Communicators and marketers frequently use multiple communication tools (e.g., advertising, public relations, and direct marketing) or channels (e.g., television, magazines, and the Web) within a single campaign. The ultimate goal of employing multiple communication vehicles is to have them synergize to create the greatest persuasion effect (Caywood, Schultz, and Wang 1991). Synergy is defined as "the interaction of two or more agents or forces so that their combined effect is greater than the sum of their individual effect" (American Heritage College Dictionary 1997). Synergy is the fundamental concept of integrated marketing communications (IMC) and can be achieved through any of four hierarchical levels: unified image, consistent voice, good listener, and world-class citizen (Duncan 1993).

BACKGROUND AND STUDY OBJECTIVE

Many communication professionals believe in the benefits of synergy, yet demonstrating its effects in laboratory or field settings and identifying how synergy operates has proved elusive. Few, if any, studies have examined whether multiple sources really bring about the effects that communicators and marketers expect. No adequate information-processing model is available to explain how people process advertising messages in a multiple-source condition, nor have there been any studies comparing information-processing models in a campaign using multiple sources versus a campaign that uses repetition from one source with the same number of exposures. The objectives of the present study, therefore, were to test the existence of different synergistic effects, as well as to compare the information-processing model of synergy with that of repetition. This study focuses on the integration of television and Web advertising because more and more often the two media are included together (Gilbert 2000; Riedman 1999).

LITERATURE REVIEW AND HYPOTHESES

An advertisement is usually viewed more than once. The effect resulting from repeated exposure to the same advertisement is called the repetition effect and is assumed to be the incremental effect of each additional advertising exposure (Pechmann and Stewart 1988). The effect resulting from exposure to coordinated advertisements is called the synergy effect.

Television-Web Synergies, Attention, and Elaboration

Several studies have directly and indirectly suggested that people would be more motivated to pay attention to and process multiple-source messages than repetitive messages (Brock, Albert, and Becker 1970; Edell and Keller 1999; Grass and Wallace 1969; Harkins and Petty 1981a, 1981b, 1987; McCullough and Ostrom 1974).

Grass and Wallace (1969) tested the satiation effects of television commercials and suggested that viewers would pay more attention to a slightly different commercial than the same repeated message. They speculated that

   viewers become more and more satiated with a commercial
   they have seen many times. But introduction of a visually
   new commercial--even though the approach or concept of
   the commercial is the same--will revive interest level.

Brock, Albert, and Becker (1970) documented that people paid greater attention to a new communication advocating a known stand than to a previously encountered message. Putrevu and Lord (2003) also suggested that a second exposure to a novel stimulus containing similar information will attract more attention than exposure to the same stimulus.

In television--Web synergy situations, both media provide unique features absent in the counterpart, and used together in a campaign can attract more attention than they could by themselves. For example, moving images such as animated graphics can grasp audiences' attention to Internet communications (Blackwell, Miniard, and Engel 2001; Rossiter and Bellman 1999). The interactive nature of the Web also engages audiences and allows them to be active in the marketing communication process (Allen, Kania, and Yaeckel 1998). Television, on the other hand, contains both intensely moving images and sounds that are effective attention-getting devices (Blackwell, Miniard, and Engel 2001; Rossiter and Bellman 1999), which are absent in most Web advertising.

Therefore, audiences are expected to pay more attention to the slightly different Web advertising than to the previously encountered television commercial and vice versa.

H1 : Multiple message sources will result in higher attention from audiences compared with repeating the same message the same number of times.

McCullough and Ostrom (1974) exposed participants to five similar magazine ads for two products (i.e., with the same basic arguments but different order and phrasing in the message arguments) and found that repetition with variation resulted in a significant and immediate positive effect on the viewer's cognitive response activity.

Harkins and Petty (1981a, 1981b, 1987) found in a series of studies that increasing the number of message sources intensified information-processing activity (i.e., as the elaboration-enhancement hypothesis suggests).

Edell and Keller (1999) documented that multiple sources could elicit more ad evaluative thoughts, brand evaluative thoughts, overall evaluative thoughts, and total thoughts when they compared a print-reinforcement sequence (TV-P) with a print-repetition sequence (P-P), and a print-teaser sequence (P-TV) with a television-repetition sequence (TV-TV).

The Web has the quality of print media (DeFleur et al. 1992; Sundar and Nass 1996), but is not identical with it (Eveland and Dunwoody 2002; Karson and Korgaonkar 2001; Sundar et al. 1997). Karson and Korgaonkar (2001) failed to adapt the Elaboration Likelihood Model (ELM) to Internet processing; the ELM has been mostly supported in print media settings. Sundar et al. (1997) concluded that the psychological distinction between the Web and print is not rooted in modality but in some other factors central to the media. Eveland and Dunwoody (2002) demonstrated that the Web elicited more elaboration compared with print because the content structure of the Web is similar to the thought process of the human mind: it makes connections among related pieces of information (i.e., as the structural isomorphism theory posits).

Therefore, when audiences view a television commercial, which is then followed by a corresponding Web advertisement, the visual elements in the Web advertisement will activate and associate with the existing visual and audio memories of the brand with unrestricted processing time. The content structure of the Web that is similar to the way the human mind works will facilitate processing. In contrast, when audiences view Web advertising, which is then followed by a corresponding television commercial, the visual cues in the television commercial will activate the existing and strongly encoded visual nodes of the brand memory. After combining the visual messages with the new audio messages from the television advertising, the brand memory will be extended, which may allow future processing (Keller 1996).

H2: Multiple message sources will result in more cognitive responses from audiences compared with repeating the same message the same number of times.

Television--Web Synergies and Information-Processing Models

According to the Elaboration Likelihood Model (ELM), a person's elaboration likelihood is high when he or she has the motivation to scrutinize the arguments, and vice versa. When elaboration likelihood is high, a person will be attentive to message arguments and his or her attitude will change via a central route to persuasion. When elaboration likelihood is low, a person will be attentive to peripheral cues such as source credibility and his or her attitude will change via a peripheral route to persuasion (Petty and Cacioppo 1986).

Building on previous speculations, people in the television--Web synergy situation are expected to have higher motivation to pay attention to (Allen, Kania, and Yaeckel 1998; Blackwell, Miniard, and Engel 2001; Brock, Albert, and Becker 1970; Grass and Wallace 1969; Putrevu and Lord 2003; Rossiter and Bellman 1999) and scrutinize persuasive arguments than do those in the repetition situation (Edell and Keller 1999; Harkins and Petty 1981a, 1981b, 1987; McCullough and Ostrom 1974). People exposed to a television--Web synergy campaign should form or change their attitudes through the central processing route, whereas people exposed to repeated messages should form their attitudes through the peripheral processing route (Petty and Cacioppo 1986, 1996a).

When individuals engage in central processing, they focus and elaborate arguments relevant to the true merits of the brand (Petty and Cacioppo 1986, 1996a). Social psychologists propose that the greater the number of sources perceived to advocate a position, the more credible the message is perceived to be (Petty and Cacioppo 1986, 1996b; Zimbardo and Leippe 1991). Harkins and Petty (1987) and McLuhan (1964) suggested that multiple sources could result in greater perceived message diversity and higher message credibility. This acts as a motivator to enhance processing (Petty and Cacioppo 1986). Higher perceived message credibility, in turn, determines the number of positive thoughts that affect brand attitude and purchase intention (MacInnis and Jaworski 1989; Petty and Cacioppo 1986) (see Figure 1).

[FIGURE 1 OMITTED]

Perceived message credibility also takes another route to influence brand attitude. People who perceive a message as credible infer that the ad and the brand are credible after integrating the salient and nonsalient ad elements into their working memory (MacInnis and Jaworski 1989). The more credible the brand is perceived to be, the greater the number of positive thoughts generated.

This path is supported by the traditional cognitive response theory, which holds that positive thoughts influence brand attitudes partially through perceived brand credibility (MacInnis and Jaworski 1989; Olson, Toy, and Dover 1982; Petty and Cacioppo 1979). The expectancy-value attitude model suggests that the integration of beliefs (e.g., brand credibility) and evaluation (e.g., cognitive thoughts) determine brand attitude (Darley and Smith 1993; Fishbein and Ajzen 1975).

In contrast, people who are only exposed to repetitive ads focus and elaborate on peripheral cues, such as source credibility, to form brand attitude (Petty and Cacioppo 1996a). The credibility of the medium that carries the advertisements affects the credibility of the ad, which, in turn, affects the credibility of the brand (Hovland and Weiss 1951; Kelman and Hovland 1953; Schiffman and Kanuk 1994). Higher perceived ad credibility may lead to a greater number of positive cognitive thoughts through a lower level of processing, such as the pure affect transfer or heuristic evaluation (MacInnis and Jaworski 1989). Positive thoughts combined with perceived brand credibility will lead to a more positive brand image (Fishbein and Ajzen 1975) and higher purchase intention (see Figure 2).

[FIGURE 2 OMITTED]

Based on the two information-processing models, it is further hypothesized that

H3: Multiple message sources will result in higher perceived message credibility compared with repetitive ads with the same number of exposures.

H4: Multiple message sources will result in higher perceived ad credibility compared with repeated ads with the same number of exposures.

H5: Multiple message sources will result in higher perceived brand credibility compared with repeated ads with the same number of exposures.

H6: Multiple message sources will result in more positive cognitive responses compared with repeated ads with the same number of exposures.

H7: Multiple message sources will result in more positive attitude toward the ad compared with repeated ads with the same number of exposures.

H8: Multiple message sources will result in more positive attitude toward the brand compared with repeated ads with the same number of exposures.

H9: Multiple message sources will result in higher purchase intention compared with repeated ads with the same number of exposures.

METHOD

An experiment was conducted to test the hypotheses posited above. The design was a 4 (media conditions; a between-subjects factor) X 2 (product category; a within-subject factor) mixed design. The four media exposure conditions are (1) WW (two exposures to the target Web ads); (2) TT (two exposures to the target television ads); (3) WT (initial exposure to the target Web ads, followed by an exposure to the target television ads); and (4) TW (initial exposure to the target television ads followed by an exposure to the target Web ads).

Pretest

According to previous studies, prior knowledge influences people's information processing and message elaboration (Celsi and Olson 1988). Therefore, only ads with unfamiliar brands were selected in this study. Two target television commercials and six filter ads were selected from a pool of 20 British commercials. A total of 16 undergraduate students, consisting of 8 males and 8 females, from an "Introduction to Advertising" class were recruited for the pretest. The criteria for selecting the ads for this study were similar ratings for attitude toward the ad, attitude toward the product, and attitude toward the brand, with the target ads having higher [A.sub.ad] than the filter ads. As a result, a soda ad (target ad 1) and a condom ad (target ad 2) were selected as the target ads. The filter ads involved products such as a life insurance company (filter ad 1), cough syrup (filter ad 2), chocolate (filter ad 3), perfume (filter ad 4), beer (filter ad 5), and a telecommunication company (filter ad 6). Different product categories were chosen to avoid the competition effect.

Participants

One hundred eighty-five students from undergraduate and graduate classes were recruited for the study using the extra grade point incentive (n = 173) and a few with monetary compensation (n = 12). The analysis of variance (ANOVA) test shows no difference between the nonmonetary- and monetary-compensation groups in terms of their responses.

Of these participants, 91 students participated under the repetitive ad condition (WW = 41 and TT = 50), and 94 students participated under the multiple-source condition (WT = 46 and TW = 48).

The majority of these participants were females (72%) in their junior (39%) or senior year (36%) of college. Most used the Internet almost every day (69%). Their ages ranged from 18 to 39 with a mean age of 22.

Statistics show that participants' gender, [chi square](1,182) = .01, p > .05, and age, F(1, 171) = 22.33,p > .05, were similar across the two conditions. Whether participants' year in college and level of Web usage differed among the three conditions is uncertain because more than 20% of the cells had an expected count of less than five, which violates the [chi square] test assumption.

Procedure

Before the experiment, each participant was asked to read and sign a human subject consent disclaimer that indicated there were no right or wrong answers. Participants could leave at any time and their answers would be kept in the strictest confidence.

A total of four news stories and eight ads were shown in each condition. News stories were included became they could create a natural viewing atmosphere for the advertisements. The four news stories were from CNN and involved a biography of Connie Chung, an art exhibition by Robert Rauschenberg, a Yohji Yamamoto fashion show, and a publication on the Harry Potter series.

The four Web news stories were revised from the audio portion from the corresponding television news stories to make them more like professional print news. Each news page contained a CNN logo, a photo about the story, a story, and a link to a product site. The first page of each product site contained a brand name, an animated image rotating four major still frames from the television commercial, and a link to the second page. The second page of each product site consisted of the four still frames on the top of the page, the copy transcribed from the spoken audio from the corresponding television commercial, and a link to the next page (Edell and Keller 1999). The order of the product sites was also counterbalanced to eliminate carryover effects. The Web versions of the news stories and advertisements were designed to be as close as possible to the television versions.

In the television and Web integration conditions, participants watched the news and ads from a television or computer first, and then shifted to the other medium for subsequent exposures. Participants were asked to browse the Web in their usual manner and were requested to view every Web page until the stop page appeared. The stop page either told the participants to switch their attention to the television advertising or to start answering the questionnaire.

The target ads always appeared at the end of the stimuli and the questionnaires were answered after all programs were viewed. The time between the last exposure to the target ads and questionnaire completion was the same for all conditions (Edell and Keller 1999).

The experiment took about 30 to 40 minutes to complete. When the participants were finished they were thanked and dismissed.

Measures

Measures were ordered in a way that would minimize carryover and contamination effects. The specific measurement order was as follows: message credibility, source credibility, cognitive thought listing, attitude toward the product, attention, manipulation check (i.e., prior knowledge of the ad and the brand), ad credibility, attitude toward the ad, brand credibility, attitude toward the brand, purchase intention, and personal information.

Message Credibility

To determine the respondents' perceived message credibility, a three-item, five-point semantic differential scale (i.e., reliable/ not at all reliable, credible/not credible, believable/not at all believable) was used to answer the questions, "In terms of the information you received about (soft drink/condoms), how credible is the information you received?" (Cronbach's [alpha] = .78).

Source Credibility

To determine the perceived credibility for television and the Web, respectively, a three-item, five-point semantic differential scale (i.e., reliable/not at all reliable, credible/not at all credible, believable/not at all believable) was used to answer the questions, "How would you rate (television/the Web) as a source of product information in general?" (television: Cronbach's [alpha] = .75; Web: Cronbach's [alpha] = .75).

Cognitive Responding

Thought listing has been used to assess the number and valence of the respondents' thoughts during ad exposures. First, respondents were requested to list all of the thoughts that went through their minds while they were watching the ad for Tango brand soft drink or Mates condoms. Second, respondents were asked to rate each of the thoughts as positive, negative, or neutral. The total number of cognitive responses and the total number of positive, negative, and neutral thoughts were used for data analysis.

Attitude Toward the Product (the Covariate)

Previous studies showed that the attitude toward the product category affected the advertising effect (Arias-Bolzmann, Chakraborty, and Mowen 2000). Respondents were asked to describe their general attitudes toward the soft drink or condoms along a four-item, five-point semantic differential scale (i.e., very likable/not very likable, interesting/not interesting, good/bad, and appealing/not appealing) (Cronbach's [alpha] = .87).

Attention

Two statements, "The advertisement caught my interest" and "I paid close attention to the advertisement," were used to estimate respondents' attention to the ad, with one equaling "strongly disagree" and five equaling "strongly agree" (r = .47, p < .00).

Prior Knowledge About the Ad and Brand (Manipulation Check)

Participants were asked to respond to three statements to assess their prior knowledge of the ads and brands, with one equaling "strongly disagree" and five equaling "strongly agree." The three statements were, "I have never seen this ad before," "I have never seen this brand before," and "The advertisement was new to me."

Ad Credibility

To assess the perceived credibility of the ad, a three-item, five-point semantic differential scale (i.e., reliable/not at all reliable, credible/not at all credible, believable/not at all believable) was used to answer the questions, "How would you rate the ad for Tango soft drink/Mates condoms?" (Cronbach's [alpha] = .84).

Attitude Toward the Ad

To estimate the attitudes toward the ad, a four-item, five-point semantic differential scale (i.e., very likable/not very likable, interesting/not interesting, good/bad, and appealing/ not appealing) was used to respond to the question, "How would you rate the ad for Tango soft drink/Mates Condoms?" (Cronbach's [alpha] = .93).

Brand Credibility

To assess the perceived credibility of the brand, a three-item, five-point semantic differential scale (i.e., reliable/not at all reliable, credible/not at all credible, believable/not at all believable) was used to answer the questions, "How would you rate Tango soft drink/Mates condoms along these scales?" (Cronbach's [alpha] = .88).

Attitude Toward the Brand

To estimate the attitudes toward the brands, a four-item, five-point semantic differential scale (i.e., very likable/not very likable, interesting/not interesting, good/bad, and appealing/ not appealing) was used to respond to the question, "How would you rate Tango soft drink/Mates Condoms along these scales?" (Cronbach's [alpha] = .93).

Product Purchase Intention

To determine respondents' intentions to purchase the product categories, a four-item, five-point semantic differential scale (i.e., very likely/not very likely, very probable/very improbable, very possible/very impossible, very existent/very nonexistent) was used to answer the questions, "How likely do you feel it is that you would purchase the Tango soft drink/Mates condoms if you were in the market for it?" (Cronbach's [alpha] = .97).

Personal Information

At the end of the questionnaire, each participant's gender, age, class (freshman, sophomore, junior, senior, or graduate student), and Web usage were measured. Web usage was assessed through the question, "How often do you usually use the Internet to browse Web pages?" with the selections of "Almost every day," "About once or twice a week," "About once or twice a month," and "Rarely."

Data Screening

Proper data screening procedures were employed before the data were analyzed. A nearly 100% data file accuracy rate was found. Few missing values existed in the data. Missing data values were replaced by group mean values (Tabachnick and Fidell 1996, p. 62). The data were normally distributed and complied with the "homogeneity of regression" assumption of analysis of covariance (ANCOVA) for each of the dependent variables (Tabachnick and Fidell 1996).

RESULTS

Manipulation Check

Among the 185 participants in this study, only one participant reported awareness of the ad for the Mates brand condoms. Only the data for Tango soft drink were considered for this participant; the data for Mates condoms were treated as missing.

Evaluation of Covariate

The variable attitude toward product ([A.sub.pr) was demonstrated to be a good covariate statistically. Results showed that A correlated significantly with most of the dependent measures (except total number of thoughts) and was not correlated with the independent variable (Tabachnick and Fidell 1996).

Test of the Hypotheses

An ANCOVA test for each of the dependent variables was performed with attitude toward the product as the covariate (see Table 1).

H1: Higher Attention

The first hypothesis is supported. The ANCOVA test shows that people in the multiple-source (MS) condition did pay more attention to the test advertisements than did those in the repetitive (Rep) ad condition, MS = 3.95, Rep = 3.88; F(1, 93) = 2.77,p = .05.

H2: More Cognitive Responses

The second hypothesis is supported. People in the multiple-source condition generated more cognitive thoughts when exposed to the advertisements than those in the repetitive ad condition, MS = 3.45, Rep = 2.98; F(1, 94) = 12.07, p < .0.

H3: Higher Perceived Message Credibility

The third hypothesis is also supported. People in the multiple-source condition rated the product information that they received as more credible than those in the repetitive ad condition, MS = 2.98, Rep = 2.63; F(1, 36) = 9.03, p < 0.

H4: Higher Perceived Ad Credibility

The fourth hypothesis was rejected. The test shows that people in the multiple-source condition did not perceive the ad as being more credible than those in the repetitive ad condition, MS = 3.13, Rep = 3.10; F(1, 94) = 1.81, p = .09.

H5: Higher Perceived Brand Credibility

The fifth hypothesis was rejected. The test shows that people in the multiple-source condition did not have higher perceived brand credibility than those in the repetitive ad condition, MS = 3.11, Rep = 3.05; F(1, 93) = 2.48, p = .06.

H6: More Positive Thoughts

The fourth hypothesis was supported. People in the multiple-source condition generated more cognitive responses and more positive thoughts than their counterparts in the repetitive ad condition, MS = 1.60, Rep = 1.31; F(1, 93) = 9.30, p = 0.

H7: More Positive Attitude Toward the Ad

The seventh hypothesis was rejected. The multiple source did not produce more positive attitudes toward the ad, MS = 3.62, Rep = 3.58; F(1, 94) = 1.90, p = .08.

H8: More Positive Attitude Toward the Brand

The eighth hypothesis was rejected. People in the multiple-source condition did not form a more positive attitude toward the brand than those in the repetitive ad condition, MS = 3.40, Rep = 3.37; F(1, 94) = 1.23, p > .05.

H9: Higher Purchase Intention

Hypothesis nine was rejected. People in the multiple-source condition did not have higher purchase intention than their counterparts in the repetitive ad condition, MS = 2.74, Rep = 2.86; F(1, 94) = .03, p > .05.

TEST OF THE MODELS

This study hypothesized that people in the multiple-source condition would form their brand attitude via a central processing route, whereas people in the repetitive ad condition would form their brand attitudes via a peripheral processing route.

AMOS was used to conduct a path analysis to test each attitude model toward the brand. The maximum likelihood estimation, the most popular estimation method in structural equation modeling (Tabachnick and Fidell 1996, p. 748) was employed to estimate the two models.

The test indicates that the data fit our model for the multiple-source condition very well: [chi square](8, N = 94) = 15.27, p > .05; NFI [normed fit index] = .99; CFI [comparative fit index] [approximately equal to] 1 (Bentler 1995; Tabachnick and Fidell 1996); RMSEA [root mean square error of approximation] = .099 (Steiger 1989). In all, the model accounted for 65% of the variance in brand attitude and 50% in purchase intention. Significant parameter estimates for every path (p < .01) supported our expectations for the relationships between each pair of variables (see Figure 3).

[FIGURE 3 OMITTED]

The value next to each path indicates its [beta] weight. The [beta] weight suggests that message credibility ([beta] = .37) is a better predictor for the number of positive thoughts generated than is brand credibility ([beta] = .31). Conversely, brand credibility ([beta] = .67) is a better predictor for brand attitude than is the number of positive thoughts ([beta] = .22).

The test also indicates that the data fit our model for the repetitive ad condition well: [chi square](9, N = 91) = 18.84, p < .03, [chi square]/df < 5 (Joreskog and Sorbom 1996; McLeod and Perse 1994); NFI = .99; CFI [approximately equal to] 1; RMSEA = .108. In all, the model accounted for 64% of the variance in brand attitude and 50% in purchase intention. The significant parameter estimates for every path supported our expectations for the relationships between each pair of variables (see Figure 4). Similarly, the [beta] weight values suggest that brand credibility ([beta] = .69) is a better predictor for brand attitude than the number of positive thoughts ([beta] = .23).

[FIGURE 4 OMITTED]

CONCLUSION

The hypothesis tests showed that the television--Web synergy did produce an effect that was superior to the repetitive ad condition. Indeed, the television--Web synergy led to higher attention, higher perceived message credibility, and a greater number of total and positive thoughts. The synergy effect finding on the number of total and positive thoughts supported Petty and Cacioppo's view (1986) that the number of sources could serve more than a simple peripheral cue that would lead to a conformity effect. The synergy effects on perceived ad credibility, brand credibility, attitude toward the ad, attitude toward the brand, and purchase intention were not found. This further suggested that the television--Web synergy had relatively little impact on the consumer's affective and conative state as compared with its impact on cognition. When the advertised products were low involvement in nature (e.g., soda and condoms in this study), the impact on cognition was insufficient to produce greater attitudinal and behavioral effects.

The laboratory experiment also revealed what it was about multiple sources that produced greater communication effects than the same number of repeated ads. The model fit indices and significant standardized path estimates clearly indicated that processing in the multiple-source and the repetitive ad conditions was via distinct processing routes. Multiple sources not only elicited a greater number of total and positive thoughts; they also motivated a higher processing level (i.e., central processing) than did repetition. The greater numbers of positive thoughts were generated by perceived brand credibility and message credibility, which resulted from the combined influence of both cognitive and social psychological factors.

Processing in the repetitive ad condition, however, was through a peripheral processing route. The positive thoughts were caused by the perceived ad credibility, which was the result of the transfer or heuristic processing effect influenced by the perceived source credibility.

Implementation

Communicators and marketers that aim to increase the awareness of proposed issues or brands should apply the multiple-source strategy to increase attention to the subsequent persuasive messages. Previous studies demonstrated that the more attention people paid to a message, the stronger the encoding and association (Keller 1987, 1993a, 1993b).

Communicators and marketers that strive for better recall of their persuasive messages should also apply the multiple-source strategy. According to the encoding specification principle (Tulving 1974), successful recall depends on the extent of the match between the information during the initial encoding and later during retrieval. Presenting information in varied contexts causes information to be encoded in a slightly different way, which improves retrieval ability and increases information accessibility.

Moreover, the multiple-source strategy can also be used to motivate message elaboration. Health communicators can use the multiple-source strategy to enhance health message processing.

Limitations and Caveats

This study has its limitations, as do all experiments. This study involved two humorous ads with unfamiliar brands and with low-ticket products. How well the findings of this study can be generalized to other products cannot be concluded. Although the experiment was designed to produce as high an ecological and external validity as possible, the unnatural viewing environment might have caused participants to respond differently. Participant responses were measured directly after exposure to stimuli. Any generalization to delayed responses should be made with great caution.

Future Research

Although this study found a synergy effect with attention, whether the attention was caused by the novelty of the subsequent source, as proposed by Harkins and Petty (1981a, 1981b, 1987), was beyond the scope of this study and was not ascertained. Moreover, multiple sources were expected to produce greater persistence, greater resistance to counterpersuasion, and greater behavior prediction than attitudes formed or changed via repetition (Petty and Cacioppo 1986). Studies investigating these effects are also recommended. Examining the multiple-source effect in areas other than advertising, such as public relations and health communications, is also strongly recommended.

TABLE 1
Summary of Results Based on the Second Proposed Criterion
[SIGMA] ([E.sub.ms] > [SIGMA][E.sub.rep])

         Dependent       [Mean.sub.ms]/   [Mean.sub.rep]/  F (group)/
         variables             SD               SD          p value

H1  Attention                 3.95             3.88        F = 2.77
                               .74              .74        p = .05
                            (n = 93)         (n = 91)
H2  Total thoughts            3.45             2.98        F = 12.07
                               .93              .98        p = .00
                            (n = 94)         (n = 91)
H3  Message credibility       2.98             2.63        F = 9.03
                               .85              .81        p = .00
                            (n = 36)         (n = 27)
H4  Ad credibility            3.13             3.10        F = 1.81
                               .72              .59        p = .09
                            (n = 94)         (n = 90)
H5  Brand credibility         3.11             3.05        F = 2.48
                               .65              .63        p = .06
                            (n = 93)         (n = 90)
H6  Positive thoughts         1.60             1.31        F = 9.30
                              1.03              .93        p = .00
                            (n = 93)         (n = 91)
H7  [A.sub.ad]                3.62             3.58        F = 1.90
                               .90              .84        p =.08
                            (n = 94)         (n = 90)
H8  [A.sub.br]                3.40             3.37        F = 1.23
                               .83              .76        p = .13
                            (n = 94)         (n = 91)
H9  PI                        2.74             2.86        F = .03
                              1.06              .97        p = .43
                            (n = 94)         (n = 91)

         Dependent                  Adjusted   Effect  Observed
         variables       [R.sup.2]  [R.sup.2]   size     power

H1  Attention               .21        .20      .02       .38
H2  Total thoughts          .07        .06      .06       .93
H3  Message credibility     .25        .23      .13       .84
H4  Ad credibility          .22        .21      .01       .27
H5  Brand credibility       .19        .18      .01       .35
H6  Positive thoughts       .22        .21      .05       .86
H7  [A.sub.ad]              .25        .24      .01       .28
H8  [A.sub.br]              .20        .19      .01       .20
H9  PI                      .15        .14      .00       .05

REFERENCES

Allen, Cliff, Deborah Kania, and Beth Yaeckel (1998), Internet World: Guide to One-to-One Web Marketing, New York: Wiley.

American Heritage College Dictionary, 3d ed. (1997), Boston: Houghton Mifflin.

Arias-Bolzmann, Leopoldo, Goutam Chakraborty, and John C. Mowen (2000), "Effects of Absurdity in Advertising: The Moderating Role of Product Category Attitude and the Mediating Role of Cognitive Responses," Journal of Advertising, 29 (1), 35-49.

Bentler, Peter M. (1995), EQS: Structural Equations [program manual], Encino, CA: Multivariate Software.

Blackwell, Roger D., Paul W. Miniard, and James F. Engel (2001), Consumer Behavior, 9th ed., Orlando, FL: Harcourt.

Brock, T. C., S. M. Albert, and L. A. Becker (1970), "Familiarity, Utility, and Supportiveness as Determinants of Information Receptivity," Journal of Personality and Social Psychology, 14, 292-301.

Caywood, Clarke L., Don E. Schultz, and Paul Wang (1991), A Survey of Consumer Goods Manufacturers, New York: American Association of Advertising Agencies.

Celsi, Richard L., and Jerry C. Olson (1988), "The Role of Involvement in Attention and Comprehension Processes," Journal of Consumer Research, 15 (September), 210-214.

Darley, William K., and Robert E. Smith (1993), "Advertising Claim Objectivity: Antecedents and Effects," Journal of Marketing, 57 (4) (October), 100-113.

DeFleur, M. L., L. Davenport, M. Cronin, and M. DeFleur (1992), "Audience Recall of News Stories Presented by Newspaper, Computer, Television and Radio," Journalism and Mass Communication Quarterly, 69 (4), 1010-1022.

Duncan, Tom (1993), "Integrated Marketing? It's Synergy," Advertising Age, 64 (10), 22.

Edell, Julie A., and Kevin Lane Keller (1999), Analyzing Media Interactions: The Effects of Coordinated TV-Print Advertising Campaigns, Cambridge, MA: Marketing Science Institute.

Eveland, William P., Jr., and Sharon Dunwoody (2002), "An Investigation of Elaboration and Selective Scanning as Mediators of Learning from the Web Versus Print," Journal of Broadcasting and Electronic Media, 46 (1), 34-53.

Fishbein, Martin, and Icek Ajzen (1975), Belief, Attitude, Intention and Behavior: An Introduction to Theory and Research, Reading, MA: Addison Wesley.

Gilbert, Jennifer (June 2000), "Special Report: Interact and Integrate," in Vol. 2000: Advertising Age.

Grass, Robert C., and Wallace Wallace (1969), "Satiation Effects of TV Commercials," Journal of Advertising Research, 9 (3), 3-8.

Harkins, Stephen G., and Richard E. Petty (1981a), "Effects of Source Magnification of Cognitive Effort on Attitudes: An Information-Processing View," Journal of Personality and Social Psychology, 40 (3), 401-413.

-- (1981b), "The Multiple Source Effect in Persuasion: The Effects of Distraction," Personality and Social Psychology Bulletin, 7 (4), 627-635.

-- (1987), "Information Utility and the Multiple Source Effect," Journal of Personality and Social Psychology, 52 (2), 260-268.

Hovland, Carl I., and Walter Weiss (1951), "The Influence of Source Credibility on Communication Effectiveness," Public Opinion Quarterly, 15 (4), 635-650.

Joreskog, Karl G., and Dag Sorbom (1996), LISREL 8 User's Reference Guide, Chicago: Scientific Software International.

Karson, Eric J., and Pradeep K. Korgaonkar (2001), "An Experimental Investigation of Internet Advertising and the Elaboration Likelihood Model," Journal of Current Issues and Research in Advertising, 23 (2), 53-72.

Keller, Kevin Lane (1987), "Memory Factors in Advertising: The Effect of Advertising Retrieval Cues on Brand Evaluations," Journal of Consumer Research, 14 (3), 316-333.

--(1993a), "Conceptualizing, Measuring, and Managing Customer-Based Brand Equity," Journal of Marketing, 57 (1), 1-22.

--(1993b), "Memory Retrieval Factors and Advertising Effectiveness," in Advertising Exposure, Memory, and Choice, Andrew A. Mitchell, ed., Hillsdale, N J: Lawrence Erlbaum.

--(1996), "Brand Equity and Integrated Communication," in Integrated Communication: Synergy of Persuasive Voices, E. Thorson and J. Moore, eds., Mahwah, NJ: Lawrence Erlbaum.

Kelman, H. C., and C. I. Hovland (1953), "'Reinstatement' of the Communicator in Delayed Measurement of Opinion Change," Journal of Abnormal and Social Psychology, 48, 327-335.

MacInnis, Deborah J., and Bernard J. Jaworski (1989), "Information Processing from Advertisements: Toward an Integrative Framework,"Journal of Marketing, 53 (4), 1-23.

McCullough, J. Lee, and Thomas M. Ostrom (1974), "Repetition of Highly Similar Messages and Attitude Change," Journal of Applied Psychology, 59 (3), 395-397.

McLeod, Douglas M., and Elizabeth M. Perse (1994), "Direct and Indirect Effects of Socioeconomic Status on Public Affairs Knowledge," Journalism Quarterly, 71 (2), 433-442.

McLuhan, Marshall (1964), "The Medium Is the Message," in Understanding Media, 6th ed., New York: McGraw-Hill.

Olson, Jerry C., Daniel R. Toy, and Philip A. Dover (1982), "Do Cognitive Responses Mediate the Effectiveness of Advertising Content on Cognitive Structure?" Journal of Consumer Research, 9 (December), 245-262.

Pechmann, Cornelia, and David W. Stewart (1988), "Advertising Repetition: A Critical Review of Wearin and Wearout," Current Issues and Research in Advertising, 11 (2), 285-329.

Petty, Richard E., and John T. Cacioppo (1979), "Issue Involvement Can Increase or Decrease Persuasion by Enhancing Message-Relevant Cognitive Responses," Journal of Personality and Social Psychology, 37, 1915-1926.

-- (1986), Communication and Persuasion: Central and Peripheral Routes to Attitude Change, New York: Springer-Verlag.

-- (1996a), Attitude and Persuasion: Classic and Contemporary Approaches, Boulder: Westview Press.

-- (1996b), "The Message-Learning Approach," in Attitudes and Persuasion: Classic and Contemporary Approaches, Boulder: Westview Press.

Putrevu, Sanjay, and Kenneth R. Lord (2003), "Processing Internet Communications: A Motivation, Opportunity, and Ability Framework," Journal of Current Issues and Research in Advertising, 25 (1), 46.

Riedman, Patricia (June 1999), "Marketers Score with Integration," in Vol. 2000: Advertising Age.

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 (Spring), 13-31.

Schiffman, Leon G., and Leslie Lazar Kanuk (1994), Consumer Behavior, 5th ed., Englewood Cliffs, NJ: Prentice Hall.

Steiger, J. H. (1989), EzPATH: Causal Modeling, Evanston, IL: SYSTAT.

Sundar, S. S., and C. Nass (1996), "Source Effects in Users' Perception of Online News," paper presented at the 46th annual conference of the International Communication Association, Chicago, May.

--, Sunetra Narayan, Rafael Obregon, and Charu Uppal (1997), "Does Web Advertising Work? Memory for Print Vs. Online Media," paper presented at the annual conference of the Association for Education in Journalism and Mass Communication, Chicago, August.

Tabachnick, Barbara G., and Linda S. Fidell (1996), Using Multivariate Statistics, 3rd ed., New York: Harper Collins College Publishers.

Tulving, Endel (1974), "Cue-Dependent Forgetting," American Scientist, 62, 74-82.

Zimbardo, Philip G., and Michael R. Leippe (1991), The Psychology of Attitude Change and Social Influence, New York: McGraw-Hill.

Yuhmiin Chang (Ph.D., University of Missouri-Columbia), is an assistant professor in the Department of Advertising and Public Relations, Fu-Jen Catholic University, Taiwan.

Esther Thorson (Ph.D., University of Minnesota) is associate dean for graduate studies and research at the School of Journalism, University of Missouri-Columbia.