行政院國家科學委員會專題研究計畫 成果報告
企業對企業平面廣告中代言人的效果：跨文化的比較
計畫類別： 個別型計畫 計畫編號： NSC91-2416-H-009-002- 執行期間： 91 年 08 月 01 日至 92 年 07 月 31 日 執行單位： 國立交通大學管理科學學系 計畫主持人： 黃仁宏 計畫參與人員： 林佳燕、董正枚、蔡雨潔 報告類型： 精簡報告 處理方式： 本計畫可公開查詢中 華 民 國 92 年 10 月 16 日
行政院國家科學委員會專題研究計畫成果報告
企業對企業平面廣告中代言人的效果：跨文化的比較
A Cross-Cultural Comparison of Endorser Effects in
Business-to-Business Print Advertising
計畫編號：NSC
91-2416-H-009-002-SSS執行期限：91 年 8 月 1 日至 92 年 7 月 31 日
主持人：黃仁宏 國立交通大學管理科學系
一、中文摘要 企業對企業廣告中，代言人是否運用 得當，對廣告效果有很大的影響。本研究 探討不同的廣告代言人，對廣告效果的影 響。本研究所探討的代言人分別為：公司 的最高階主管、研究部門主管、客戶公司 的主管和沒有任何代言人。因為代言人的 效果，可能受到文化的影響，因此本研究 比較台灣和美國的代言人效果。研究的對 象為美國和台灣的工程師，廣告則為高科 技機械設備的廣告。研究結果顯示，在美 國，客戶公司的經理人是最佳的選擇，而 在台灣，製造該產品公司的研發主管是最 佳的選擇。本文最後提供未來研究的建議。 關鍵詞：廣告、代言人、企業對企業、跨 文化 AbstractEndorsers can profoundly determine the effectiveness of business-to-business advertising. This study compares advertising effectiveness of four different types of endorsers: a company CEO, an R&D
manager, a manager of a user firm, and no endorser. Since endorser effects may vary across national boundaries, this study also compares endorser effects between two distinct cultures: the U.S. and Taiwan. Engineers from the U.S. and Taiwan are the subject of this study of high-tech machinery advertising. The results indicate that a manager of a user firm is the best choice as an endorser in the U.S., while an R & D manager from the advertiser’s company is more effective in Taiwan. Suggestions for future research are provided.
Keywords: Advertising, business-to-business, endorser, cross-culture
二、目的
Advertising effectiveness has been a concern of business-to-business advertisers and a subject of inquiry in many research studies (e.g., [1, 2]). Many factors of concern to business-to-business advertising have been examined, such as identifying the ad’s objectives and target markets [3],
communications effects [4], dimensions of an effective print ad [1, 2], layout styles [5],
color usage [6, 7], sexism [8] and appeals from cross-cultural perspectives [9].
However, to our knowledge, endorser effects for business-to-business advertising have not been examined.
Endorsers in consumer advertising are used quite extensively, and the literature is rich in studies that have assessed the effectiveness of various communication sources. The effectiveness of an endorser has been examined from the perspective of source credibility [10, 11], source physical attractiveness [11, 12], source expertise [11, 13], and celebrity status [14]. Many theories such as the Elaboration Likelihood Model [14], associative learning [15], social adaptation theory [12], attribution theory [16], cultural-meaning transfer [17], and
identification [18] have been advocated to explain the effects of sources of
communications.
In business-to-business advertising, endorsers are used less frequently. An examination of popular business and trade magazines quickly demonstrates that only a small percentage of the business-to-business ads use endorsers. In trade magazines, quite a few ads for components and instrument products feature mainly pictures of products. Is business-to-business advertising
employing only straightforward presentation and reasoning more effective than using an endorser? What types of endorsers are more effective if they are effective at all? One of the purposes of this study is to answer these questions.
Culture is likely to play an important role in the effectiveness of endorsers in business-to-business advertising. Source attractiveness, source credibility, source expertise and status carry different meanings across cultures [19, 20]. For example, all cultures stereotype on the basis of physical attractiveness but the content of a stereotype relies on cultural values [21, 22]. An attractive model in an ad may be perceived as more concerned for others in a collective society, while perceived as higher in potency in a more individualistic society. The same type of endorser is likely to be perceived to have different levels of attractiveness, trustworthiness, and expertise and will result
in having different impacts on the effectiveness of advertising in different cultures. Hence, the other purpose of this study is to compare the effectiveness of endorsers across two very distinctly different cultures, specifically between the U.S. and Taiwan.
三、文獻探討和假設
Endorser Effects
Four major types of endorsers are: celebrities, experts, company CEOs and typical users. The celebrity endorser is any individual who enjoys public recognition and who uses this recognition on behalf of a company and/or a product by appearing with it in an ad [17]. Sports heroes, TV
personalities and movie stars are celebrities used quite frequently on TV advertising; for example, Michael Jordan for Nike and McDonald’s, Candice Bergen for MCI long distance service and Tiger Wood for Buick automobiles. An expert is a recognized authority on the product class endorsed whose expertise is considered superior to general users. Examples of expert
endorsers are dentists for toothpaste, and R & D managers for machinery. The company CEO is the person in charge of the company whose product is being advertised.
Examples of CEOs as endorsers include Lee Iacocca for Chrysler automobiles and Frank Perdue for Perdue chickens. In this study, ads with an R & D manager as the expert, ads with the company CEO, ads with a manager of a user company and ads without an endorser are examined. Since employing a celebrity for particular industrial machinery demands a quite different approach in terms of the design of the picture and the message of the ad, use of a celebrity as an endorser is not examined in this study.
Research examining source persuasiveness has been extensive.
However, many studies aim at explaining the structure of source credibility and use only descriptive scenarios. Far less effort has been made to identify types of endorsers versus no endorser on credibility.
with endorsers and without endorsers. They found that ads with celebrity endorsers have higher credibility. Respondents consider the ads with celebrity endorsers as more distinctive, easily being recalled, conveying their messages easily, and as being more attractive. Friedman et al. [24] examined the effect of a celebrity, typical consumer, professional expert and company president used as endorsers on the expected selling price, probable taste, believability and intention of purchasing of wine. The use of an actor and company president had better results than those ads which used experts and general consumers in terms of believability, probable taste and intention of purchasing. Atkin and Block [25] found that for alcohol products, ads featuring famous persons are highly effective with teenagers, as compared with almost-identical versions of an ad with a non-celebrity. The celebrity figure is
perceived as more competent and trustworthy. However, the impact of famous endorsers on older persons is limited. Freiden [26] examined the effects of endorser type and gender by two audiences on TV sets and found that the CEO and the expert fair better in terms of perceived endorser knowledge and product quality for student subjects, but not for adult sample. Rubin et al. [27] examined the effectiveness of ads using a company president as endorser versus an unknown endorser in television commercials for a furniture store. They found that the company president was more effective than unknown endorser, especially in the
dimension trustworthiness. The results generally indicate that using a celebrity, CEO or expert is more effective than using an unknown person or lacking any person.
Many researches point to the
effectiveness of celebrity endorsers used in consumer ads in several dimensions. Celebrities make ads attention-gaining and entertaining [28], believable [29], aid in recognizing the brand name [14], and enhance message recall [30]. Physical attractiveness of message sources influences attitude and purchase intentions [12].
Ohanian [31] and O’Mahony and Meenaghan [28] found that when celebrity endorsers are perceived to have expertise, they increase the
purchase intentions. Wilson and Sherrell [32] found that source treatments account for 9% of explained variance among studies reporting significant findings. Agrawal and Kamakura [33] even showed that the
announcement of celebrity endorsement increases stock return.
The effectiveness of an endorser is moderated by the involvement of the product in the ad. Several studies found that for a low-involvement product, source credibility is the primary cue for credibility of the communication; while for high-involvement products, source credibility and effectiveness of advertising are not positively related [34]. This is consistent with the Elaboration
Likelihood Model [35], which proposed two distinct routes to persuasion. When
recipients are highly motivated, attitudes are formed and changed through the central route to persuasion. Advertising messages in the form of attribute description and reasoning evoke central route processing and will be more effective than peripheral cues. On the other hand, when recipients are unmotivated, attitudes are formed and changed through the peripheral route of persuasion. The
message source or endorser characteristics are used as peripheral cues that affect the effectiveness of persuasion. Since many industrial products are highly involved products, endorsers as peripheral cues are expected to be ineffective. This may explain the dearth of research of endorser effects in academia and the much less frequent use of endorsers in practice in business-to-business advertising.
However, the effectiveness of endorsers in ads may depend on the amount of
information displayed. A magazine ad for industrial products serves the purpose of informing, increasing preference and stimulating the interest of obtaining further information. The ad may not contain as much information as the readers would like to know. At this time, the endorser may provide important cues for the advertised products. In other words, although endorsers are usually used to invoke
processing through peripheral routes, under certain conditions endorsers may also invoke central route processing. Hence, the
endorser can still enhance the effectiveness of business-to-business ads for highly involved products. In consideration of the expert source, Kelman’s [36] contention that persuasion through the internalization process is applicable to business-to-business ads. This perspective is consistent with social adaptation theory [12], which implies that the adaptive significance of information will determine its impact. Information based on salience may be processed, and its influence may be based on usefulness for adaptation. If the information is useful, a considerable amount of time will be used to process the information.
Heath et al. [37] provide another support for the possibility that endorsers may be
effective for business-to-business advertising. Their experiments show that an endorser
does not influence attitudes in
noncompetitive settings when consumers engage in issue-relevant thinking. However, an endorser proved effective in the context of competition when brands were homogenous or characterized by large price-quality
trade-offs. Their results imply that decision difficulty can defuse the effects of
substantive reasoning, increase indecision, and promote perceptual contrast. These effects then empower an endorser to serve as an heuristic and reduce the risk of
post-preference regret. Similarly, a new and highly priced industrial product involves quite a lot of decision intricacy; the effects of an endorser can increase substantially.
H1: Using an endorser in business-to-
business advertising will create more favorable attitudes toward (a) the ad, (b) the advertised product, and (c) create higher purchase intention than similar advertising without an endorser.
The Effectiveness of Different Types
of Endorsers
McCracken [17] considered that an endorser may carry a variety of meanings, such as distinctions of status, class, gender, age, etc. An endorser is something richer and more complicated than a merely credible
or attractive individual. The success of the endorsement process hinges on the transfer of meaning from the endorser to the product, and then to the customer. The question then is what type of endorser can transfer a
particular meaning more effectively. Most previous research examines the influence of attractiveness and expertise on the
effectiveness of an endorser. These
researches usually vary the attractiveness and expertise of the endorsers by using
manipulative treatments before the experiment. However, this may not be consistent with the ways by which a reader processes an ad. In reading a magazine, the reader is not given a lot of information about the endorser in an ad sufficient to establish the desirable level of expertise of the endorser. The present research does not manipulate these two variables before the experiment; instead, only the title of the endorser varies. The variations of the pictures of the endorser are kept at the minimum. The persuasive impact of attractiveness and source expertise was independent and additive [38]. Thus, one question the present study seeks to examine comes down to “What title of the endorser will be perceived as more credible, that is, being more trustworthy and having more expertise?”
Expertise is defined as the perceived ability of the source to make valid assertions, one who knows the correct stand on an issue, or one whose statements have been verified empirically [38] . Expertise has the greatest impact on respondents’ reactions to celebrity endorsements [11, 28, 38]. Wilson and Sherrell [32] in a meta-analysis found that expertise has the greatest effect on
persuasion. The expert versus non-expert manipulation accounts for an average of 16% of the variance. However, the question is: “Who has the most expertise to recommend a machine as perceived by readers?”
For many industrial products, especially for advanced high-tech products, the R & D manager can be considered as an expert. The CEO of the manufacturing company may also be a candidate for the endorser. Hence, within the company that develops the product, the CEO and the R & D manager
may be used as endorsers. The CEO as an endorser may convey an impression other than those of having advanced technology, such as power, prestige, service and
commitment; while the R & D manager may give the impression of expertise and
credibility of the advanced technology of the machine. Another candidate for
endorsement is a manager from a user company, who has used the product in a real environment. In other words, the endorser conveys the message that the benefits of the machine have been empirically tested. On the other hand, an R & D manager, involved in research and development, may possess skills in advanced technology, but lack of the hands-on experience in running a day-to-day operation. Hence, in the U.S., in which a culture emphasizes empirical evidence [39], a manager from a user company is likely to be considered as having more expertise for recommending a machine. Furthermore, target audience can identify with the manager of the user company better than the other types of endorsers.
Several studies point to the important role of identification as a mediator of celebrity effects. Basil [18] suggests that a spokesperson with whom the audience identifies ensures the greatest likelihood of achieving lasting attitudes or behavior changes. The basis of communication effectiveness for drama was an audience member’s identification with a fictional character. A person perceives that he or she shares values or perceptions of reality with the character, and a bond is established with that character. The audience members are more likely to be affected by the performance. Kelman (1961) theorized that an attractive source influences attitude changes through identification processes. McGuire [40] also proposed that sources who are similar to the consumer are attractive and persuasive. Bandura [41] proposed social cognitive theory, which states that a person’s likelihood of enacting a behavior depends on that person’s identification with the model. When people perceive themselves as similar to the source in terms of attitudes, opinions, activities, background, social status or lifestyle, they are more likely to achieve
identification and enact whatever behavior is modeled by that person [42]. Comparing with other types of endorsers, a manager from a user company can achieve a higher level of identification with a target audience for a machine for two reasons. First, the audience and the endorser are similar in that they all use the machine for doing their jobs. Second, their companies are similar to each other because both companies have to use the same machines. Hence, a manager from a user company is likely to be more effective than other types of endorsers in the U.S. Furthermore, a source may be most effective when he or she is slightly better rather than very superior, as when children are more influenced by those just a little older than themselves than by age peers or much older children [40].
H2: In the U.S., using a manager from a customer firm as an endorser in advertising will create more favorable attitudes toward (a) the ad, (b) the advertised product, and (c) create higher purchase intentions than do other types of endorsers.
The Influence of Culture on the
Effectiveness of Endorsers
Most of the research in source persuasiveness was conducted in the U.S, using U.S. consumers as the subjects for the studies. The generalizability of the results to other cultures is questionable. Hofstede’s [39] cultural dimensions are employed here to explain source persuasiveness in different cultures. Cultural differences could be captured in four dimensions: power distance, individualism-collectiveness,
masculinity-femininity, and uncertainty avoidance. Since the four dimensions were developed in a western culture, some
dimensions unique to oriental cultures may be missing. A group of scholars [43] further identified another dimension named “Confucian Dynamism,” which is long-term versus short-term orientation. Hofstede’s framework has been discussed and employed in marketing [44], and specifically in
five dimensions, i.e., power-distance,
individualism-collectiveness, and uncertainty avoidance, are related to the issue of
persuasiveness.
The basic issue involved in power distance is human inequality. Inequality can occur in areas such as prestige, wealth, knowledge, and power. Power distance refers to the less powerful members of
organizations and institutions (like the family) accepting and expecting an unequal
distribution of power. The inequality is defined from below, not from above,
suggesting that a society’s level of inequality is endorsed by its followers as much as by its leaders.
It is likely that larger inequalities in power are reflected in larger inequalities in other areas. Such differences in other areas in turn feed back into power distance and reinforce it. We can therefore expect to find that differences between societies in the power distance norm might be associated with a difference in their response to ads with endorsers. In a high power-distance country, persons in high positions and experts such as a CEO and an R & D manager are well respected and are likely to be regarded as having expertise, which in turn increase the persuasiveness of the ad.
The dimension of individualism versus collectivism refers to the degree to which individuals are integrated into groups. In an individualist society, ties between individuals are loose. Everyone is expected to take care of oneself. In a collectivist society, strong family and extended family ties exist. People in the group are expected to look after each other. Collectivism here does not refer to the political system; it refers to the social group, not the country.
Some past research compared
advertising effectiveness across cultures from the perspective of the
individualism-collectiveness dimension. Han and Shavitt [49] conducted a content analysis of existing magazine ads to reveal that ads in the U.S. employ individualistic appeals more frequently than do those in Korea. Their experiment found that
advertising appeals vary in their effectiveness across cultures. Appeals that emphasize
individualistic benefits are more persuasive in the United States than in Korea, while appeals stressing family or in-group benefits are less persuasive in the United States than they are in Korea. Differences in cultural orientation also influence perceptions of the in-group versus out-group, attribution styles, patterns of emotions and behaviors [45].
The dimension of individualism versus collectivism is related to locus of control, which refers to the belief that rewards, reinforcements or outcomes in life are controlled either by one's own actions or by outside forces such as fate and power [48, 50]. Persons with internality drives locus of control believe that outcomes are influenced chiefly by personal factors, implying
individualism. At the opposite extreme, externality is the belief that outcomes are caused by external factors, meaning
collectivism. People with internality drives believe in individual decisions. Everyone is considered to have the right to and to have to hold personal opinions. For a highly
involved product in an individualistic society, a careful reasoning process and personal opinion should be given. In a collective society, opinions of persons in high positions and opinions of experts weigh heavily in the judgment. Hence, we would predict that a CEO or expert such as an R & D manager as endorser would be effective in a collective society, but ineffective in an individualistic society.
The other dimension that may be related to endorser effects is uncertainty avoidance. Uncertainty about the future is something with which people in all cultures must cope. People try to cope with uncertainty by
employing technology, law, and religion. Tolerance for uncertainty varies considerably between people in different countries. In a society in which avoidance of uncertainty is strong, people tend to be more resistant to change, more fearful of failure, and less likely to take risks. They consider conflict to be undesirable and have less tolerance for criticism than people scoring low on
uncertainty avoidance.
People in high uncertainty avoidance cultures tend to look for security, and to avoid risk taking. They tend to search for
ultimate, absolute truths and values, and not relativism and empiricism. They believe in experts [39]. Dawar et al. [51] found that individuals from high uncertainty avoidance cultures were more likely to seek product information from personal sources than impersonal sources, relative to individuals from low uncertainty avoidance cultures. Similarly, Pornpitakpan and Francis [48] found that individuals from high uncertainty avoidance cultures are more influenced by source expertise, relative to individuals from low uncertainty avoidance cultures. Thus, we can predict that experts such as an R & D manager as an endorser would be effective in a high uncertainty avoidance society. A manager from a user company who actually has used the product in a real environment will be judged as having the expertise to recommend the machine. However, an R & D manager, involved in research and
development, possessing skills in advanced technology, will be considered as having even more expertise than a manager from a user company in recommending the machine. Hands-on experience
in using the machine and empirically verifying the superiority of the machine is not
considered important; instead, the level of
professional achievement is the focus. Hence, in Taiwan, a high uncertainty avoidance culture, a manager from a user company would be considered to have less expertise for recommending a machine than an R & D manager, and to be less effective. H3: In a high power-distance, collective and
high uncertainty avoidance culture, an R & D manager as endorser in advertising will create more favorable attitudes toward (a) the ad, (b) the product, and (c) create higher purchase intention than do other types of endorsers
四、研究方法
Design and Subjects
The experiment employed a 2 (Taiwan/US) x 4 (CEO/expert/user/no endorser) between-subjects factorial design. An R & D manager is utilized and considered to be an expert, while a manager from a user company is employed as a user. Engineers were asked to participate in this research without any compensation. A total of 120 and 225 usable responses were obtained from the U.S. and Taiwan respectively. The engineers from Taiwan work in Taiwan Semiconductor Manufacturing Corporation (TSMC) while the US engineers work in a plant of Wafertect, a US subsidiary of TSMC. Such a homogeneous sample facilitates
comparison across cultures and is desirable for theory-testing studies. The design and the number of respondents within each cell are given in Table 1.
TABLE 1
The Experimental Design and the Number of Respondents in Each Cell
Stimuli
The product in the ad is a laser repair machine used in the process of making semiconductor chips. Engineers in a fabrication plant know this type of machine and may engage in the process of choosing a machine. The name of the manufacturer, the model name of the product and endorsers are fictitious. The ads with endorsers include a headline, a picture of the endorser, a picture of the machine, and copy stating the benefits of this machine. The picture of the endorser accounts for about 1/3 of the area of the ad. The picture of the R & D manager is the same as the picture of the manager from a user firm. Under the picture of the endorser, the name and the title of the
Endorser Country CEO Expert User No endorser Total
Taiwan 47 54 67 57 225
endorser and the messages of the endorsers were given. The picture of the machine is placed on the left-bottom corner of the ad. The benefits of the machine are list on the right of the picture of the machine. In the ad without endorsers, the machine and the copy listing the benefits of the machine in the ad are enlarged. Since English is the
official language of the company, the engineers in the Taiwan plant have the capacity to read the language. To avoid error from translation, the contents of the ads and the questionnaire are not translated into Chinese. This is consistent with the practices adopted by many U.S. and European firms, who simply place their English version of their ads in local magazines. Furthermore, many U.S. and European firms operating in Taiwan show their English versions of promotional
materials to engineers, since the total number of people in the island involved in the
purchase is too small to justify producing local promotional materials. Furthermore, people involved in the purchase may prefer to read specifications of the machine in English.
Procedures
A document explaining the purpose of the study, ads and a questionnaire were put into a booklet. Engineers from the Taiwan plant were conveniently selected and
randomly place into one of the four cells of the experimental design. The respondents were asked to read the document and the ad, and then answer the questionnaire. The U.S. sample and the Taiwan sample were selected and
administered in the same way.
Dependent
Variables
The questionnaire consists of five sections. The first section measures respondents’ attitude toward the endorser. This section consists of 15 items,
encompassing three dimensions:
attractiveness of the endorser, trustworthiness of the endorser and expertise of the endorser.
The items are taken from Ohanian [31]. The second section, consisting of 15 items, measures respondents’ attitude toward the ad and are taken from Atkin and Block [25]. The third section measures respondents’ attitude toward the product in the ad. This section consists of four items which are taken from Atkin and Block [25]. The fourth section measures respondents’ purchase intention and are taken from Rubin, Mager, and Friedman [27]. All attitude and intention items are measured with Likert scales fro 1 to 7, with 1 on the scale being “strongly disagree” and 7 being “strongly agree”. The fifth section solicits
demographic information about the respondent, including sex, nationality, and years as an engineer.
The Cronbach alpha of the measures are as follows: .8570 for attitude toward the attractiveness of the endorser, .9248 for attitude toward the trustworthiness of the endorser, .9604 for attitude toward the expertise of the endorser, .9524 for attitude toward the ad, .8855 for attitude toward the product, and .7334 for purchase intention, indicating that the measures have internal consistency.
The nationality, gender and years as engineer of the respondents are listed on Table 2. Males account for a higher percentage in Taiwan than that in the U.S.
TABLE 2
The Nationality, Gender and Years as Engineer of the Respondents
六、結果
Attitude toward the Endorser
The 15 items concerned with
respondents’ attitude toward the endorsers are first factor analyzed and then rotated. The results showed three distinct dimensions Gender Years as Engineer
Male Female 0.5~4 5~8 9~12 Over 12 Taiwan 170 55 143 63 17 2
that are related to attractiveness,
trustworthiness, and expertise of the endorser. Subsequently, each item loading on the same dimension is averaged to obtain the overall measure of that dimension. The means and standard deviations for each treatment are shown on Table 3. ANOVA analysis of each dimension is shown on Table 4. Figure 1 shows the mean attitude as a function of country and endorser type.
TABLE 3
Means and Standard Deviations of Attitude toward Endorsers
The number with a frame indicates that the number is the highest on the row as expected. Items are measured using a scale from 1 to 7, with 1 being “strongly disagree” and 7 being “strongly agree”.
* Significant at alpha = 0.05 level when comparing with the user treatment on the row for the U.S. sample and comparing with the expert treatment on the row for the Taiwan sample.
‡ Significant at alpha = 0.1 level when comparing with the user treatment.
On the attractiveness dimension, country, endorser type and the interaction effects between country and endorser type is insignificant. Separate analysis of the
Taiwan and the U.S. samples also shows that there is no difference in attractiveness among the three types of endorsers. Since the models in the ads were maintained to be as close as possible and the picture of the manager from a user company and the picture of the R & D manager are the same, we would expect that types of endorser do not differ much in the attractiveness dimension.
On the trustworthiness dimension, endorser type is significant at the alpha = .1 level, while country and the interaction between country and endorser are not significant. Both U.S. and Taiwan respondents considered the user to more trustworthy than the expert, who in turn is more trustworthy than the CEO. For the
U.S. sample, by examining the means, the user is perceived as more trustworthy than the CEO or the expert, although the differences are not significant. For the Taiwan sample, ANOVA analysis shows that the effect of endorser type is significant at the alpha = .05 level. Pairwise comparison, using a Bonferroni adjustment, shows that the effect of the user as endorser in creating trustworthiness is significantly higher than using the CEO as endorser at the alpha = .1 level. Although the difference in means between user and expert is not significant at the alpha = .1 level, it is close to this level. Hence, we can conclude that the user is considered more trustworthy than other types of endorsers for the Taiwan sample, and the U.S. sample, though not statistically
significant, is moderately so.
Endorser Country CEO Expert User
Taiwan 3.19 (1.04) 3.25 (.87) 3.16 (.14) Attractiveness of Endorser U.S. 3.07 (0.87) 3.02 (.88) 3.37 (1.10) Taiwan 3.53* (1.09) 3.99 (.97) 4.10 (1.18) Trustworthiness of Endorser U.S. 3.79 (1.02) 3.67‡ (.98) 3.98 (1.06) Taiwan 4.06 (1.20) 4.31 (1.16) 4.19 (1.48) Expertise of Endorser U.S. 3.77 (1.27) 3.36* (.88) 3.90 (1.18)
TABLE 4
Effects of Attitude toward Endorser
FIGURE 1
Mean Attitude toward the Endorser as a Function of Country and Endorser Type
Attractiveness of Endorsers
2 2.5 3 3.5
CEO Expert User
Taiwan U.S. Trustworthiness o 2 2.5 3 3.5 4 4.5 CEO Exp Taiwan Expertise of Endorsers 2 2.5 3 3.5 4 4.5
CEO Expert User
Taiwan U.S.
On the expertise dimension, country effect is significant at the alpha = .01 level. It appears that the endorsers are perceived to have more expertise by Taiwan respondents than that perceived by their U.S. counterparts. The means indicate that Taiwan respondents considered the R&D manager to have Source Df Mean Square F Sig.
Country 1 .122 .119 .730 Endorser 2 .489 .475 .622 Attractiveness of Endorser Country *Endorser 2 1.022 .992 .372 Country 1 .197 .174 .677 Endorser 2 3.177 2.799 .063 Trustworthiness of
Endorser Country * Endorser 2 1.813 1.597 .204 Country 1 15.851 10.116 .002 Endorser 2 .907 .579 .561 Expertise of
slightly more expertise than other types of endorsers, while U.S. respondents
moderately regard the user as having more expertise. ANOVA analysis shows that there is no significant difference among different types of endorsers for the Taiwan sample and for the U.S. sample, separately. However, the t-test of the means of the R&D manager between the Taiwan sample and the U.S. sample was significant at the alpha = 0.01 level. That an R & D manager is judged to be somewhat more lacking in expertise for recommending the machine than other types of endorsers for the U.S. sample is a surprise.
Overall, there are differences and similarities in attitudes toward endorsers between the two countries. Respondents from both countries considered the endorsers as having about the same degree of
attractiveness. Respondents from both countries also considered the user to be more trustworthy than other types of endorsers. However, U.S. respondents considered the user to have somewhat more expertise; while Taiwan respondents considered an R & D manager to have slightly more expertise. In terms of endorser’s trustworthiness, and expertise, the user as endorser is the clear choice for the U.S. target. Since expertise is the most important dimension for an endorser, an R & D manager seems to be a better
choice for a Taiwan audience.
Attitude toward the Ad
Attitudes toward the ad are measured with fifteen items. These items are factor analyzed. The results indicate that there is only one dimension. The average of the 15 items is used as the dependent variable, which is shown on Table 5. The mean attitude toward the ad as a function of
country and endorser type is shown on Figure 2. An ANOVA analysis with country and type of endorser as independent variables is conducted. The results are shown on Table 6. For the U.S. sample, respondents
considered the ad with the user as endorser to be the most effective. ANOVA analysis of the U.S. sample indicates that the type of endorser affects significantly respondents’ attitude toward the ad at alpha = .01.
Pairwise comparison, with no adjustment, shows that the effect of the user as endorser in creating a favorable attitude is not significantly higher than using the R & D manager and using the CEO, due to the small sample size. However, by examining the means, using the user as endorser creates the most favorable attitude toward the ad, and the size of the difference is considerable. In the Taiwan sample, respondents consider the ad with an R & D manager as endorser to be the most effective. ANOVA analysis of the Taiwan sample indicates that the type of endorsers affects significantly respondents’ attitude toward the ad at alpha = .01.
Pairwise comparison, without any adjustment, shows that the effect of the expert as
endorser in creating a favorable attitude is significantly higher than using other types of endorsers at alpha = .01. Adjustment for multiple comparisons is considered
unnecessary, since we are testing hypothesis, not comparing all possible pairs or
comparing with a control treatment. The difference in means between using any type of endorser can create a significantly more favorable attitude toward the ad than that of using no endorsers for both the U.S. and Taiwan samples.
Attitude toward the Advertised
Product
Attitude toward the advertised product was measured with four items. The average of the four items is used as the dependent variable, of which means and variances are shown on Table 5. An ANOVA analysis with country and type of endorser as independent variables is conducted. The results are shown on Table 6. The mean attitude toward the product as a function of country and endorser type is shown on Figure 2.
TABLE 5
Means and Standard Deviations for Attitude toward Ads,
The number with a frame indicates that the number is the highest on the row as
expected.
Items are measured using a scale from 1 to 7, with 1 being “strongly disagree” and 7 being “strongly agree”.
* Significant at alpha = 0.05 level when comparing with the user treatment on the row for the U.S. sample and comparing with the expert treatment on the row for the Taiwan sample.
The type of endorser and the interaction effect of endorser type and country are significant. In the U.S. sample, respondents consider the product in the ad with the user as endorser to be the most effective.
ANOVA analysis of the U.S. sample indicates that type of endorsers affect
significantly respondents’ attitude toward the product at the alpha = .01 level. Pairwise comparison shows that the effect of using any endorser in creating favorable attitude toward the product is significantly higher than using no endorser at the alpha = .01 level. For the Taiwan sample, respondents consider the advertised product with an R & D manager as endorser to be the most favorable. ANOVA analysis of the Taiwan sample indicates that type of endorsers affect significantly respondents’ attitude toward the product at alpha = .01. Pairwise
comparison shows that the effect of the expert as endorser in creating a favorable attitude toward the product is significantly higher than using the user or no endorser at
alpha = .05. Overall, using the R & D manager as endorser created a more favorable attitude toward the product than using other types of endorsers for the Taiwan respondents, while using any endorser generated a more favorable attitude toward the product than using no endorser for the U.S. sample.
Purchase Intention
Intention to purchase the advertised product was measured with four items. The average of the four items is used as the dependent variable, of which means and variances are shown on Table 5. An ANOVA analysis with country and type of endorser as independent variables is conducted. The results are shown on Table 6. The mean attitudes of intentions to purchase the product as a function of country and endorser type are shown on Figure 2. The type of endorser and the interaction effect of endorser and country are significant. For the Taiwan sample, respondents have the highest intention of purchasing the product in the ad with an R & D as endorser. However, ANOVA analysis of the Taiwan sample indicates that type of endorsers do not affect significantly respondents’ intention to purchase the product. For the U.S. sample, respondents have the highest intention of purchasing the product in the ad with the user as endorser. ANOVA analysis of the U.S. sample indicates that type of endorsers affect significantly respondents’ intention to
purchase the product at alpha = .01.
Pairwise comparison shows that the effect of Endorser Nationality CEO Expert User No endorser
Taiwan 3.05* (.74) 3.54 (.98) 3.09* (.99) 2.95* (.84) Attitude toward Ad U.S. 3.43 (.75) 3.29* (.79) 3.72 (1.11) 2.35* (1.55) Taiwan 3.42 (1.02) 3.71 (1.10) 3.28* (1.20) 3.22* (.90) Attitude toward Product U.S. 3.74 (.96) 3.73 (1.22) 3.88 (1.20) 2.25* (1.59) Taiwan 3.87 (1.07) 3.98 (.88) 3.67 (1.46) 3.94 (1.21) Intention to Purchase U.S. 4.32 (1.20) 3.63 (1.40) 4.19 (1.49) 2.33* (1.72)
the user as endorser in creating purchase intention is not significantly higher than using other types of endorsers. Using any type of endorser can create a significantly more favorable attitude toward the product than using no endorsers at the alpha = .01 level. Overall, using an R & D manager as endorser does not create more intention to purchase the product than using other types of endorsers for the Taiwan sample, while using the user as endorser does not generate higher intention to purchase the product than using other types of endorsers in the U.S. sample. In the Taiwan sample, the result of purchase intention is different from the result of other attitude measures in that the ad with no endorser creates a relatively high intention to purchase the product. This is not as expected and deserves careful scrutiny.
TABLE 6
Effects of Attitude toward Ad and Product
六、結論和討論
There are differences as well as
similarities between the two cultures in terms of effectiveness of endorsers. In both cultures, an ad without any endorser often gets the worst results, as indicated in Table 5. The only exception is intention to purchase for the ad with no endorser. Even in an ad for a machine, the persuasiveness of the ad containing only the machine and the substantiate advantages of the machine is minimal. Regardless of culture, depicting a real human with a title for a cool machine can increase the effectiveness of persuasion. Thus, hypothesis H1 that using an endorser in
a business-to-business ad is more effective
than a similar ad without an endorser is generally supported.
Endorsers who are perceived to be more credible tend to be more persuasive, i.e., increasing the preference for the ad and the product. For example, using the R & D manager consistently delivers the highest score for attitude toward the endorser,
attitude toward the ad and attitude toward the product in the Taiwan sample. Sternthal et al. [52] categorized subjects on the basis of their favorability toward the position advocated in a communication. Subjects were then presented a persuasive message attributed to either a highly credible or moderately credible source. The highly credible source was found to be more
FIGURE 2
Mean Attitude toward the Ad and Product, and Intention to Purchase as a Function of
Country and Endorser Type Source df Mean Square F Sig.
Country 1 .134 .153 .696 Endorser 3 7.470 8.498 .000 Attitude toward Ad
Country * Endorser 3 5.430 6.177 .000
Country 1 0.005 .004 .949
Attitude toward Product Endorser 3 11.524 9.405 .000 Country * Endorser 3 7.162 5.846 .001
Country 1 4.490 2.743 .099
Intention to purchase Endorser 3 10.786 6.590 .000 Country * Endorser 3 15.346 9.376 .000
Attitude toward Advertisement 2 2.5 3 3.5 4
CEO Expert User No Endorser
Taiwan U.S.
Attitude toward Product
2 2.5 3 3.5 4
CEO Expert User No Endorser Taiwan U.S. Intention to Purchase 2 2.5 3 3.5 4 4.5
CEO Expert User No Endorser
Taiwan U.S.
persuasive than the moderately credible source among subjects opposed to the message position. In contrast, the less credible source was more influential for subjects favoring the message stance. Harmon and Coney [53] also found that a highly credible source was more effective than a moderately credible source when the message was viewed unfavorably. The
moderately credible source was more
persuasive when the message advocated was viewed favorably. Based on the findings of these previous researches, we can predict that the more trustworthy endorser would be more effective in an ad for machinery for which respondents do not have a favorable predisposition. The results of this study support this contention.
There exist tremendous differences between the two cultures. What type of endorsers should be used in an ad is culturally bound. For the Taiwan sample, an R & D manager from one’s own company is a good choice as an endorser for the ad. Attitude toward the ad and attitude toward the product are significantly better than other types of endorsers or no endorsers at all. Hence, hypothesis H2a and H2b are supported.
For the U.S. sample, using a manager from a user company is a good choice for endorser in an ad. Attitude toward the ad is
significantly better than when other types of endorsers or no endorsers are used. Hence, hypothesis H3a is supported. Although the
data fails to support H3b statistically, the
relative magnitudes of the means are as hypothesized.
The data fails to support the contention that an R & D manager as an endorser would increase purchase intention of the Taiwan respondents. It also fails to support the contention that a manager from a user firm as an endorser would increase purchase
intention of the U.S. respondents. Thus, H2c
and H3c are not supported. This
substantiates previous literature that attitude and action tendency are different dimensions. Despite these results, a manager from a user firm, capable of creating a more favorable attitude toward the endorser and toward the ad, and no worse in other effectiveness measures is still a good choice as an endorser for the U.S market. Since using an R & D manager creates a more favorable attitude toward the endorser, toward the ad, and toward the product, and no worse in other effectiveness measures for the Taiwan respondents, an R & D manager is a better choice than other types of endorsers for the Taiwan market.
The score of the purchase intention for the ad with no endorser in the Taiwan sample is relatively high, as compared with other attitude measures. In other words, attitude measures for the group without any endorser are relatively low, while the purchase
intention scores high. Several studies have indicated that highly credible sources are less persuasive than a less credible source in a choice situation [34, 54]. This may be explained by using self-perception theory [55]. According to self-perception theory, internal causes and external messages do not sum up to determine ultimate motivation of choice, but rather both interact in such a way that external messages may undermine internal causes. In an effort to attribute causation to their behaviors, subjects under a highly credible source may be uncertain about the underlying causes of their
compliance. It may be attributed either to some internal causes or to the credible external sources. On the other hand,
subjects under the influence of a low credible source may feel at ease in taking actions. In the context of advertising with endorsers, subjects in the treatment with no endorser in the ad may feel that they are competent to take action, while those in the treatment with an expert endorser may feel uncertainty as to their acceptance of the product, due to the internal causes or to the credible source, hence, they may hesitate to take action.
The U.S. sample does not show that an ad without an endorser enhances choice. This result is quite different from the result obtained in the Taiwan sample. This may be due to the fact that the U.S. is a much more individualistic society and respondents did not feel that they are influenced by any authoritative figures in the first place. They feel that their choices are theirs to make. Self-perception theory thus does not apply in this situation.
Graham [54], in a research on coupon effectiveness, supports the contention of self-perception theory that high-value coupon offers may undermine brand repurchasing to a greater degree than do low-value coupon offers. Furthermore, the results also present evidence that allowing consumers to choose between two different value-level coupons
may increase the repurchase rates of the brand. If self-perception theory is applicable to advertising, then using an expert as an endorser and offering audiences choices may alleviate the audiences’ feeling of being controlled by credible sources in Taiwan. Future research in this avenue may prove to be fruitful.
For the present study, a few limitations point to the directions of future research. First, it is related to the issue of
generalization. The findings are limited to the product and the type of endorsers examined in this study. Future research with other types of endorsers and products can shield light on the generalizability of these findings. For example, what type of endorsers would be more effective in persuasion and soliciting the intention of purchasing in advertising for services in different cultures? Additionally, this study did not examine the impact of the celebrity endorser. Since this type of endorser is sometimes used in business-to-business image ads, examining the effectiveness of this type of endorser may prove rewarding from both an academic as well as a practical perspective. Furthermore, this study examines only two cultures. It is desirable to examine more than two cultures so that findings can be generalized [56].
Second, by using a fictitious brand name, unfamiliar model and exposing the ad only once, the effect of previous brand knowledge obtained through multiple exposures to ads, reading reports about the products and the endorsers are not taken into account. Future research may address these issues.
Despite the limitations, the present research clearly indicates that culture has a bearing on the effectiveness of endorsers for business-to-business advertising.
Advertisers have to adjust endorsers in their ads accordingly. Localization of
business-to-business advertising with different types of endorsers is a prudent approach.
七、計劃成果自評
本計劃依原計劃內容執行，搜集了台 灣和美國的資料，執行統計分析，並驗證
假設。本計劃具學術上和實務上的價值。
九、參考文獻
1. Hanssens, Dominique M., and Weitz, Barton A., The Effectiveness of Industrial Print Ads Across Product Categories, Journal of Marketing
Research 17, 294-306 (1980).
2. Lohtia, Ritu, Johnston, Wesely J. and Aab, Linda, Business-to-Business Advertising: What Are the Dimension of an Effective Print Ad?, Industrial
Marketing Management 24, 369-378
(1995).
3. Hartley, Steven W. and Patti, Charles H., Evaluating Business-to-Business
Advertising: A Comparison of Objectives and Results, Journal of
Advertising Research 25, 21-27 (1988).
4. Gilliland, David I. And Johnston, Wesley Jr., Toward a Model of Business-to-Business Marketing Communications Effects, Industrial
Marketing Management 26, 15-29
(1997).
5. Chamblee, Robert and Sandler, Dennis M., Business-to-Business Advertising: Which Layout Style Works Best,
Journal of Advertising Research 29,
39-46 (1992).
6. Clarke, Irvine, Ⅲ and Honeycutt, Earl D., Jr., Color Usage in International Business-to-Business Print Advertising,
Industrial Marketing Management 29,
255-261 (2000).
7. Huang, Jen-Hung, Color in U.S. and Taiwanese Industrial Advertising,
Industrial Marketing Management 22,
195-198 (1993).
8. LaTour, Michael S., Henthorne, Tony L. and Williams, Alvin J., Is Industrial Advertising Still Sexist, Industrial
Marketing Management 27, 247-255
(1998).
9. Albers-Miller, Nancy D. and Gelb, Betsy D., Business Advertising appeals as a Mirror of Cultural Dimensions: A study of Eleven Countries, Journal of
Advertising 25, 57-70 (1996).
10. Lafferty, Barbara A., and Goldsmith, Ronald E. (1999), Corporate
Credibility’s Role in Consumers’ Attitudes and Purchase Intentions: When a High versus a Low Credibility Endorser Is Used in the Ad, Journal of
Business Research 44, 109-116.
11. Ohanian, Roobina, The Impact of Celebrity Endorsers’ Perceived Image on Consumers’ Intention to Purchase,
Journal of Advertising Research 31,
46-54 (1991).
12. Kahle, Lynn R. and Homer, Pamela M., Physical Attractiveness of the Celebrity Endorser: A Social Adaptation
Perspective, Journal of Consumer
Research 11, 954-961 (1985).
13. Wood, Wendy and Kallgren, Carl A., Communicator Attributes and
Persuasion: Recipient’s Access to Attitude-Relevant Information in Memory, Personality and Social
Psychology Bulletin 14, 172-182 (1988).
14. Petty, Richard E., Cacioppo, John T., and Schmann, David, Central and Peripheral Routes to Advertising Effectiveness: The Moderating Role of Involvement, Journal of Consumer
Research 10, 135-146 (1983).
15. Till, Brian D. and Shimp, Terence A, Endorsers in Advertising: The Case of Negative Celebrity Information, Journal
of Advertising 27, 67-82 (1998).
16. Tripp, Carolyn, Jensen, Thomas D. and Carlson, Les, The Effects of Multiple Product Endorsements by Celebrities on Consumers’ Attitudes and Intentions,
Journal of Consumer Research 20,
535-547 (1994).
17. McCracken, Grant, Who Is the Celebrity Endorser? Cultural Foundations of the Endorsement Process, Journal of
Consumer Research 16, 310-321 (1989).
18. Basil, Michael D., Identification as a Mediator of Celebrity Effects, Journal
of Broadcasting & Electronic Media 40,
478-495 (1996).
19. Aaker, Jennifer L., Accessibility and Diagnosticity? Disentangling the Influence of Culture on Persuasion Process and Attitude, Journal of
20. Grier, Sonya A. and Brumbugh, Anne M., Noticing Cultural Differences: Ad Meaning Created by Target and Non-Target Markets, Journal of
Advertising 28, 79-93 (1999).
21. Wheeler, Ladd and Kim, Youngmee, What Is Beautiful Is Culturally Good: The Physical Attractiveness Stereotype Has Different Content in Collectivistic Cultures, Personality and Social
Psychology Bulletin 23, 795-800 (1997).
22. Shaffer, David R., Crepaz, Nicole, and Sun, Chien-Ru, Physical Attractiveness Stereotyping in Cross-Cultural
Perspective, Journal of Cross-Cultural
Psychology 31, 557-582 (2000).
23. Nataraajan, Rajan and Chawla, Sudhir K., ’Fitness’ Marketing: Celebrity or Non-Celebrity Endorsement?, Journal
of Professional Services Marketing 15,
119-129 (1997).
24. Friedman, Hershey H., Termini,
Salvatore, and Washington, Robert, The Effectiveness of Ads Utilizing Four Types of Endorsers, Journal of
Advertising 5, 22-24 (1977).
25. Atkin, Charles and Block, Martin, An Experiment revealed the Effectiveness of Celebrity Endorsers, Journal of
Advertising Research 23, 57-61 (1983).
26. Freiden, Jon B., Advertising
Spokesperson Effects: An Examination of Endorser Type and Gender on Two Audiences, Journal of Advertising
Research 24, 33-41 (1984).
27. Rubin, Vicki, Mager, Carol, and Friedman, Hershey H., The
Performance of Company President versus Endorser in Television
Commercials, Journal of Advertising
Research 22, 31-33 (1982).
28. O’Mahony, Sheila and Meenaghan, Tony, The Impact of Celebrity Endorsements on Consumers, Irish
Marketing Review 10, 15-24 (1997).
29. Kamins, Michael A, Brand, Meribeth J., Hoeke, Stuart A., and Moe, John C., Two-Sided Versus One-Sided Celebrity Endorsements: The Impact on
Advertising Effectiveness and
Credibility, Journal of Advertising 18, 4-10 (1989).
30. Friedman, Hershey and Friedman, Linda, Endorser Effectiveness by Product Type,
Journal of Advertising Research
19,63-71 (1979)
31. Ohanian, Roobina, Construction and Validation of a Scale to Measure Celebrity Endorsers’ Perceived Expertise, Trustworthiness, and Attractiveness, Journal of Advertising
19, 39-52 (1990).
32. Wilson, Elizabeth J. and, Sherrell, Daniel L., Source effects in
communication and persuasion research: A meta-analysis of effect size, Journal
of the Academy of Marketing Science 21,
101-112 (1993).
33. Agrawal, Jagdish and Kamakura, Wagner A., The Economic Worth of Celebrity Endorsers: An Event Study Analysis, Journal of Marketing 59, 56-62 (1995).
34. Dholakia, Ruby R. and Sternthal, Brian, Highly Credible Sources: Persuasive Facilitators or Persuasive Liabilities,
Journal of Consumer Research 3,
223-232 (1977).
35. Petty, Richard E. and Cacioppo, John T., Source Factors and the Elaboration Likelihood Model of Persuasion, in
Advances in Consumer Research, Vol.
11, T. C. Kinnear, ed. Provo, UT: Association for Consumer Research, 668-672 (1984).
36. Kelman, Herbert C., Processes of Opinion Change, Public Opinion
Quarterly 25, 57-78 (1961).
37. Heath, Timothy B., McCarthy, Michael S., and Mothersbaugh, David L., Spokesperson Fame and Vividness Effects in the Context of Issue-Relevant Thinking: The Moderating Role of Competitive Setting, Journal of
Consumer Research 20, 520-34 (1994).
38. Maddux, James E. and Rogers, Ronald W., Effects of source expertness,
Physical Attractiveness, and Supporting arguments on Persuasion: A Case of Brains Over Beauty, Journal of
Personality and Social Psychology 39,
235-244 (1980).
39. Hofstede, Geert (1980), Culture’s
Differences in Work-Related Values,
Sage, Beverly Hills, CA.
40. McGuire, William J., Attitudes and Attitude Change, in Handbook of Social
Psychology, Vol. 2, eds. Gardner
Lindzey and Elliot Aronson, Random House, New York, 233-346, 1985. 41. Bandura, Albert, Social Foundations of
Thought and Action: A Social-Cognitive Theory, Prentice-Hall, Englewood Cliffs,
N.J., 1986.
42. Aaker, David A, Batra, Rajeev, and Myers. John G., Advertising
Management, 4th ed., Prentice Hall,
Englewood Cliffs, N.J., 1992. 43. Chinese Culture Connection, Chinese
Values and the Search for Culture-Free Dimensions of Culture, Journal of
Cross-Cultural Psychology 18, 143-64
(1987).
44. Clark, Terry, International Marketing and National Character: A Review and Proposal for an Integrative Theory,
Journal of Marketing 54, 66-79 (1990).
45. Aaker, Jennifer L. and Maheswaran, Durairaj, The Effect of Cultural Orientation on Persuasion, Journal of
Consumer Research 24, 315-328 (1997).
46. Cutler, Bob D, Erdem, S Altan, and Javalgi, Rajshekhar G., Advertiser's Relative Reliance on
Collectivism-Individualism Appeals: A Cross-Cultural Study, Journal of
International Consumer Marketing 9,
43-55 (1997).
47. Leach, Mark P. and Liu, Annie H., The Use of Culturally Relevant Stimuli in International Advertising, Psychology
and Marketing 15, 523-546 (1998).
48. Pornpitakpan, Chanthika and Francis, June N. P., The Effect of Cultural Differences, Source Expertise, and Argument Strength on Persuasion: An
Experiment with Canadians and Thais,
Journal of International Consumer Marketing 13, 77-101 (2001).
49. Han, Sang-Pil and Shavitt, Sharon, Persuasion and Culture: Advertising Appeals in Individualistic and Collectivistic Societies, Journal of
Experimental Social Psychology 30,
326-350 (1994).
50. Mueller, Stephen L. and Thomas, Anisya S., Culture and Entrepreneurial Potential: A Nine Country Study Of Locus Of Control And Innovativeness,
Journal of Business Venturing 16, 51-75
(2001).
51. Dawar, Niraj, Parker, Philip M and Price, Lydia J., A cross-cultural study of
interpersonal information exchange,
Journal of International Business Studies 27, 497-516 (1996).
52. Sternthal, Brian, Dholakia, Ruby, and Leavitt, Clark, The Persuasive Effects of Source Credibility: Tests of Cognitive Response, Journal of Consumer
Research 4, 252-260 (1978).
53. Harmon, Robert R. and Coney, Kenneth A., The Persuasive Effects of Source Credibility in Buy and Lease Situations,
Journal of Marketing Research 19,
255-260 (1982).
54. Graham, Judy F., Increasing Repurchase Rates: A Reappraisal of Coupon Effects,
Psychology & Marketing 11, 533-547
(1994).
55. Bem, Daryl J., Self-Perception Theory, in L. Berdowitz ed., Advances in
Experimental Social Psychology, Vol. 6,
Academic Press, New York, 1972. 56. Samiee, Saeed and Jeong, Insik,
Cross-Cultural Research in Advertising: An Assessment of Methodologies,
Journal of the Academy of Marketing Science 22, 205-217 (1994)
