to the dependent measures. Subjects were told they were participating in a study on Internet Web design. After the study, subjects were debriefed as to the purpose of the study and told that all materials were fictitious. Three months after the study, 56 subjects participated in a follow-up interview.

Materials

Each printout contained a series of hot link buttons that were constant across conditions. The center section of the printout varied across conditions. In the control condition, subjects were exposed to an article that provided information on getting organized, keeping up on the latest movies, or credit card interest rates. The experimental banner condition had the same article with a banner advertisement for the related company above the article; the experimental sponsored content condition used a tag line of an author's name followed by a company icon of the related company at the end of the article.

Subjects

Two hundred seventy subjects from an undergraduate business school introductory marketing course participated in the study for partial course credit. Sixty-two percent of the subjects were men and the average age was 21. Fifty-six subjects agreed to participate in a follow-up interview. Thirty-nine percent of the follow-up respondents were male. Student subjects were used because of the study content: an informational college Web site community. It is estimated that 4.4 million college students will have on-line access by 2004 and 62% of them will regularly engage in on-line purchasing, with total segment spending greater than $2.1 billion ("Marketing to the Internet Generation" 2000). In addition, the 18- to 24-year-old segment consists of the heaviest Internet users with the highest average on-line expenditures ("Marketing to the Internet Generation" 2000). As with any study, these results can only be generalized to this population, but it is expected that many other groups of consumers will not differ in h ow they respond to content that is of interest to them.

Design and Stimuli

The design is a completely randomized 3 (sponsored content, banner advertising, and control) X 3 (company exemplar) between-subjects design and is diagrammed in Table 1.

In condition 1, subjects are exposed to a banner ad from the given company and a click-through icon to that company. In this condition, the article has no identified sponsor. In condition 2, subjects are exposed to the same article with an identified sponsor and a click-through icon next to the company name. In condition 3, subjects are exposed to the same article with no identified sponsor and no banner ad. In conditions 2 and 3, the header to the article is sized larger to take up the space that would have been used for the banner ad.

Pretests

Two pretests were performed to identify three areas of equal interest to college students and find three companies of equal liking and familiarity that represented those areas of interest. A sample of students was given a list of eight topics and asked to rate them according to level of interest (seven-point Likert scale with endpoints of very interesting/not very interesting). The three most interesting topics were technology, entertainment, and financial information ([M.sub.technology] = 6.77, [M.sub.entertainment] = 6.86, and [M.sub.financial information] = 6.46).

A separate sample of students was given a list of three companies offering products or services in the three different product areas (e.g., technology-Dell, Palm, and Nokia) and asked to rate them in terms of liking and familiarity. The three firms that were most liked, M = 6.35, M = 6.46, and M = 6.33, F(1, 23) < 1, and familiar, M = 6.65, M = 6.64, and M = 6.68, F(1, 23) < 1, were Palm (technology), Sony (entertainment), and MasterCard (financial information), respectively. Liking and familiarity were each measured using three seven-point semantic differential scales: good/bad, favorable/unfavorable, negative/positive, and familiar/unfamiliar, have heard of/have not heard of, recognize/do not recognize. Cronbach's [alpha] = .96 and .95 for liking and familiarity, respectively. As noted by Speed and Thompson (2000), a prior positive attitude is necessary for positive attitudinal and behavioral response. Thus, this pretest not only equalizes the companies to ensure that the differences are not related to prior company attitudes, but also ensures that the prior attitudes of the experimental companies are positive.

A fit manipulation pretest revealed that the companies were perceived as fitting with the interest area (Palm, M = 6.78; Sony, M = 6.65; and MasterCard, M = 6.83) and did not differ by company (F < 1). Fit was measured on seven scaled items: similar/dissimilar, consistent/inconsistent, typical/ atypical, representative/unrepresentative, complementary/not complementary, low fit/high fit, and makes sense/does not make sense. Cronbach's [alpha] = .98.

Measures

The manipulation checks for fit, familiarity, and liking employed the same measures as in the pretest.

Attitudes and beliefs were measured in two ways: thoughts generated in response to the experimental materials and closed-end measures. Initial thought listings show the extent to which inferences are made spontaneously while thinking only about the Web site, whereas the closed-end measures are included to test specific beliefs identified in the literature as direct benefits of sponsored content advertising ("Advertising on the Internet: Implications for Marketers and Advertisers" 1997; Logan 2000; Mara 2000).

Overall attitude toward the Web site and company is measured with three items: negative/positive, unfavorable/favorable, and bad/good, Cronbach's [alpha] = .92 and .94 for the Web site and company attitudes, respectively. Behavioral response or reward for the firm is measured by the rated likelihood of three behaviors: consider, purchase, and recommend, Cronbach's [alpha] = .96. Behavioral response to the Web site is measured by the rated likelihood of two favorable behaviors: visit and recommend, Cronbach's [alpha] = .88.

A list of 21 scaled items was included to test specific beliefs regarding the company and Web site. The 21 statements were derived from the specific beliefs discussed in the anecdotal accounts (Logan 2000; Mara 2000). Responses to the 21 items were subjected to an exploratory factor analysis. The principal-factor method was used to extract the factors, followed by a promax rotation. The factor analysis shows that these measures load on five factors for the company and four factors for the Web site and are consistent with a priori expectations. Only factors with eigen-values greater than one were examined. To create a sufficient sample and increase reliability for the factor analysis, respondents from both studies 1 and 2 were included. The results for the factor loadings for company-related variables are provided in Table 2, and the results for the Web site-related variables are shown in Table 3.

As shown in Table 2, factor 1, which accounts for 37.43% of variance and represents a customer-related dimension, referred to as customer responsiveness, includes the following variables: offers vital information to students, acts in the best interests of students, plays a significant role in students' lives, responds to the needs of students, is interested in helping students, cares about its customers, and offers a vital product/service to students. Accounting for 11.21% of the variation, factor 2 appears to revolve around product quality and includes the following: delivers a credible product, has reliable products/services, has high-quality products/services, and products/services need to be carefully researched. Factor 3, which accounts for 4.6% of the variation, concerns category leadership and includes: is a small-time player (reverse scaled in analysis), is a--leader, is a cutting-edge company, is an important company, is a market leader, is an innovator, and has the financial ability to offer the bes t products/services. Factor 4, accounting for 4% of the variation, is related to corporate credibility and trust, and includes makes claims that are hard to believe and is a company that you can trust. The last factor has only one variable that loads on it, offers credible products, and accounts for 1.08% of the variation. This variable also loads nicely on the product-related factor (.543) and is managerially more relevant on that factor. For the purposes of this analysis, the variable is included on factor 2.

As noted in Table 3, factor 1, which accounts for 50.34% of variance, again represents the customer responsiveness dimension, and includes the following variables: offers vital information to students, acts in the best interests of students, plays a significant role in students' lives, responds to the needs of students, is interested in helping students, and cares about its customers. Factor 2, which accounts for 18.34% of the variation, is centered around product quality dimensions, and includes delivers a credible product, has reliable products/services, is an important Web site, has high-quality products/services, offers a vital product/service to students, and products/services need to be carefully researched. Accounting for 15.06% of the variation, factor 3 concerns category leadership and includes is a small-time player (reverse scaled in analysis), is an information leader, is a cutting-edge Web site, is a market leader, is an innovator, and has the financial ability to offer the best products/ service s. Factor 4, accounting for 5.77% of the variation, is related to Web site credibility and trust and includes makes claims that are hard to believe (reverse scaled for analysis), service can be used with confidence, and is a Web site that you can trust.

Results

Generally the results support the company-related portion of the hypotheses, H1, H2, and H3, which state that companies experience the greatest benefits when engaging in sponsored content, followed by banner advertising, and that both are significantly better than no advertising. The results for the Web-related attitudes, beliefs, and behavioral intentions are mixed. Web sites do experience the greatest attitudinal benefit when there is no advertising, but the results for specific beliefs appear to be somewhat contradictory. In addition, H1 and H2, which predict greater cognitive elaboration in the sponsored content condition, are supported.

The manipulation check for perceived fit shows that it does not vary across the experimental conditions, [M.sub.sponsor] = 6.48 and [M.sub.banner] = .42, F(l, 174) < 1. Because there were no company interaction effects, results for the companies are collapsed across all experimental exemplars. All means ate displayed in Table 4.

Effects on the Company

As predicted, sponsored content leads to more favorable evaluations for the overall attitudinal measure, F(2, 263) = 36.54, p < .01; three of the belief measures--customer responsiveness, F(2, 263) = 28.65,p < .01, product quality; F(2, 263) = 29.86, p < .01, and category leadership, F(2, 263) = 34.04,p < .01; and the reward measure--purchase intention, F(2, 263) = 24.06, p < .01. Furthermore, the banner ad leads to more favorable attitudes than the control condition, F(1, 174) = 18.43, p < .01. The banner ad, when compared with the no-advertising control condition, leads to more favorable belief measures for customer responsiveness, F(1, 174) = 28.56, p < .01, and product quality, F(1, 174) = 28.49, p < .01. However, there is no difference between the banner ad and the control condition for category leadership, F(1, 174) = 3.97, p > .37, or purchase intention, F(1, 174) = 2.58,p > .42. There are no differences across any of the conditions for corporate trust, F(2, 263) = 4.23, p > .21.

Effects on the Web Site

As predicted, the Web site experiences the most positive attitude when no advertising is present, F(2, 263) = 16.49, p < .01. There is a moderate dilution effect for customer responsiveness, F(2, 263) = 12.35, p < .01, across both experimental conditions. Within the experimental conditions, sponsorship is perceived as a more customer-responsive form of advertising than is banner advertising, F(1, 174) = 14.48, p < .01. For product quality, sponsorship leads to the strongest beliefs, F(2, 263) = 7.23, p < .01, with no significant difference between the banner and control conditions, F(1, 174) = 2.35, p > .42. Both category leadership and trust show a significant improvement across the experimental conditions, F(2, 263) = 39.35, p < .01 and F(2, 263) = 41.52, p < .01, respectively, and have no differences within the experimental conditions, F(1, 174) = 1.35, p > .41 and F(1, 174) = 1.46, p > .37, respectively. Revisit intention is diluted with the sponsored content, F(2, 263) = 21.07, p < .01.

Initial thoughts were coded by two coders as positive company, negative company, positive Web site, negative Web sire, and other, with an intercoder agreement of 78%. Discrepancies were resolved by discussion. Company thoughts include thoughts about the products (I like my Palm organizer, I cur up my Visa card, and I have a Sony DVD player), the company (this is a great company and it is good that MasterCard warns about credit card fraud), and the advertising choice (I think this is a great way to get their name out and Sony does a lot of banner advertising). Thoughts about the Web site include thoughts about the site content (this has some really cool stuff, I would like to read the article on cheating, and great hot buttons), and advertising on the site (I hate so many banner ads, it is sneaky that they have MasterCard writing the article, and good thing the banner ad is so small). All other thoughts (I never use a credit card, this is dumb that we have to do this, I would never have time to sit and play on the Internet, and I want to see that movie) are not valences. The mean number of thoughts and valences are reported in Table 5.

The thought listings support the closed-ended findings, indicating that respondents experience greater cognitive elaboration in the sponsored content condition and that the greatest positive affect is generated in this condition.

As predicted, respondents engaged in more processing in the sponsored content condition than in either of the other conditions, F(2, 226) = 8.19, p < .03; [M.sub.sponsored content] = 4.04, [M.sub.banner ad] = 2.39, and [M.sub.control] = 2.55. The type of thoughts within the positive company category differ across conditions. In the sponsored content condition, the thoughts tend to be more focused on company credibility, category leadership, and the quality of the article content. Thus, positive affect appears related to company character or image and the credibility of the information presented. In the banner advertising condition, however, positive thoughts tend to focus more directly on the product offering, providing some preliminary evidence that sponsored content messages are processed differently than are banner advertising messages. Consumers may think more holistically about the company when presented with sponsored content and be more product focused when exposed to banner advertising. Negative thoug hts in the sponsored content condition were related to company motivation, whereas negative thoughts in the banner ad condition focused on the company's product offering, thus suggesting that greater cognitive elaboration may lead consumers to question the motivation of the company. Consistent with prior work on sponsorship (Becker-Olsen and Simmons 2002), it is expected that, in cases of lower fir, this elaboration will lead to even greater questioning of company motivation, and it may ultimately negatively affect consumer choice (Barone, Miyazaki, and Taylor 2000).

The content of negative Web site thoughts also differed across the experimental conditions. In the sponsored content condition, the negative Web site thoughts focus on issues of trust, lack of expertise, and eventual lack of editorial control. This suggests that the negative affect is engendered by thoughts related to credibility and motivation, and that the additional cognitive elaboration may have sparked consumers' persuasion knowledge (Friestad and Wright 1995). The negative thoughts in the banner ad condition, in contrast, are more related to issues of how much advertising the community member will be subjected to in the future and other product/Web site quality issues. Thus, the negative affect seems to be more related to processing time and is not offset by positive affect generated through increased knowledge. None of the positive Web site thoughts mentioned learning or increased knowledge related to the experimental content.

Finally, the control condition with no advertising elicited more favorable thoughts about the Web site than did the other conditions, which is expected because respondents were only presented with Web site information prior to the thought listing exercise and were not exposed to any advertising. In this condition, many of the positive Web site thoughts related to increased learning (e.g., this site has some good information and I liked this article, I would like to read the one on cheating), indicating that in the control condition, the positive learning--induced affect is attributed to the site.

The follow-up interviews confirm the findings and support a persuasion knowledge argument. Respondents noted that because "this was an informational Web site, it should not have sponsored content," and if it was going to use this advertising format, then "it needs to be more clearly noted and not just dropped in." This idea of being tricked by the Web site was echoed by several respondents. Respondents noted that they "expect companies to get to customers however they can," but that the Web site should be clear as to what is advertising and what is editorial content. Furthermore, respondents confirmed that although they "do not like banner advertising, it is a necessary evil of a Web site" and they can "easily disregard" it. In addition, they would prefer to have banner ads than pay for a subscription because the banner ads can be "clicked away" or mentally "skipped over." Many of the respondents thought the Web site had solicited the content from the sponsoring company and did not think the company had paid for any advertising space. Thus, they thought it was "very clever" of the sponsoring companies.

Discussion

The results from this experiment provide evidence that sponsored content is more likely than banner advertising to increase positive affect and positively influence beliefs and behavioral intentions for sponsored content advertisers and that it is processed differently than banner advertising. This experiment also provides evidence that consumer response toward the Web community is very involved. Although the Web site experienced the greatest overall attitudinal benefit, as well as the strongest customer responsiveness measure, in the control condition, when respondents were directed to think about the relationship, both forms of advertising provided benefits in terms of creating category leadership for the Web site. This may be a function of the unfamiliarity of the Web site and that the companies were large, well-stablished, highly familiar, and liked by the respondents.

STUDY 2

In an effort to increase external validity, more clearly identify the sponsor, and test the robustness of the effects, a second study was conducted. In this experiment, the same procedures and similar materials were used; however, the experiment was run on a computer rather than as a paper-and-pencil exercise. The sponsored content article header was put in bold and used a larger size type to clearly identify to the reader that the article content came from a sponsor. Finally, because banner ads are often criticized for being static, the banner ad was changed to include an animated graphic and a promotional inducement (respondents were told they would be entered in a contest for a free laptop computer). Many companies are using banner ads with promotional inducements to create higher levels of involvement and processing (Krishnamurthy 2000; Sparta 2000), as well as to boost click-through rates. This change therefore was made to determine if the sponsored content would be as successful when compared with a mor e entertaining, realistic banner ad. In this study, subjects in the banner ad condition were able to click on the ad, enter personal information, and be entered in a drawing for a free laptop computer (respondents were debriefed at the end of the experiment and told there was not a laptop being given away). All other aspects of the materials remained the same.

One hundred eighty subjects in an undergraduate computer programming course participated in the study without any compensation or course credit. The average age was 20.5 and 49% of the respondents were men.

The response forms were completed on-line after respondents viewed the Web site for not longer than 10 minutes. For this study, cognitive effort is measured by amount of time spent reviewing the stimuli, amount of time spent filling in the thought listings, and the number of thoughts generated. These measures are consistent with previous work (Bettman, Johnson, and Payne 1990; Garbarino and Edell 1997). All other measures are the same as in the first study, except for the addition of a question in the sponsored content condition to determine whether respondents were aware that the content was provided by the corporate sponsor. Again, there are no differences across company exemplars. The results are presented in Tables 6, 7, and 8.

Results

Although there were changes from the first experiment (e.g., medium, promotional offer, subject pool, and company icon size in sponsored content condition), the results for the sponsor condition are robust, in that effects for the attitudinal measure and several of the belief measures are replicated in the second study. As in Study 1, the company clearly benefits the most in the sponsored content condition and least when no advertising message is present. In this study, 92% of the respondents in the sponsored condition were aware that the content was provided by the sponsor. Similar to Study 1, the Web site experiences the strongest attitudinal measures in the control condition; however, beliefs about customer responsiveness, product quality, and trust are best in the case of sponsored content. This may be reflective of the change in stimuli and the large portion of the respondents that was aware of the sponsorship arrangement. As in Study 1, cognitive effort is greatest in the sponsored content condition.

Effects on the Company

Similar to Study 1, sponsored content leads to more favorable evaluations for sponsoring companies than do banner advertisements or no advertising for the overall attitudinal measure, F(2, 177) = 29.48, p < .01; two belief measures--product quality, F(2, 177) = 8.18, p < .01, and category leadership, F(2, 177) = 14.05, p < .01; and the reward measure--purchase intention, F(2, 177) = 16.25, p < .01. In contrast to the first study, trust, F(2, 177) 177) = 13.36, p < .01, greatest in the sponsored content condition. Although customer responsiveness is directionally as predicted, the difference between the experimental conditions is not significant, F(1, 118) = 4.12, p > .06. This finding is likely the result of the promotional inducement offered in the banner ad condition. As in Study 1, the banner ad led to a more favorable attitude than did the control condition, F(1, 118) = 8.51, p < .01, and more favorable beliefs regarding product quality, F(1, 118) = 5.98, p < .01, and trust, F(1,118) = 5.49, p < .01. Also as in Study 1, category leadership appears unchanged from the banner ad to the no-advertising condition, F(1,118) = 2.86, p > .48.

Effects on the Web Site

Similar to Study 1, attitude toward the Web site is strongest in the control condition, F(2,177) = 7.46, p < .02. Two of the belief measures, customer responsiveness, F(2,177) = 11.67, p < .01, and category leadership, F(2,177) = 6.38, p < .01, exhibit significantly stronger effects in both experimental conditions than in the control condition. Thus, the changes in the size of the corporate icon in the sponsored content condition and the promotional inducement in the banner advertising condition seem to have the greatest effect on Web site evaluations. Product quality is identical to the first study, in that sponsored content leads to significantly stronger effects than do either banner advertising or no advertising, F(2,177) = 6.49, p < .01. Unlike Study 1, trust, F(2,177) = 14.59, p < .01, is significantly stronger in the sponsored content condition. This may be reflective of the change in testing materials, which incorporated a larger company icon and led to fewer respondents feeling "tricked." The reward measure shows revisit intention is strongest in the case of sponsored content, F(2,177) = 5.98, p < .01. This intuitively makes sense, in that this is the condition in which respondents felt the Web site was the most customer responsive, that the product quality was strongest, and that the Web site was trustworthy.

Cognitive Elaboration

As is consistent with the first study, thoughts were coded by two coders as positive company, negative company, positive Web site, negative Web site, and other, with an intercoder agreement of 86%. Discrepancies were resolved by discussion. Again, company thoughts include thoughts about the products, the companies, and the advertising. Thoughts about the Web site include thoughts about the site content and advertising on the site. All other thoughts are not valenced.

As in Study 1, the thought listings support the closed-ended findings and H1 and H2, which predict that cognitive elaboration will be greatest in the sponsored content condition. This elaboration is associated with greater positive company affect in the sponsored content condition and stronger positive Web site affect in the control condition.

Respondents engaged in significantly more processing in the sponsored content condition than in the other conditions, F(2,177) = 12.48, p < .01, as evidenced by the greater number of overall thoughts, [M.sub.sponsored content] = 4.61, [M.sub.banner ad] = 3.13, and [M.sub.control] = 2.54. Similar to Study 1, the type of thoughts in the positive company category differ across conditions. In the sponsored content condition, the positive affect is driven by company character and positive learning associated with the article. In this study, there are specific thoughts reflective of social identification, in that they specifically note that "I like the Web site and know how hard it is to get advertisers to new sites, so I will support the advertisers; this is a great way to get people to support your company; and people are more likely to buy products when they learn about them on sites of interest." In the banner advertising condition, company thoughts are again focused on the specific product.

Furthermore, when the sponsored content is clearly identified, respondents feel less tricked and experience less negative affect toward the Web site. Thus, there are fewer negative Web site thoughts in the sponsored content condition for Study 2 than in Study 1. In Study 1, the negative Web site thoughts in the sponsored content condition focused on issues of trust; in Study 2, the negative thoughts are related to site design. Finally, as in Study 1, the control condition is the only condition in which there is evidence that the positive learning--induced affect is directed toward the Web site. The positive thoughts generated in the banner condition tend to be more related to the design of the Web site, not to the content of the experimental article.

GENERAL DISCUSSION

Sponsored content can be an effective advertising tool for engendering positive response toward an advertiser and increasing feelings of customer responsiveness, product quality, category leadership, and even purchase intention. The data suggest that this may occur, at least in part, because the increased cognitive effort associated with sponsored content results in positive affect generated from increased learning. The positive affect also may be reinforced through social identification with the Web site and its contributors. Because respondents were only exposed to the stimuli once, whereas most sponsorship programs rely on repeat exposure, the effects from these studies are encouraging.

Furthermore, contrary to some anecdotal accounts that banner ads are ineffective in terms of building positive corporate effects, these results suggest that high-fit banner ads can significantly improve attitude toward the company, customer responsiveness, and perceived product quality. These effects may be the result of the high degree of fit between the banner ad and the Web site content in this particular study, reinforcing the notion that targeted banners are most effective. Therefore, the problem might not be the banner ads themselves, but rather the fir between the banner ads and the Web sire. The results from Study 2, which tested the effects of a high-fit banner ad with a promotional inducement, indicate that banner advertising can also lead to increased feelings of corporate trust and purchase intention. The purchase intention finding in Study 2 may be the result of the promotional inducement, which triggered feelings of reciprocity. Category leadership seems to be the only belief measure unaffected by banner advertising. This may be because the companies chosen for this study are all large, well-known, and well-liked companies.

The implications for the Web community are nor quite as clear. In both studies it appears that the Web site experiences the greatest overall attitudinal benefit in the case of no advertising, but the specific belief measures show a different picture. In both studies, customer responsiveness and category leadership are enhanced when the Web community is associated with the test companies, which might serve as either editorial contributors or advertisers. Although this seems contrary to the attitudinal findings, when respondents were directed to think about the Web site and the specific companies, they may have realized that the partnership between the two had advantages, such as expert content and increased ad revenue, which would benefit the Web site and its members. In contrast, when they first evaluated the relationship, they did so on the basis of prior expectations. It is therefore reasonable to assume that consumers' initial expectations are that an informational Web community should have limited adverti sing and present unbiased content. As they think more elaborately about the relationship, however, they may begin to realize the benefits of sponsored content, as well as banner advertising, for the site and its members. In addition, the ability of the site to entice such large companies as Sony and MasterCard as editorial contributors may make the Web community seem more prominent. This is supported by the results for product quality, which is strongest in the case of sponsored content. Because informational Web community users seek reliable content and could easily see that the content came from a well-liked, highly familiar company, respondents might value the information.

In Study 1, revisit intention was lowest in the case of sponsored content, thus suggesting that respondents may have been suspicious of the motives of the Web site and questioned its sincerity. According to the thought listings and responses from the follow-up interviews, it is reasonable to assume that these findings are a function of consumers' persuasion knowledge. When respondents engaged in cognitive processing, they began to use their persuasion knowledge and felt tricked by the Web community. Respondents suggested that they did not feel negatively toward the company, because they expect that a company will use whatever advertising venue it has to promote itself, but they do not want or expect the Web site to trick them. In further support of this argument are the results of Study 2, in which respondents exhibited the strongest revisit intention in the sponsored content condition when the editorial contributor was clearly noted as a corporate sponsor.

In summary, the closed-ended sponsored content findings are consistent with the Kalorama ("Advertising to the Internet Generation" 2000) study and Logan's (2000) and Mara's (2000) accounts, which suggest that sponsored content is an effective way to reach consumers and to be perceived as helpful (in this case, responsive to customer needs). It is reasonable to expect that by further integrating the company into the site and having it develop a long-term relationship with the site and the community, perceived sincerity and trust will not be questioned, any negative Web site-related evaluations will be diminished, and positive company evaluations will be enhanced.

The research also supports other findings that banner advertising can have positive company effects, as is evidenced by several of the belief measures. This is consistent with the findings of Dynamic Logic, which found that consumers were willing to trade the intrusive nature of banner ads for free information (Cheng 2001). Study 2 further reinforces the lesson that banner ads need to be highly relevant to the site and the audience, as well as provide some added value, such as entertainment or, in this case, promotional inducement.

The thought listings and interviews provide a glimpse into the processing that occurs with these various types of advertising. There is preliminary evidence that suggests sponsored content messages are processed differently than banner advertising messages. As noted, respondents in the banner ad condition focused more on the product, whereas those in the sponsored content condition focused on more corporate-level thoughts, such as credibility, trust, and category leadership. This is consistent with other sponsorship findings that suggest sponsorship goals should focus on corporate objectives rather than specific product attributes and associations (Harvey 2001). In addition, the overall number of thoughts is greater in the sponsored content condition, which indicates that respondents are more involved with the material. Again, this is consistent with some preliminary findings of Harvey (2001), who suggests that sponsorships tend to have a gratitude effect in which the company behind a brand becomes more attra ctive than the brand itself. This effect positively affects consumers' willingness to think about and buy the brand. Thus, depending on the objectives of an advertising campaign, a communications manager may be more inclined to select a campaign that focuses on sponsored content rather than banner advertising. In the L'Oreal campaign, for example, the company was interested in generating interest among a given target market, not necessarily developing specific product beliefs, and the sponsorship campaign was very successful.

SUMMARY AND CONCLUSIONS

Internet advertising is here to stay, and the level of expenditures for Internet advertising will increase. What is in question is how corporate communications managers can best leverage their spending and how Web sites can best serve their customers. This research begins to provide a framework for the Internet advertising options available to communications managers. The study provides evidence that companies can use on-line sponsorship, including sponsored content, as a viable alternative to banner advertising. Sponsorship in the long run may become a value-added alternative or supplement to banner advertising for companies looking to develop strong customer relationships. The study also confirms the positive effects of targeted banner advertising. However, the effects of integrated campaigns that use both targeted banner advertising and sponsored content or more integrated on-line sponsorship programs involving targeted e-mail, questionnaires, and chat rooms have yet to be studied systematically.

Although this research provides preliminary empirical evidence on the effectiveness of two Internet advertising options, it points to several areas for future research. First, though the study has generality across three companies, the results may differ for less well-known or liked companies. The companies in this study may represent a boundary condition,

in that only well-known, well-liked companies were used. Specifically, the category leadership dimension for the Web site may not be significantly affected because a priori beliefs about less well-known and liked companies may not be as strong as those associated with the companies used in this study. There may even be a dilution effect for less well-known and well-liked companies. Furthermore, according to prior work, category leadership or prominence bias is strengthened in more cluttered environments (Pham and Johar 1999). Therefore, a sponsorship program with a smaller company may not enhance either the company or the Web site on this dimension.

Second, only one Web site was tested with one specific target audience. Although there is no reason to believe that 18- to 24-year-olds will respond differently than other groups of consumers with similar familiarity and prior expectations, additional research is needed to test whether the results can be replicated with other types of Web sites and audiences. If prior expectations of on-line advertising vary significantly, it is reasonable to assume that the level of cognitive elaboration associated with the types of advertising may differ, and thus, the responses may differ. These results also should be replicated across other types of Web communities (e.g., shopping portals) in which expectations about advertising may be different.

Third, because the Web community was fictitious, respondents were not actual members and thus did not have a strong relationship or vested interest in seeing the community benefit. Although revisit intention was significant across all conditions, additional research that looks at similar types of measures with real Web community members may provide deeper insight into the issue of reciprocity and social exchange. The effects on the company should be magnified as feelings of reciprocity are stimulated in the community membership. Related to this notion of the relationship among a Web community, its members, and sponsor is the need for studies that consider the persistence of effects when respondents have been repeatedly exposed to the community and the sponsor. This study used a single exposure in both experiments and thus can not determine the effects of repeated exposure. It is expected that repeated exposure will magnify the effects found in this study.

Fourth, further work is needed to better understand the processing differences across these advertising options. As has been theorized in other work (Harvey 2001), the cognitive processes that consumers use to process sponsorship and advertising messages are different. In other words, what causes a consumer to be persuaded by a sponsorship program is different than what causes a consumer to be persuaded by an advertising message. What underlies these differences has yet to be determined.

In light of changing media consumption patterns, fragmentation of consumers, and lifestyle changes, a continual stream of research is needed to develop a framework for understanding when specific types of advertising offerings are most effective.

APPENDIX

College Life 101

Everything You Need to Know to Stay on Top

WELCOME TO THE FINANCIAL CORNER

Where we give you the latest information on all your money matters.

Budgeting Basics  Paying Tuition - scholarships,  Investment Strategies
and Not so Basic    loans, and other resources       and the Market

  Buying a car           All about Credit            Insurance needs

Prepare for the Credit Card Deluge

by Kim Ross

Ready Set Charge You're young, ready to assert your independence, and you are being offered credit cards wherever they turn. It has become a college certainty that before you are settled into your freshman housing, you will be offered a credit card. Most will bring limited credit lines and relatively high interest rates, they are easy to get and there are few income requirements to worry about. Used well, credit cards can help a student establish a respectable credit history. The best reference you'll find on a credit report is a major credit card paid on time, all the time. However, they can also can leave financial bruises that don't heal until long after the diploma has yellowed on the wall.

Watch those rates Many student cards come with high rates, the average rate in 1999 was 17.51 percent, about a half a percentage point above the average for all variable-rate cards. Students seeking credit cards should view teaser rates -- those incredibly low introductory rates that last about six months -- with caution. What will the APR be after the introductory period? Also be aware that one late payment and the rate, whether it's during an intro period or not, often will be jacked up to 20 percent or more, and there will be a hefty late fee to boot. While cash advances may be tempting, check out those interest rates. They can be as high as 22%. And watch the annual fees. Remember that all cards are not the same, you need to shop around.

Play it smart The best way to avoid having to explain that bill to mom and dad is to learn to get by with one or two low-limit credit cards. Keep those balances down. A credit limit of $l,000 is plenty for most students. A good rule of thumb is if you can eat it, drink it or wear it, then it's not an emergency. Another caution: If a lender gives you cards with $1,000 or $2,000 limits, that doesn't mean you can afford to carry a $1,000 or $2,000 balance. Remember, this is not like cutting a class - the credit report folks are here to stay and missed payments will be on your report for seven years, so play it smart.

WHAT'S HOT

WHAT'S NOT

College Life 101

One stop convenience for all your college needs

CLASS NOTES

Library links, discounted textbooks, study tips and course specific study guides

FINANCIAL CORNER

Creative tuition programs, credit cards, and market information

TECHNOLOGY INSIDER

Latest technology reviews - must haves and fads and practical assistance for your next class multimedia presentation

ENTERTAINMENT ZONE

Music, movies, concerts and more

CAREER STOP

Resume help, job postings and interview tips

FASHION ALLEY

New fashions with comfort and style, Steve Madden does it again and what to wear to keep warm and look good

MYWORLD

Binge drinking on campuses - a new report by Harvard's School for Public Health, health resources, surviving stress and helping fellow students in the middle east

TABLE 1

Experimental Design and Company Exemplars

                       Condition 2:         Condition 3:
Condition I:           Sponsored            Control--No
Banner ad                content            advertising

Technology--Palm      Technology--Palm      Technology--Palm

Entertainment--Sony   Entertainment--Sony   Entertainment--Sony

Financial services--  Financial services--  Financial services--
 MasterCard            MasterCard            MasterCard

TABLE 2

Factor Loadings for Company-Related Variables

                                         Component factor loadings

Variable                                1          2          3

Offers vital information              .678      -.376       .305
Is a small-time player                .437      8.751E-02   .650
Makes claims that are hard to         .481       .460       .213
 believe
Is a leader in their industry         .431      9.439E-02   .639
Acts in the best interests of         .695      -.351       .257
 students
Has a significant role in students'   .742      -.523       .160
 lives
Delivers a credible product           .131       .841       .132
Responds to the needs of students     .785      -.292       .276
Is a cutting-edge company             .505      8.546E-02   .644
Interested in helping students        .828      -.207       .142
Products can be used with             .211       .543       .230
 confidence
Has reliable products/services       6.331E-02   .763      1.535E-02
Is a company I can trust              .496       .423       .236
Is an important company               .506      -.118       .595
Cares about its customers             .763      -.142       .422
Is a market leader                    .543       .122       .712
Has high-quality products/services    .137       .900       .166
Offers a vital product/service        .736      -.281       .339
Is an innovator                       .478       .358       .610
Products/services need to be          .172       .854       .179
 carefully researched
Has the financial ability to offer    .427       .500       .681
 the best products/services

                                         Component factor
                                             loadings

Variable                                 4           5

Offers vital information              9.228E-02  -9.863E-02
Is a small-time player               -2.306E-02   5.349E-03
Makes claims that are hard to         -.602       8.619E-02
 believe
Is a leader in their industry        -4.372E-02  -3.056E-02
Acts in the best interests of         -.120       -.215
 students
Has a significant role in students'  -1.593E-02   8.684E-02
 lives
Delivers a credible product            .232       -.345
Responds to the needs of students      .307       6.272E-02
Is a cutting-edge company            -3.708E-03  -5.006E-02
Interested in helping students         .243        .113
Products can be used with              .272        .619
 confidence
Has reliable products/services         .257        .495
Is a company I can trust               .650        .165
Is an important company               -.279      -1.803E-02
Cares about its customers              .115       8.051E-02
Is a market leader                    6.708E-04   6.011E-03
Has high-quality products/services     .190       -.279
Offers a vital product/service         .284       -.229
Is an innovator                       -.231       -.133
Products/services need to be           .212       -.300
 carefully researched
Has the financial ability to offer    2.623E-03   5.500E-02
 the best products/services

Notes: Extraction method was principal-component analysis. Varimax
rotation with Kaiser normalization. Rotation converged in five
iterations. Items in boldface reflect the highest factor loading.

TABLE 3

Factor Loadings for Web Site- Related Variables

                                         Component factor loadings

Variable                                 1          2          3

Offers vital information               .962       .118     8.525E-02
Is a small-time player               5.625E-02    .108        .984
Makes claims that are hard to          .306       .313        .130
 believe
Is a leader in their industry        5.577E-02    .100        .984
Acts in the best interests of          .960       .116     9.507E-02
 students
Has a significant role in students'    .959       .117     9.078E-02
 lives
Delivers a credible product            .106       .955        .155
Responds to the needs of students      .959     9.730E-02  8.158E-02
Is a cutting-edge Web site             .337       .186        .806
Interested in helping students         .961       .116     8.254E-02
Service can be used with confidence    .403       .405     -1.460E-02
Has reliable products/services         .109       .954        .162
Is a Web site I can trust              .347       .478     3.294E-02
Is an important Web site               .242       .605     1.184E-02
Cares about its customers              .952       .120     9.267E-02
Is a market leader                   4.490E-02    .127        .970
Has high-quality products/services     .162       .954        .107
Offers a vital product/service         .109       .954        .162
Is an innovator                        .341       .248        .821
Products/services need to be           .111       .953        .164
 carefully researched
Has the financial ability to offer   9.014E-02    .146        .955
 the best products/services

                                      Component
                                       factor
                                      loadings

Variable                                 4

Offers vital information               .227
Is a small-time player               3.954E-02
Makes claims that are hard to          .620
 believe
Is a leader in their industry        3.780E-02
Acts in the best interests of          .219
 students
Has a significant role in students'    .223
 lives
Delivers a credible product            .215
Responds to the needs of students      .223
Is a cutting-edge Web site             .148
Interested in helping students         .225
Service can be used with confidence    .689
Has reliable products/services         .217
Is a Web site I can trust              .722
Is an important Web site               .670
Cares about its customers              .236
Is a market leader                   4.713E-02
Has high-quality products/services     .217
Offers a vital product/service         .243
Is an innovator                        .213
Products/services need to be           .208
 carefully researched
Has the financial ability to offer   4.948E-02
 the best products/services

Notes: Extraction method was principal-component analysis. Vatimas
rotation with Kaiser normalization. Rotation coverged in five
iterations. Items in boldface reflect the highest factor loading.

TABLE 4

Effects for Sponsored Content and Banner Advertising (Study I)

                                           Company-related

                                  Sponsored
                                   content    Banner ad  Control

Attitude toward company/Web site    6.43 a     6.02 b    5.45 c
Customer responsiveness             6.67 a     5.78 b    5.44 c
Product quality                     6.53 a     6.40 b    5.09 c
Category leadership                 6.45 a     5.66 b    5.59 b
Trust                               5.93 a     5.87 a    5.91 a
Purchase/visit intention            6.35 a     5.97 b    5.95 b

                                           Web site-related

                                               Sponsored
                                    Content    banner ad  Content

Attitude toward company/Web site    5.41 d      5.46 d    5.71 e
Customer responsiveness             5.70 d      4.91 e    5.86 f
Product quality                     5.61 d      5.19 e    5.17 e
Category leadership                 5.30 d      5.28 d    4.70 e
Trust                               5.76 d      5.72 d    5.77 d
Purchase/visit intention            5.86 d      6.53 a    6.56 e

Note: Means in the same row within the same condition with different
superscripts are significantly different at the .05 level.

TABLE 5

Mean Number of Thoughts per Condition (Study 1)

Sponsored           Content  Banner ad  Control

Company, positive    1.46       .46        -
Company, negative     .17       .51        -
Web site, positive    .32       .27      1.72
Web site, negative   1.24       .67       .24
Other                 .85       .48       .59

TABLE 6

Effects for Sponsored Content and Banner Advertising (Study 2)

                                          Company-related

                                  Sponsored
                                   content   Banner ad  Control

Attitude toward company/Web site  6.27 (a)   5.91 (b)   4.96 (c)
Customer responsiveness           6.76 (a)   6.61 (a)   4.91 (b)
Product quality                   6.57 (a)   5.26 (b)   5.01 (c)
Category leadership               6.42 (a)   5.29 (b)   5.23 (b)
Trust                             6.66 (a)   5.77 (b)   5.46 (c)
Purchase/visit intention          6.46 (a)   6.01 (b)   5.43 (c)

                                          Web site-related

                                  Sponsored
                                   content   Banner ed  Control

Attitude toward company/Web site  5.42 (d)   5.37 (d)   5.73 (e)
Customer responsiveness           6.61 (d)   6.57 (d)   5.32 (e)
Product quality                   5.76 (d)   5.01 (e)   4.96 (e)
Category leadership               5.47 (d)   5.48 (d)   4.78 (e)
Trust                             5.92 (d)   4.96 (e)   5.02 (e)
Purchase/visit intention          5.89 (d)   5.12 (e)   5.07 (e)

Note: Means in the same row within the same condition with different
superscripts are significantly different at the .05 level.

TABLE 7

Cognitive Elaboration (Study 2)

                                Sponsored
                                 content   Banner  Control

Processing time--
 stimuli (in min.)                9.36      6.29    6.31
Thought listing time (in min.)    8.26      6.34    5.41
No.of thoughts                    4.61      3.13    2.54

TABLE 8

Mean Number of Thoughts per Condition (Study 2)

Sponsored           Content  Banner ad  Control
Company, positive    2.01      1.12      --

Company, negative     .28       .36      --
Web site, positive    .84       .86      1.56
Web site, negative    .52       .48       .43
Other                 .96       .31       .55

REFERENCES

Aaker, David A., and Kevin Lane Keller (1990), "Consumer Evaluations of Brand Extensions," Journal of Marketing, 54 (January), 27-41.

"Advertising on the Internet: Implications for Marketers and Advertisers" (1997), Kalorama On-line Information.

Barone, Michael, Anthony D. Miyazaki, and Kimberly A. Taylor (2000), "The Influence of Cause-Related Marketing on Consumer Choice: Does One Good Turn Deserve Another?" Academy of Marketing Science, 28 (Spring), 248-262.

Bayan, Roby (2001), "Banner Ads--Still Working After All These Years?" Link-up, 3 (November/December), 14-16.

Becker-Olsen, Karen, and Carolyn J. Simmons (2002), "When Do Social Sponsorships Enhance or Dilute Equity? Fit, Message Source and the Persistence of Effects," working paper, New York University.

Belk, Russell (1998), "Possessions and the Extended Self," Journal of Consumer Research, 15 (September), 139-168.

Bettman, James R., Eric Johnson, and John W. Payne (1990), "A Componential Analysis of Cognitive Effort in Choice," Organizational Behavior and Human Decisions Processes, 45 (1), 111-139.

Broniarczyk, Susan M., and Joseph W. Alba (1994), "The Importance of the Brand in Brand Extension," Journal of Marketing Research, 31 (May), 214-228.

Cheng, Kipp (2001), "Intrusive Ads OK with Web Users," Adweek (November 12), 7.

Cialdini, R. B. (1993), Influence: Science and Practice, New York: HarperCollins.

Cornwell, T. Bettina, and Isabelle Maignan (1998), "An International Review of Sponsorship Research," Journal of Advertising, 27 (1), 1-21.

-----, Donald Ray, and Edward Steinhard, II (2002), "Exploring Managers' Perceptions of the Impact of Sponsorship on Brand Equity," Journal of Advertising, 30 (2), 41-51.

Dietz, Nancy (1998), "Survey: Banners Losing Effectiveness," Advertising Age's Business Marketing, 83 (9), 40.

Fisher, Robert, and Kirk Wakefield (1998), "Factors Leading to Group Identification: A Field Study of Winners and Losers," Psychology and Marketing, 15 (January), 23-44.

Fiske, Susan T., and Shelley Taylor (1984), Social Cognition, Reading, MA: Addison Wesley.

Friestad, Marian, and Peter Wright (1995), "Persuasion Knowledge: Lay People's and Reseachers' Beliefs About the Psychology of Advertising," Journal of Consumer Research, 22 (June), 62-74.

Garbatino, Ellen, and Julie Edell (1997), "Cognitive Effort, Affect and Choice," Journal of Consumer Research, 24 (2), 147-158.

Gwinner, Kevin P. (1997), "A Model of Image Creation and Image Transfer in Event Sponsorship," International Marketing Review, 14 (3), 145-158.

-----, and John Eaton (1999), "Building Brand Image Through Event Sponsorship: The Role of Image Transfer," Journal of Advertising, 28 (4), 47-57.

Harvey, Bill (2001), "Measuring the Effects of Sponsorships," Journal of Advertising Research, 41(1), 59-65.

Ianni, Drew, Marissa Gluck, Marc Johnson, Michele Slack, and Vipul Patel (1999), "On-line Advertising Through 2003: On-line Growth as a Catalyst for Changes in Traditional Business," Jupiter Media Metrix, Digital Content (July).

Isen, Alice M., and Barbara Means (1983), "The Influence of Positive Affect on Decision Making Strategy," Social Cognition, 2 (1), 18-31.

Johar, Gita Venkataramani, and Michel Tuan Pham (1999), "Relatedness, Prominence, and Constructive Sponsor Identification," Journal of Marketing Research, 36 (August), 299-312.

Keller, Kevin, and David A. Aaker (1992), "The Effects of Sequential Introduction of Brand Extensions, "Journal of Marketing Research, 24 (February), 35-50.

Kennerdale, Caspar (2001), "Banner Advertising: Still Alive, but Fundamentally Flawed?" EContent, 24 (2), 56-58.

Kokkinaki, Flora, and Peter Lunt (2000), "The Effect of Advertising Message Involvement on Brand Attitude Accessibility," Journal of Economic Psychology, 20 (1), 41-51.

Krishnamurrhy, Sandeep (2000), "Deciphering the Internet Advertising Puzzle," Marketing Management, 34.

Lawler, Edward J. (2001), "An Affect Theory of Social Exchange," American Journal of Sociology, 107 (2), 321-335.

Leuning, Erich (2001), CNET News.com.

Logan, Toni (2000), "And Now a Word from Our Sponsor," Business2.0 (September).

Loken, Barbara, and Deborah Roedder John (1993), "Diluting Brand Beliefs: When Do Brand Extensions Have a Negative Impact?" Journal of Marketing, 57 (July), 71-84.

Madrigal, Robert (2000), "The Influence of Social Alliances with Sports Teams on Intentions to Purchase Corporate Sponsors' Products," Journal of Advertising, 29 (4), 13-24.

Mara, Janis (2000), "The Halo Effect," Adweek, 41 (May 22), 88-92.

"Marketing to the Internet Generation" (2000), Kalorama Online Information.

Meenaghan, Tony (1983), "Commercial Sponsorship," European Journal of Marketing, 7 (7), 5-71.

Myers-Levy, Joan, and Alice Tybour (1989), "Schema Congruity as a Basis for Product Evaluations," Journal of Consumer Research, 16 (June), 39-54.

Neufeld, Evan, Juliana Deeks, and Rudy Grahn (2002), "Online Advertising: Crafting Adverrorial Opportunities to Attract CPGs," Jupiter Research, Digital Content Vision Report, vol. 2 (January 14).

Pack, Thomas (2001), "Contextual Commerce: Shopping for Long-Term Viability," Econtent, 24 (2), 20-24.

Payne, John W. (1982), "Contingent Decision Behavior," Psychological Bulletin, 92 (14), 382-402.

Peach, Andrew, Jared Blank, and Rudy Grahn (2002), "Online Advertising and E-mail Marketing Through 2007," Jupiter Research, Digital Content Vision Report, vol. 2 (September 16).

-----, Vipul Patel, Corina Matiesanu, David Card, and Joe Laszlo (2002), "Profiling the On-line User,"Jupiter Research, Digital Content Vision Report, vol. 2 (September 16).

Simonin, Bernard, and Julie Ruth (1998), "Is a Company Known by the Company It Keeps? Assessing the Spillover Effects of Brand Alliances on Consumer Brand Attitudes," Journal of Marketing Research, 35 (February), 30-42.

Sinnereich, Aram (1999), "Paid Content: Maximizing Limited Opportunity with Net Value," Jupiter Media Metrix Vision Report, Digital Content (August).

Snyder, Beth (2001), "Not by Ads Alone," EBusiness, 3 (February), 9.

Sparta, Christine (2000), "Incentive Banner Ads Entice E-Shoppers with Primo Offers," Adweek, 41(15), 100.

Speed, Richard, and Peter Thompson (2000), "Determinants of Sports Sponsorship Response," Journal of the Academy of Marketing Science, 28 (2), 226-238.

Stipp, Horst (1998), "The Impact of Olympic Sponsorship on Corporate Image," International Journal of Advertising, 17 (1), 75-87.

Weiner, Bernard (1985), "'Spontaneous' Causal Thinking," Psychological Bulletin, 97 (1), 75-87.

Karen L. Becker-Olsen (Ph.D., Lehigh University) is currently a visiting assistant professor at New York University's Leonard Stern School of Business, 44 West Fourth Street, New York, NY 10012. The author thanks Carolyn Simmons for her review of this manuscript.