轉碼廣告的說服力—語言熟悉程度、消費者認同與產品類別的干擾效果

科技部補助專題研究計畫成果報告

期末報告

轉碼廣告的說服力—語言熟悉程度、消費者認同與產品類別的

干擾效果(第3年)

中　華　民　國　106　年　09　月　07　日

中 文 摘 要 ： 本三年期計畫欲逐步探討影響轉碼廣告效果的因素。所謂轉碼廣告 ，指的是在同一則廣告中插入外國語言於文案中的廣告，由於不同 語言會影響消費者的認知與偏好，因此，轉碼廣告的語言使用會影 響轉碼廣告的效果。 許多研究已證實轉碼廣告會影響消費者的廣告態度，然而，對於轉 碼廣告中語言的使用，如該使用何種語言?不同語言的配對與比例 ，或可能影響廣告效果的因素等相關研究仍較少，因此，本計畫以 三個子研究分別探討影響轉碼廣告效果的干擾變項—語言熟悉度、 消費者認同與產品類別，並探討其中的中介變數。本研究分三部分 進行探討: (1)驗證不同產品類別的轉碼廣告說服效果，並驗證該效 果是否能同時存在於非雙語精通者(non-bilinguals)與雙文化雙語 使用者(bicultural bilinguals)中；(2) 確認消費者認同對轉碼廣 告效果的影響效果；(3)驗證語言熟悉度對轉碼廣告的干擾效果，同 時提出語言理解度為其中重要的中介變數。 藉由理解影響轉碼廣告的重要因子與其中的影響歷程可以幫助理解 消費者心理，並能提供於採用轉碼廣告時的實務建議，其研究結果 能貢獻於消費者行為、廣告效果與國際行銷等相關領域。 中 文 關 鍵 詞 ： 轉碼廣告、語言熟悉度、消費者認同、廣告效果評估

英 文 摘 要 ： This research investigates the moderating role of language familiarity, consumer identity and product categories in code-switched ad effectiveness for non-bilingual consumers. Code-switched ads refer to ads that insert foreign words or phrases into a sentence in an advertising copy (e.g.,

headline and slogan), resulting in a mixture of native and foreign languages in the ads.

Code-switched ad has been found to have great impact on consumers attitudes toward the ad and it has sprung up over the advertising research with time (Luna and Peracchio, 2005b; Luna and Peracchio, 2001; Luna, Lerman, and Peracchio, 2005) Nevertheless, there is scant research literature available on the use of CS in advertising. The persuasion effect of a code switching ad could be

influenced by both intrinsic, such consumer identity, and extrinsic, such as language familiarity and product

category aspects. Therefore, in this research project, three different aspects will be further examined: (1) the critical effect of product category on persuasion of a code-switched ad between non-bilinguals, and bicultural bilinguals; (2) The moderating role of and consumer

identity on a code-switching ad; (3) the moderating effect of language familiarity on persuasive of a code-switching ad and the underlying mechanism mediating role of language comprehension on persuasive of a code-switching ad.

By advancing our understanding of the mechanism and

applications of CS advertising, this research project will provide a valuable contribution to the fields of consumer

psychology and advertising, as it will unveil the

phenomenological mechanisms involved in consumer-related phenomena and provide useful insights for managerial

applications. The proposed studies will provide advertisers with guidelines for developing advertising strategies in Taiwan.

英 文 關 鍵 詞 ： Code-switched ad; consumer identity; foreign language familiarity, ad effectiveness

本三年期計畫欲逐步探討影響轉碼廣告效果的因素，分三部分進行探討: (1)驗證不同產品 類別的轉碼廣告說服效果，並驗證該效果是否能同時存在於非雙語精通者(non-bilinguals)與雙 文化雙語使用者(bicultural bilinguals)中；(2) 確認消費者認同對轉碼廣告效果的影響效果；(3) 驗證語言熟悉度對轉碼廣告的干擾效果，同時提出語言理解度為其中重要的中介變數。茲將三 部分內容分述如下：

子計畫一：Language Choice in Advertising for Multinational Corporations and Local Firms: A Re-Inquiry Focusing on Monolinguals

This research involved conducting 2 studies to investigate whether the asymmetric language effects observed by Krishna and Ahluwalia (2008) among bilinguals can be replicated among monolinguals. In Study 1, we observed asymmetric language effects for local firms, but not for multinational corporations (MNCs), which differs from Krishna and Ahluwalia’s observations. Based on the Study 1 results, Study 2 further proposes and tests two routes (language expectation and language-based) that lead to consumer slogan evaluations. The findings of Study 2 suggest that slogan evaluations for MNCs were determined via the language expectation route, whereas

evaluations for local firms were determined via the language-based association route. This research provides evidence that monolinguals have different responses to language choices in advertising than bilinguals do. Possible explanations and implications are discussed, and future research directions are outlined for this underexplored area.

子計畫二：Local or Global Image? The Role of Consumers’ Local-Global Identity in Code-Switched Advertising Effectiveness among Monolinguals

Local and global brands alike have increasingly used code-switching to enhance advertising persuasion. Although this tactic is widely used, previous studies have focused on bilinguals, but not monolinguals. Because of the emerging use of code-switching in advertisements in monolingual markets, more research efforts are required to understand its effectiveness and boundaries among monolinguals. This study investigated whether the consumers’ local-global identity plays a moderating role in the effectiveness of code-switched advertisements among monolinguals. The consumers’ local-global identity refers to the combination of local and global identities possessed by individuals that affect how they define themselves in relation to the social environment. Study 1

(manipulating consumers’ local–global identity) demonstrated that the local-global identity moderated the effect of code-switched advertisements. The results indicated that the congruence between code-switching and the consumers’ local-global identity enhanced persuasiveness, and that advertising involvement mediated this effect. Study 2 replicated the findings of Study 1 by using a local-global identity measure. These findings provide implications for branding and advertising strategies.

子計畫三：Creating an Effective Code-Switched Ad for Monolinguals: The Influence of Brand Origin and Foreign Language Familiarity

This study investigates the influence of brand origin and foreign language familiarity in code-switched (CS) ad effectiveness for monolingual consumers. CS ads refer to ads containing foreign words or phrases in an advertising copy (e.g., headline and slogan), resulting in a mixture of native and foreign languages. In this research, we conducted a pilot study to show the increasing trend of using code-switching in ads, regardless of whether the brands are local or foreign, in a monolingual market. We further examine if the effectiveness of CS ads is contingent on the brand origin among monolinguals. Study 1 showed that non-CS ads were perceived more favorably than CS ads for advertising a local brand. However, not all of the CS ads were perceived more favorably than non-CS ads when a foreign brand was advertised. The results of Study 2 showed that when a foreign brand was advertised, CS ads using a high exposure foreign language were evaluated more favorably than CS ads using a low exposure foreign language. Foreign language familiarity played a mediating role in the observed effects. We provide evidence that the research findings on CS ads among bilinguals cannot be applied to monolinguals. We discuss implications for international marketing and suggest advertising strategies for practitioners.

本計畫藉由理解影響轉碼廣告的重要因子與其中的影響歷程可以幫助理解消費者心理，並 能提供於採用轉碼廣告時的實務建議，其研究結果能貢獻於消費者行為、廣告效果與國際行銷 等相關領域。

Language Choice in Advertising for Multinational Corporations and Local Firms: A Re-Inquiry Focusing on Monolinguals

Language choice in advertising serves as a marketing tool in positioning a firm’s brand (local vs. global; Alden, Steenkamp, and Batra 1999; Chang 2008). Krishna and Ahluwalia (2008) examined language choice in advertising by multinational corporations (MNCs) and local firms with a focus on bilinguals. Their results revealed that language choice in ad slogans did not influence evaluations of ads used by local firms to market necessity and luxury goods. However, for MNCs, a foreign language (i.e., English) was more effective than the local language in marketing luxury goods, whereas the local language (i.e., Hindi) led to more favorable evaluations of ads for marketing necessity goods. These findings have crucial managerial implications for MNCs and local firms in marketing communications. Our research aims to answer the question: Can the asymmetric language effect among bilinguals mentioned above be generalized to a monolingual population?

Because of the growth of bilingual populations worldwide, code-switched ads which contain two or more languages have been widely adopted by marketers to attract the attention of bilinguals. The use of this advertising tactic has emerged as a trend in monolingual countries such as China, South Korea, and Taiwan. Longitudinal research involving content analysis has suggested that

code-switched ads are widely used in Taiwan. Specifically, the use of English brand names in

code-switched ads in the Taiwanese market has gradually increased (over 80% in total in 2007, 2009, and 2012, Lin and Yang 2013), particularly compared with the 53.90% observed in magazine ads sampled from 2003 to 2004 (Chang 2008).

Grosjean (1992, p. 51) stated that “[b]ilingualism is the regular use of two (or more) languages, and bilinguals are those people who need and use two (or more) languages in their everyday lives.” Our target populations, called monolinguals in this research, do not communicate in two languages

regularly. Previous research has identified differences (e.g., brain and cognition) between bilinguals and monolinguals (see Bialystok, Craik, and Luk 2012 for a review). For instance, bilinguals expend more effort retrieving a common word than monolinguals (Jones et al. 2012); however, bilinguals have better executive control than monolinguals (Bialystok et al. 2012). Research has also shown that the language switching mechanism of bilinguals is different from that of monolinguals (Costa and Santesteban 2004). As such, the manner in which monolinguals process code-switched ads should be different from that of bilinguals; thus, it is not appropriate to assume that findings observed among bilinguals can be generalized to monolinguals. This research entailed conceptually replicating Krishna and Ahluwalia’s (2008) research in a monolingual context. Specifically, we reinvestigated the effects of asymmetric language on ad effectiveness and the underlying mechanisms among monolinguals. Examining language choice in advertising targeting monolinguals not only fills the research gap but also provides marketing communication implications for practitioners.

In contrast to Krishna and Ahluwalia’s findings, we found that asymmetric language effects were not observed for MNCs but were for local firms. In Study 1 for MNCs, ad slogans with foreign language elements were evaluated more favorably by monolinguals than those with local language elements, regardless of the product type. For local firms, English slogans were evaluated more favorably than Chinese in marketing luxury goods. However, Chinese slogans were more effective than English slogans in marketing necessities. Based on the Study 1 results, we further proposed and tested two routes (language expectation and language-based association) that lead to slogan

evaluations among monolinguals in Study 2. The results suggest that slogan evaluations for MNCs were determined via the language expectation route, whereas evaluations for local firms were determined via the language-based association route. Due to in-group favoritism, consumers are content focused when processing ad messages from local firms with which they share a group

identity (i.e., nationalism). Consumers pay attention to the advertised content, rather than the language used. Attitudes toward an ad slogan are more favorable if the association of the advertised product matches the language used in the ad. On other hand, when evaluating ad slogans by MNCs, consumers focus on peripherals (i.e., language used) instead of content. Monolinguals have less attentional control for incongruent stimuli, so the use of a language that meets expectation is preferred. Given this, our study goes beyond replicating previous advertising research on code switching and provides managerial implications for MNCs and local firms that can be used to inform their language choices in marketing products to monolingual populations (Carlson 2015).

THEORETICAL BACKGROUND

Numerous studies have investigated language choice in advertising to bilinguals over the past decade (e.g., Bishop and Peterson 2010; Luna and Peracchio 2005a; 2005b), particularly in terms of the impact of code-switching on ad persuasion. For example, picture–text congruity (Luna and Peracchio 2001) and motivation (Luna and Peracchio 2002) have been shown to facilitate second language processing and, thus, enhance ad message recall and recognition. Consumer language attitude (Luna and Peracchio 2005a), attitude toward code-switching (Luna and Peracchio 2005b), and type of processing (Luna, Lerman, and Peracchio 2005) have been identified as moderating the effect of code-switched ads. Recent research has also shown that brand origin and product category (Krishna and Ahluwalia 2008) as well as ad content (Carroll and Luna 2011) and medium context (Bishop and Peterson 2010; 2011) should be considered in language choice in advertising.

Some researchers have shifted the focus to monolingual populations. Chang (2008) has shown that using the local language in a brand name enhances brand trust and liking among Taiwanese consumers across three product categories (telecommunication, fashion and food). Villar, Ai, and Segev (2012) found that Chinese consumers prefer the language used in the brand name to match its

brand origin. Due to different research designs and foci, the generalizability of these research findings across two populations appears to be an important research gap.

Bilingualism research has shown differences in cognition and effort expended on such tasks as reading between bilinguals and monolinguals. Reading words demands more work from the brain of bilinguals than that of monolinguals (Jones et al. 2012). Bilinguals have a smaller receptive

vocabulary size (Bialystok and Luk 2012) and slower responses to naming pictures (Costa and Santesteban 2004), comprehending (Ransdell and Fischler 1987), and producing words (Ivanova and Costa 2008) than monolinguals. In contrast, bilinguals have better executive control and linguistic performance than monolinguals due to their lifetime experience managing attention to two languages (Bialystok et al. 2012). Bilinguals’ language switching mechanism is different from that of

monolinguals (Costa and Santesteban 2004) in that switching from the weaker language to the more dominant language is harder than vice versa for monolinguals, but not for bilinguals. Given these differences, it is likely that the asymmetric language effects observed among bilinguals will not be replicated among monolinguals.

To compare the results of our research with those of Krishna and Ahluwalia (2008), we

replicated all of their studies, including the pilot study, four pretests, and two main experiments, as well as their procedures and measures. In the interest of brevity, we do not repeat all of the details in this paper. In the pilot study, we report how language favorability and use of language differ between India and Taiwan. We develop the stimuli used for the two experiments in the pretests and then present our main studies and results. Finally we discuss our findings and implications.

PILOT STUDY

The purpose of the pilot study was to understand the language use, proficiency and perceived favorability of both languages. Fifty undergraduate students (Mage = 20.41, 36% male) from a

university in Taiwan participated in the study. Table 1 shows that Chinese is the dominant language used in Taiwan (both at home and at school). In addition, a self-assessment of language proficiency showed that the subjects were more proficient in Chinese than in English, which indicates that our target population consisted of monolinguals.

To evaluate language favorability, the subjects were required to assess two languages. Table 2 summarizes responses to open-ended and rating-based items regarding language favorability. An open-ended question prompted subjects to describe their thoughts, images, and feelings when hearing someone speaking English (Chinese). The results showed that both languages were considered favorable because positive thoughts were significantly more numerous than negative thoughts (Chinese: M = .6 vs. .0, t(49) = 5.42, p < .001; English: M = .68 vs. .3, t(49) = 2.91, p < .01). Consistent with these findings, the subjects reported that their feelings toward both languages were favorable (neutral midpoint = 4; English = 5.4, Chinese = 6.16, both differed from 4.00 at p < .001). For the two items regarding negative associations (inferior and embarrassing), both languages received low scores (midpoint = 3, inferiority: M = 2.5 and 1.26 for English and Chinese; embarrassing: M = 2.6 and 1.24 for English and Chinese; all four differed from 3.00 at p < .02), indicating that neither language was evaluated negatively.

The results (Table 3) indicate that English exhibited significantly stronger associations with all of the words representing sophistication (all p < .001) except “exclusivity,” whereas Chinese

exhibited significantly stronger associations with most of the items representing belongingness (all p < .001), except “personal.” In addition, English was associated more strongly than Chinese with “distant” and “upper class” (p < .001). However, the languages did not differ in the extent to which they were perceived as polite or stern, or associated with the middle class (all p > .1).

Discussion

In Taiwan, both Chinese and English were evaluated favorably. Nevertheless, the subject population was more fluent in Chinese than in English. These results differ from Krishna and Ahluwalia, in which the bilingual population in India was fluent in both the foreign (English) and local (Hindi) languages. Still, our pilot study revealed a language association pattern similar to that observed by Krishna and Ahluwalia, namely that the local language (Chinese) was associated with belongingness (family, closeness, sense of belonging, and caring), whereas the foreign language (English) was associated with sophistication (globalness, cosmopolitanism, professionalism, and prestige) and the upper class.

PRETESTS AND DEVELOPMENT OF STIMULI Pretests

Pretest 1: Language Expectations. Forty MBA students (Mage = 24.6, 37.5% male) from a large

university in Taiwan were recruited and rated the extent to which they expected (two 7-point scales, anchored by unexpected/expected and unlikely/likely) that an ad slogan from either a local company or an MNC would be in English, Chinese, or a mixed language (Cronbach’s alphas for all three language types > .94). A Chinese slogan was more highly expected (M = 6.78) for local firms than a mixed-language (M = 5.98) or English (M = 4.3, all p’s < .01) slogan. By contrast, for MNCs, English (M = 6.5) and mixed-language (M = 6.4) slogans were similarly highly expected (p > .6), while a Chinese slogan was less expected and its mean value was significantly lower (M = 4.78, both p < .01). The results (Table 4) indicate that local firms were not expected to use the foreign language and MNCs were not expected to use the local language, which is similar to those of Krishna and Ahluwalia. In contrast, a mixed language was not considered an unexpected language for an MNC among the monolinguals in our study.

Pretest 2: Relevance of Belongingness and Sophistication. Twenty-two different students (Mage =

21.09, 50% male) from the same university participated in testing the relevance of associations with belongingness (closeness and friendliness, r = .9) and sophistication (sophistication and global, r = .85) in evaluating necessity and luxury goods. The results reveal that product category had a main effect on belongingness (F(1, 20) = 30.57, p < .001) and sophistication (F(1, 20) = 51.67, p < .001). For necessity goods, belongingness (M = 6.59) was rated as significantly more important than sophistication (M = 3.45, F(1,10) = 79.88, p < .001), while for luxury goods, sophistication (M = 6.27) was rated as significantly more important than belongingness (M = 3.32, F(1,10) = 23.29, p = .001). The results (Table 5) are similar to those observed by Krishna and Ahluwalia.

Pretest 3: Product Categories. Pretest 3 was conducted to select a target product for each category. Another fifty undergraduate students (Mage = 22.62, 34% male) from the same university

were recruited. The results (Table 6) indicate that soap was perceived as a necessity (M = 1.2), whereas ice cream was perceived as a luxury (M = 3.74, F(1,49) = 174.81, p<.001). These two products fulfilled the following requirements: (a) they were frequently purchased by the students; (b) price was not the only consideration in purchasing them, and (c) the students did not consider the country of origin when purchasing them. Therefore, we used soap and ice cream as necessity and luxury goods, respectively, in our main study.

Pretest 4: Ad Slogans. We developed four language formats for ad slogans for each product category (Table 7) in accordance with Krishna and Ahluwalia. The different language versions were translated using the standard back translation technique (Hui and Triandis 1985) to ensure that they had the same meaning. An additional pretest involving 40 undergraduate students (Mage = 22.68, 35%

male) was conducted to ensure that all four ad slogans were free of confounds. No differences were observed among the four versions of the slogans in perceived flow, clarity, and appeal (all p’s >.2).

PLACE TABLES 4, 5, 6 and 7 ABOUT HERE

STUDY 1: TESTING KRISHNA AND AHLUWALIA’S FRAMEWORK

Krishna and Ahluwalia (2008) suggest that when the language of a slogan is unexpected,

bilinguals pay attention to language-based associations. Slogan evaluations were more positive when the associations matched the product category. In other words, the foreign language was associated with sophistication, which matches luxury goods, and thus enhances slogan evaluations for luxury goods. The local language was associated with belongingness, which matches necessity goods, and thus enhances slogan evaluations for necessity goods. Local and mixed-language slogans are unexpected for MNCs, whereas none of the language formats are unexpected for local firms. Thus, language-based ad effects have been observed for MNCs but not for local firms. The purpose of Study 1 is to test whether Krishna and Ahluwalia’s framework is valid among monolinguals.

Method

A 2 (corporation: multinational vs. local) × 2 (slogan language: English vs. Chinese) × 2 (product type: luxury vs. necessity) between-subjects design was used. A total of 327 students (Mage

= 20.36, 43.1% male) enrolled in marketing classes in Taiwan participated in the study and were randomly assigned to one of eight treatment conditions. The subjects read a short paragraph

describing the company that was advertising (the descriptions were identical except for information identifying the company as either a local firm or an MNC). Subsequently, a print ad (one of four versions: English ice cream, Chinese ice cream, English soap, and Chinese soap) and dependent measures (slogan evaluation: good/bad and like/dislike, r = .90) as well as covariates (language fluency: reading, writing, and speaking,  = .85; net language favorability: difference between reported evaluations of English and Chinese) were presented. The covariates had no significant effects on the dependent variable. Thus, we did not include them in the analysis.

Results

The data were analyzed as a 2 (corporation) × 2 (slogan language) × 2 (product type) factorial design. Degrees of freedom for all measures are 1 and 319 unless otherwise indicated. Treatment means are listed in Table 8. The three-way ANOVA revealed a main effect of product type (F = 9.63, p < .01) and slogan language (F = 9.15, p < .01) on slogan evaluation. The results suggest that among the different conditions, the soap ad (M = 5.10) was evaluated more favorably than the ice cream ad (M = 4.74), the English slogan (M = 5.09) was considered more favorably than the Chinese slogan (M = 4.74), and the interactions of corporation with slogan language (F = 8.83, p < .01) and product type with slogan language (F = 36.98, p < .001) were significant. These effects were qualified by a significant three-way interaction (F = 13.25, p < .001). For the MNC, the interaction of language with product type was marginally significant (F = 3.23, p =.07). The results showed that for

necessities, the English slogan (M = 5.31) was evaluated more favorably than the Chinese slogan was (M = 4.9; F = 3.09, p < .08). A similar result pattern was observed for luxury goods. The English slogan (M = 5.36) was evaluated more favorably than the Chinese slogan was (M = 4.38; F = 18.03, p < .001). For the local firm, we observed a significant two-way interaction between slogan language and product type (F = 47.41, p < .001). The subjects evaluated the Chinese slogan more favorably than the English slogan (M Chinese = 5.64, M English = 4.52, F = 23.45, p < .001) for necessity goods.

However, for luxuries, the subjects evaluated the English slogan more favorably than the Chinese slogan (M English = 4.04, M Chinese = 5.17, F = 23.96, p < .001).

PLACE TABLE 8 ABOUT HERE

Discussion

The results of our study conducted in Taiwan, a monolingual country, exhibit a different pattern than Krishna and Ahluwalia’s (2008) research. Our results suggest that for an MNC, when the

language choice is consistent with the firm’s country of origin, slogan evaluations are enhanced. Therefore, MNCs should use a foreign language to market its product regardless of the product type. However, for local firms, the language choice for ad slogans should depend on the product type. The local language should be used for marketing necessity goods, whereas a foreign language should be used for marketing luxury goods.

Based on the results of Study 1, we propose two possible routes that lead to the formation of slogan evaluation among monolinguals, language expectation and language-based association. If slogan evaluations are formed through language expectation, the expected language should be used for marketing communication. MNCs should use a foreign language, whereas local firms should use the local language. If slogan evaluations are formed through language-based associations, they are more positive when the associations match the product category. The foreign language is associated with sophistication, which matches luxury goods and, thus, enhances the slogan evaluations for this category. The local language is associated with belongingness, which matches necessity goods and, thus, enhances the slogan evaluations for this category. Figure 1 illustrates the predictions of slogan evaluation. Study 2 was conducted to test our proposed routes.

PLACE FIGURE 1 ABOUT HERE

STUDY 2: ROUTE OF SLOGAN EVALUATION AND THE UNDERLYING PROCESS

Because our findings in Study 1 differed from those of Krishna and Ahluwalia, instead of following their study and investigating the mechanism underlying the effects only for MNCs, we examined slogan evaluations for both MNCs and local firms. In addition to testing whether the findings of Study 1 could be replicated using mixed-language slogans, Study 2 tested the proposed routes to slogan evaluation formation. A total of 571 undergraduate students (Mage = 21.74, 42.2%

multinational vs. local) × 2 (product type: luxury vs. necessity) × 4 (language: English vs. most English vs. most Chinese vs. Chinese) between-subjects design. All ad versions used in this study are listed in Table 7. The procedure and key dependent variable (slogan evaluation: r = .90) were

identical to those of Study 1. Participants’ perceptions of the slogan in the dimension of belongingness were evaluated using three 7-point scales anchored by impersonal/personal, distant/close like family, and formal/friendly (α = .71), whereas perceptions of sophistication were assessed using three 7-point scales anchored by rural/cosmopolitan, middle class/upper class, and local/global (α = .76).

Results

Overall omnibus analysis. The data were analyzed as a 2 (corporation) × 2 (product type) × 4 (language) factorial design. Degrees of freedom for all measures are 3 and 555 unless otherwise indicated. The three-way ANOVA on slogan evaluation revealed the overall main effects of product type (F(1, 555) = 7.13, p < .01) and language (F = 3.62, p < .02). The main effect of product type revealed a higher slogan evaluation for necessity (M = 5.02) than for luxury goods (M = 4.8). The main effect of language showed that the Chinese slogan was evaluated lower (M = 4.69) than the mostly Chinese slogan (M = 4.93), the mostly English slogan (M = 5.07), and the English slogan (M = 4.94, p’s < .01). There were no differences among the last three slogan evaluations. The two-way interactions of corporation with language (F = 4.61, p < .01) and product type with language (F = 8.47, p < .001) were significant. These two-way interactions must be interpreted in light of a significant three-way interaction (F = 7.24, p < .001). We further examined the product type and language interaction regarding local firms and MNCs. Means, by condition, are presented in Table 9.

Multinational corporation. A 2 (product category) × 4 (language type) ANOVA revealed that the two-way interaction was significant (F = 3.14, p < .05), which indicates that slogan evaluation in

different language formats exhibited different patterns between necessity and luxury goods. When the product was a luxury good, the language format of the slogan exerted a significant effect on slogan evaluation (F = 3.41, p < .05). The pattern of slogan evaluation revealed that the Chinese slogan (M = 4.53), the mostly Chinese (M = 4.91) and the mostly English (M = 4.8) was evaluated lower than the English slogans (M = 5.25; all p’s < .05). The former three was no significantly different. However, when the product was a necessity, the language format of the slogan exerted a significant effect on the slogan evaluation (F = 6.62, p < .001). The Chinese slogan (M = 4.37) was considered the least favorable, followed by the English slogan (M = 4.85). The two mixed versions exhibited a similar evaluation (M = 5.28 and 5.3 for the mostly Chinese and mostly English slogans, respectively) and were evaluated significantly more highly than the Chinese slogans (all p’s < .05). The results suggest that for MNCs, slogans with English elements (e.g., an English or a

mixed-language slogan) were evaluated more favorably than the Chinese slogan in marketing luxury goods. The Chinese slogan was the least favorable in both product categories for an MNC. These results support the prediction of the language expectation route rather than the language-based association route.

Local firm. A 2 (product category) × 4 (language format) ANOVA revealed that the two-way interaction was significant (F = 12.30, p < .001). When the product was a necessity, the language format of the slogan exerted a significant effect on the slogan evaluation (F = 6.70, p < .001). The Chinese slogan (M = 5.63) was the most favorable, followed by the mostly Chinese (M = 5.08) and mostly English slogans (M = 5.09). The English slogan was rated the least favorable (M = 4.56). The latter three slogan evaluations were not significantly different. When the product was a luxury good, the language format of the slogan exerted a significant effect on the slogan evaluation (F = 7.02, p < .001). The English (M = 5.10) and mostly English slogans (M = 5.10) were the most favorable,

followed by the mostly Chinese slogan (M = 4.46). The Chinese slogan was rated the least favorable (M = 4.23). The results suggest that for local firms, the Chinese slogan was the most favorable for marketing necessities; however, English or mostly English slogans were more favorable for marketing luxury goods. Results support the prediction of the language-based association route.

PLACE TABLE 9 ABOUT HERE

Underlying mechanism. To test the prediction that belongingness (sophistication) perceptions mediate slogan evaluations for necessity goods (luxury goods), we conducted a moderated mediation analysis in accordance with Hayes (2013). A bootstrapping method supported the claim that

belongingness perceptions mediate the effect of slogan language on slogan evaluations for necessity goods, with the 95% confidence interval (CI) of the indirect effect excluding zero (lower limit [LL] CI = -.2237; upper limit [UL] CI = -.0356). As expected, belongingness did not mediate slogan evaluations for luxury goods, with the 95% confidence interval of the indirect effect including zero (LL CI = -.0496; UL CI = .1132). Conversely, a bootstrapping method supported the claim that sophistication perceptions mediate the effect of slogan language on slogan evaluations for luxury goods, with the 95% confidence interval of the indirect effect excluding zero (LL CI = .0215; UL CI = .1110). However, sophistication also mediated slogan evaluations for necessity goods, with the 95% confidence interval of the indirect effect excluding zero (LL CI = .0189; UL CI = .1173). This mediation effect was not expected.

GENERAL DISCUSSION

The results of the two studies in this research show that Krishna and Ahluwalia’s (2008) findings regarding bilinguals in India cannot be replicated among monolinguals. The language choice in ad slogans by MNCs and local firms targeting monolinguals seems more complex than that for slogans targeting bilinguals. In the two studies, for MNCs, the English (or mixed language) slogan was

evaluated more favorably than the Chinese slogan, regardless of the product category. The findings support the language expectation route. In contrast, the results for local firms seem to support the language-based association route. In both studies, for local firms, the English slogan was evaluated more favorably than the Chinese slogan in marketing luxuries, whereas the Chinese slogan was more favorable than the English slogan in marketing necessities. Study 2 further shows that the English slogan was considered as favorably as the mostly English slogan, and more favorably than the Chinese in marketing luxuries. Slogans with foreign language elements (except when the local language is dominant) are superior in positioning a brand as a luxury, whereas slogans with local language elements are superior in positioning a brand as a necessity.

In-group favoritism, which refers to when a group member tends to have a more favorable attitude toward other group members than people outside the group (Tajfel and Turner 1986), may explain why the two routes were taken. Researchers have indicated that messages from in-group members evoke content-focused processing (Mackie, Worth, and Asuncion 1990). Consumers share a group identity (i.e., nationalism) with local firms, but not with MNCs; thus, when they see ads from local firms, they are likely to pay attention to the advertised content (e.g., messages and product category), not simply the language used. When the product category association matches the language used, slogan evaluation is more favorable. On other hand, for MNCs, consumers focus on the peripherals (e.g., language used). Research has indicated that monolinguals possess inferior attentional control over incongruent stimuli (Bialystok et al. 2012). Thus, it is likely that when they view mixed-language or English slogans by MNCs, they prefer the use of language that meets their expectation, which supports the language expectation route. MNCs should be aware in monolingual markets that using only the local language in ad slogans to position a brand (i.e., local positioning) might not be as effective as expected (Alden, et al. 1999; Chang 2008).

Regarding the underlying mechanism, findings concerning the mediation effects of

belongingness and sophistication were mixed. As expected, belongingness mediated the effect of language on slogan evaluations for necessity goods, but not for luxury goods. Our pilot study results show that Chinese is associated with belongingness, whereas English is associated with

sophistication. Taiwan’s history of Japanese occupation and rule also enhances the Chinese– belongingness association. Using Chinese activates consumers’ belongingness association. Thus, Chinese is more relevant than English in evaluating necessities, and English is more relevant in evaluating luxury products. However, our findings show that sophistication mediates the effect of language on slogan evaluations for both necessity and luxury goods. This unexpected result has two possible causes. The ad slogans used in the studies were different for the two product categories. The mediation effect was likely caused by the slogans rather than the product categories. In addition, in accordance with Krishna and Ahluwalia’s study, the relevance of product category association to sophistication was measured on two scales, sophistication and global, whereas the mediation effect of sophistication was measured on three scales anchored by rural/cosmopolitan, middle class/upper class, and local/global. The former construct focused on the association of sophistication with product categories, whereas the latter construct was used to measure respondents’ perceptions of slogans on sophistication. The differences in the foci of the measures might be problematic.

Our findings seem consistent with the ad practices in Taiwan observed by Chang (2008), who indicated that the product category with the highest percentage of English brand names was luxury goods (e.g., fashion and cosmetics), whereas the product category with the highest percentage of Chinese brand names was necessity goods (e.g., daily consumption products and food). In addition, Chang (2008) indicated that English brand names enhanced the perceived globalness of a brand and Western models enhanced the perceived product quality, whereas Chinese brand names were

associated with brand friendliness and trust among monolinguals in Taiwan. These factors might explain the effects that we observed for local firms in our study. Further research is required.

Language–product and language–country of origin (language-based association) congruity effects may provide an alternative theoretical explanation for our research findings. For slogan evaluations to be positive, language associations must match the country of origin and product type associations. When there is a language–product or language–country of origin mismatch, evaluations are low. For monolinguals, foreign language ads are salient; monolinguals pay attention to these ads because they are unfamiliar with the language and the ads break their everyday norms. For MNCs, if the language matches the country of origin association, evaluations are high. Mixed-language slogans are more effective than English-only slogans because they are understood better by monolinguals (Study 2). Including only a small portion of English activates congruent associations. For local firms, if language associations match product associations, evaluations are high. Thus, a foreign language should be used to advertise luxury products.

This research was designed to compare Krishna and Ahluwalia’s findings with findings obtained from a monolingual population. However, the research design has limitations. First, we used different ad slogans for different product categories. The differences between ad slogan evaluations may have been caused by differences in the phrases rather than the language formats. Future research is

encouraged to use an identical ad slogan for different product categories and test the ad effectiveness. Second, this research included only monolingual samples. The asymmetric code-switching effect can be compared between bilinguals and monolinguals by using the same research design in a single study. Similarly, due to difficulties defining the term “bilingual” (Francis 1999), future research should develop a scale for measuring the level of bilingualism as a continuous variable and thereby test the moderating effect of bilingualism on code-switched ad effectiveness.

This research also provides valuable directions for future research. Our findings suggest two routes that lead to slogan evaluations for MNCs and local firms, namely language expectation and language-based association, respectively. More empirical studies should be conducted to determine how each route is taken and explore other conditions that influence the route chosen. In addition, due to significant differences in how the English and Chinese languages are written and processed (Ahn and Ferle 2008; Unger 2011), code-switching between English and Chinese is likely different from that between other languages. Future research should explore how these factors may influence ad persuasiveness based on language.

Local or Global Image? The Role of Consumers’ Local-Global Identity in Code-Switched Advertising Effectiveness among Monolinguals

Presenting a local or global image is crucial for global brands and multinational companies when developing marketing strategies (e.g., Friedman 1999; Ö zsomer 2012; Sklair 1995; Zhou and Belk 2004). One marketing tactic to achieve such positioning is to vary the language used in advertisements according to the firm’s local or global branding strategy (Alden, Steenkamp, and Batra 1999; Chang 2008). Including a foreign language in advertisements is a growing trend that has received much attention (Ahn and Ferle 2008; Bishop and Peterson 2015; Chang 2008; Krishna and Ahluwalia 2008; Lin and Wang 2016). Advertisements that include two or more languages are

referred to as code-switched advertisements (Luna and Peracchio 2005a; 2005b). English, in addition to the local language, has been widely used in code-switched advertisements in many nonnative English-speaking markets to enhance the global appeal of brands and products, thereby increasing perceived quality and social status (Chang 2008; Eckhardt 2005; Zhou and Belk 2004).

Most studies investigating code-switched advertising effectiveness have focused on bicultural bilinguals, who incorporate the values of two cultures into their identities and speak the languages of both cultures (Bishop and Peterson 2010; Luna, Ringberg, and Peracchio 2008), and monocultural bilinguals, who learn a nonnative language in an educational setting (Krishna and Ahluwalia 2008; Luna, Ringberg, and Peracchio 2008). For example, researchers have examined code-switched advertising effectiveness among Hispanic Americans who speak both Spanish and English (Luna et al. 2008) and among Asian Indians who speak both Hindi and English (Krishna and Ahluwalia 2008). In such cases, bicultural and monocultural bilinguals can speak both their native language and a foreign language fluently. The findings of the aforementioned studies are crucial for marketers, because they demonstrate that including two languages in advertisements is a reasonable marketing

strategy when targeting bilinguals. In contrast to bilinguals, who can regularly communicate in two languages (Francis 1999; Luna, Ringberg, and Peracchio 2008; Krishna and Ahluwalia 2008), monolinguals are defined as individuals who are less-than-fluent in a second language. Recent studies have emphasized the importance of investigating the influence of foreign elements (e.g., languages and fashion models) on advertising effectiveness among monolinguals (Ahn and Ferle 2008; Chang 2008; Lin and Wang 2016).

Empirical studies have revealed an increasing trend of using code-switching tactics in

advertisements in Asian markets, such as Hong Kong, Japan, South Korea, and Taiwan, where most consumers are monolinguals (Cutler, Javalgi, and White 1995). Lin, Wang, and Hsieh (forthcoming) examined six magazine issues and 773 advertisements published in October 2009 and determined that both local and global brands widely use code-switching in Taiwan to influence the consumer perceptions of a brand or product. In addition, the results indicated that the number of code-switched magazine advertisements in Taiwan is on the rise: Lin, Wang, and Hsieh (forthcoming) reported 89% of the magazine advertisements used code-switching while 53.9% was observed from 2003 and 2004 (53.90%) in Chang’s (2008) study.

Despite the popularity of code-switched advertisements in nonnative English-speaking markets, the Boston Consulting Group (2008) indicated that such advertising tactics are not always effective among monolingual consumers. The decisions of these consumers are far more complex and are influenced by factors other than beyond brand origin or nationalism. The unsuccessful outcomes of code-switched advertisements might be attributed to differences among target populations.

Bilingualism literature has identified differences in brain activity and cognitive ability between monolinguals and bilinguals (see Bialystok, Craik, and Luk 2012 for a review). In addition,

and code-switched advertisements differently than bilinguals do, which subsequently influences code-switched advertising effectiveness (Lin and Wang 2016). In order to enable marketers to use code-switching tactics more effectively in monolingual markets, the effect of code-switched advertisements and their boundary conditions must be further investigated.

Previous studies have revealed the importance of identity accessibility, or the extent to which certain mental representations are salient, in consumer persuasion among bilinguals (Deshpandé and Stayman 1994; Lau-Gesk 2003). However, little is known about the effect of identities on consumer persuasion among monolinguals. In a study of the psychology of globalization, Arnett (2002) proposed that every individual possesses a local-global identity, comprising local and global identities, which affects how the individual defines himself or herself in relation to the social

environment. In this study, according to the identity–accessibility effect, we proposed and tested that the consumers’ local-global identity moderates the effect of code-switching on advertising

effectiveness among monolinguals. We also explored the mechanism underlying the effect of code-switching on persuasion.

This study contributes to the literature on code-switched advertising, branding, and

international marketing. First, we showed that the consumers’ local-global identity moderates the effect of code-switching on advertising persuasiveness. The results indicated that the effectiveness of a code-switched advertisement is enhanced when the consumers’ local-global identity corresponds with code-switching, supporting the identity-accessibility effect. Second, the results advance our understanding of the mechanism underlying the effect of code-switching on persuasion. Specifically, we showed that the process variable of advertising involvement is associated with the effects of code-switching on advertisement and product evaluations. Third, previous studies on branding and international marketing have focused on the effect of the country of manufacture and the brand origin

on consumer persuasion. Our study findings provide additional evidence supporting the idea that the global culture (e.g., consumers’ local-global identity) of the target population should be considered in international marketing efforts (Zhang and Khare 2009). Fourth, in addition to demonstrating the consumers’ local-global identity among monolinguals, our study showed that the consumers’ local-global identity could be manipulated through priming (e.g., reading local vs. global events). Similar results were obtained from the two studies (Study 1: manipulating consumers’ local-global identity; Study 2: measuring consumers’ local-global identity). The findings enable marketers to develop marketing strategies (e.g., selecting medium context) by using code-switched advertisements when targeting monolinguals. In summary, these findings provide managerial implications for

marketers by advancing the understanding of code-switched advertisements and by demonstrating the importance of the consumers’ local-global identity in a globalized market.

The remainder of this paper is organized as follows. First, we review the literature on code-switching in advertising and investigate the moderating role of the consumers’ local-global identity in the effectiveness of code-switched advertisements. Second, we present two studies that were conducted to investigate the moderating role of the consumers’ local-global identity in the effect of code-switched advertisements and explore the mechanism underlying this effect. Finally, we discuss the implications of our findings.

THEORETICAL BACKGROUND Code-Switching in Advertising

Bilinguals in various languages worldwide practice code-switching (Grosjean 1984); that is, they use two or more languages within one sentence or conversation (Myers-Scotton 1999).

Marketers use code-switched advertisements as a marketing tactic to target bilinguals, because more than one language is used within the advertisement; typically, words from a foreign language are

used for a brand name or slogan (Luna and Peracchio 2005a; 2005b). Researchers have shown that the medium language (Bishop and Peterson 2011) and the advertisement content (Carroll and Luna 2011) influence the effectiveness of code-switched advertisements. Studies have identified that several variables, including picture-text congruity (Luna and Peracchio 2001), motivation (Luna and Peracchio 2002), language attitude (Luna and Peracchio 2005a), type of processing (Luna, Lerman, and Peracchio 2005), attitude toward code-switching (Luna and Peracchio 2005b), and brand origin (Krishna and Ahluwalia 2008), moderate the effects of code-switching in advertising persuasion among bilinguals.

Recent studies have observed a growing trend of using this marketing tactic to target monolinguals and have examined its effect on advertising effectiveness. Ahn and Ferle (2008)

examined South Korean consumers and found that the language (foreign vs. local) used for the brand name and body copy influenced the recognition and recall of the brand name and the advertisement message. Chang (2008) studied Taiwanese consumers and revealed that the use of Western fashion models and English brand names affected the perceived global feel of a brand, perceived product quality, and brand affinity. Lin and Wang (2016) demonstrated differences in the processing of language switching between Taiwanese consumers (monolinguals) and bilinguals. Many countries, including emerging and developed countries, are populated with monolinguals (e.g., Brazil, France, Turkey, China, Japan, Thailand, Korea, and Taiwan). Although bilingualism literature has identified various differences between monolinguals and bilinguals, relatively few studies have focused on monolinguals when investigating code-switched advertising despite its increasing usage.

The increasing use of code-switched advertisements in monolingual markets can be attributed to globalization (Alden, Steenkamp, and Batra 1999), which refers to “a process by which cultures influence one another and become more alike through trade, immigration, and the exchange of

information and ideas” (Arnett 2002, p. 774). Globalization affects individuals’ psychological

functioning, and identity transformation is its central effect (Arnett 2002). Specifically, Arnett (2002) proposed that everyone possesses a local-global identity that influences how individuals identify themselves in relation to their social environments. However, few studies have examined the role of the consumers’ local-global identity in code-switched advertising effectiveness. This is a substantial research gap, considering the increasing importance of code-switched advertisements in consumer persuasion and international marketing. In this study, according to the identity–accessibility effect, we proposed that the consumers’ local-global identity moderates the effect of code-switched advertisements on advertising persuasion among monolinguals.

Effect of Consumers’ Local-Global Identity on Code-Switched Advertising Persuasion

Because of globalization, consumers have developed a local-global identity, comprising local and global identities (Arnett, 2002). A local identity is defined as mental representations that cause individuals to appreciate local traditions, be interested in local events, and to feel a part of and identify with the local community. By contrast, a global identity is defined as mental representations that prompt individuals to adopt positive attitudes toward globalization, be interested in global events, and to feel a part of and identify with the global community (Arnett 2002; Zhang and Khare 2009). When both identities are equally strong, individuals may experience identity confusion (Arnett, 2002; Zhang and Khare 2009). Thus, most often, “one identity is stronger and more accessible than the other” (Zhang and Khare 2009, p. 525). Advancing Arnett’s conceptualization of the local-global identity, Tu, Khare, and Zhang (2012) developed, empirically tested, and validated a local-global identity measure. They showed that this construct is related to but different from constructs such as ethnocentrism, nationalism, and global consumption orientation.

behavior (e.g., Reed II, Forehand, Puntoni, and Warlop 2012; Strizhakova and Coulter 2013) and international marketing (e.g., Bartikowski and Walsh 2015; Guo 2013; Westjohn, Singh, and

Magnusson 2012; Zhang and Khare 2009). Among these studies, one research stream focuses on the effect of the global identity in emerging and developed markets. Strizhakova and Coulter (2013) observed that the global identity moderates the positive relationship between materialism and environmentally friendly tendencies (e.g., willingness to pay extra for environmentally friendly products) in emerging and developed markets. The positive effect of materialism becomes stronger when an individual’s global identity increases. Guo (2013) indicated that the global identity

moderates the negative relationship between ethnocentrism and the consumers’ attitudes toward global brands from developed countries as opposed to brands from emerging countries. For low-global identity consumers, ethnocentrism has a negative effect on the consumers’ attitudes toward global brands from developed countries, but not toward global brands from emerging countries. Such effects do not occur in high-global identity consumers.

The other research stream examines how the local-global identity affects consumer responses to international marketing strategies. For example, Zhang and Khare (2009) showed that an accessible local-global identity influences consumer evaluations of local and global products. Consumers with an accessible local identity prefer a local (over a global) product, whereas those with an accessible global identity prefer a global (over a local) product. Westjohn, Singh, and Magnusson (2012) found that the global identity is strongly associated with responsiveness to global consumer culture

positioning, whereas the local identity is significantly related to responsiveness to local consumer positioning. Bartikowski and Walsh (2015) demonstrated the influence of the global identity on consumer intention to purchase foreign products. In accordance with this second research stream, in this study, we examined the possible moderating role of the consumers’ local-global identity in their

responses to code-switched advertisements.

According to the identity-accessibility effect, individuals prefer stimuli that are consistent with an identity when it is accessible (i.e., its mental representations are salient) (Zhang and Khare 2009). When an identity becomes accessible, an individual’s various identities are reorganized so that the accessible identity becomes more prominent than the other identities, subsequently evoking

corresponding responses (Brewer 1991). Factors such as language cues can increase the salience of an identity and can thus influence the effect of the identity on attitudes and behavior (Reed II et al. 2012). Thus, it is likely that the languages used in code-switched advertisements influence the salience of an identity, making either the local or global identity more accessible, and subsequent advertisement evaluation.

Alden, Steenkamp, and Batra (1999, p. 77) stated that “one way for a brand to communicate global consumer culture positioning is to use English words, written and/or spoken, in its

communication. In contrast, a brand manager wanting to use local consumer culture positioning might emphasize the local language.” Noncode-switched advertisements, which contain only local language characters, evoke a perception of local culture (Alden, Steenkamp, and Batra 1999). The use of a local language in a brand name, headline, or body copy entices consumers with an accessible local identity to view the advertisement and advertised product more than when a mixture of local and foreign languages is used. The local elements are consistent with the consumers’ local identity; thus, the consumers feel more personal relevance (Brewer 1991; Wheeler, Petty, and Bizer 2005). Such consumers weigh local elements more heavily than global elements (Zhang and Khare 2009) and consequently respond more favorably to noncode-switched advertisements than to code-switched advertisements.

language characters, evoke a perception of global culture (Alden, Steenkamp, and Batra 1999). The use of a mixture of local and global languages in a brand name, headline, or body copy prompts consumers with an accessible global identity to view the advertisement and advertised product more than when the local language is used. The global elements are consistent with the consumers’ global identity; thus, the consumers feel more personal relevance (Brewer 1991; Wheeler, Petty, and Bizer 2005). Such consumers weigh global elements more heavily than local elements (Zhang and Khare 2009). Thus, they exhibit a more favorable response to code-switched advertisements than to noncode-switched advertisements. Specifically, in this paper, we propose that:

H1: Consumers with an accessible local identity evaluate noncode-switched advertisements

and advertised products more favorably than code-switched advertisements and advertised products.

H2: Consumers with an accessible global identity evaluate code-switched advertisements and

advertised products more favorably than noncode-switched advertisements and advertised products.

According to the identity verification principle (Reed II, Forehand, Puntoni, and Warlop 2012), consumers will monitor themselves to maintain their self-views when their identity is salient. They prefer that others’ perception of themselves be consistent with their perception and will take actions to defend their identities. Thus, identity-consistent advertisements are likely to be associated with more consumer advertising involvement than identity-inconsistent advertisements.

Identity-consistent information is perceived as being more relevant for processing objectives than identity-inconsistent information (Wheeler, Petty, and Bizer 2005). Advertising involvement is a crucial factor influencing advertising persuasion (Petty, Cacioppo, and Schumann 1983). Individuals are more likely to be persuaded by an advertisement if their advertising involvement is higher

(Bishop and Peterson 2010). Consequently, consumers develop more favorable attitudes toward identity-consistent advertisements than toward identity-inconsistent advertisements. Therefore, in this paper, we propose that:

H3: Advertising involvement mediates the effect of the consumers’ local-global identity on the

effectiveness of code-switched advertisements.

To test the research hypotheses, we conducted two experiments that involved manipulating (Study 1) and measuring (Study 2) the consumers’ local-global identity. The results indicated that the congruence between code-switching and the consumers’ local-global identity increased advertising persuasiveness through advertising involvement. Both studies provided evidence supporting the idea that consumers possess a local-global identity, consisting of local and global identities (Arnett 2002). Real and fictitious brands were used in Study 1 and Study 2, respectively, to test the robustness of the results of the two studies. Using a real brand increases the generalizability of the findings, and using a fictitious brand can avoid the possible confounding effects of preexisting attitudes toward a brand. We enhanced the internal and external validity of the findings with convergent results when both real and fictitious brands were used. We selected Taiwanese respondents as the study sample (i.e.,

monolinguals). Recent research has shown that Chinese is the main language spoken at home and in school in Taiwan, and that the respondents in Taiwan are more proficient in Chinese than in English (Lin and Wang 2016). In addition, researchers have shown evidence of an increasing trend of using code-switched advertising in Taiwan (Chang 2008; Lin, Wang, and Hsieh forthcoming). Thus, the respondents and the study context favorably serve our research purpose.

STUDY 1: MANIPULATING CONSUMERS’ LOCAL-GLOBAL IDENTITY

An increasing number of local firms in developing countries have attempted to take advantage of the equity of a foreign country’s image, intending to enhance the appeal of their advertisements by

using foreign (often Western) fashion models, settings, brand names and foreign-language body copy (Ahn and Ferle 2008; Chang 2008; Zhou, Yang, and Hui 2010). Similarly, international firms have devoted time and effort to localizing their brands, products, and packaging to attract local consumers and enhance product evaluations in local markets (Zhang and Schmitt 2001; Krishna and Ahluwalia 2008). Krishna and Ahluwalia (2008) indicated that consumers are likely to provide a more favorable evaluation of certain product categories when a local language is used than when a foreign language is used by foreign companies. These trends of international marketing strategies suggest that the brand origin (local vs. global) moderates the effect of code-switching on advertising persuasion. Thus, we included the brand origin in the research design to exclude alternative explanations. Global

brands refer to brands that are recognized internationally, whereas local brands refer to brands that are only recognized domestically (Zhou et al. 2010). A 2 (code-switching: noncode-switching vs. code-switching) × 3 (consumers’ local-global identity: local vs. global vs. control) × 2 (brand origin: local vs. global) between-subjects design was used in this study.

Method

Stimuli. A pretest was conducted to select a target product and to ensure that the consumer evaluation of the product was not influenced by the brand origin. To avoid the possibility of the product category affecting the persuasiveness of code-switched advertisements (e.g., Krishna and Ahluwalia 2008), we used daily necessities, products that everyone requires and frequently uses, to test our hypotheses in both studies. In the pretest, 30 undergraduate students were recruited and presented with 10 daily necessities (e.g., toothpaste, pen, laundry detergent, and shampoo). They responded to questions about whether they considered the brand origin when purchasing each of the 10 products; the responses were rated on a seven-point Likert scale (1: strongly disagree, 7: strongly agree). The results indicated that the students were less likely to consider the brand origin when

purchasing toothpaste compared with the other nine products. Thus, toothpaste was used as the target product in the two studies.

Two real brands were used in this study. Darlie and Colgate represented the local and global brands, respectively. Both brands are established with known product quality and credibility. The first page of the survey contained the manipulation of the brand origin. In the company description, the local brand condition indicated that Darlie is a Taiwanese company, whereas the global brand condition indicated that Colgate is an international company. For the code-switched advertisements, according to the method used for designing code-switched advertisements in previous studies (e.g., Ahn and Ferle 2008; Lin, Wang, and Hsieh forthcoming), code-switching was manipulated by using the foreign language for the brand name, headline, and body copy, but the local language was used in the rest of the written content. For the noncode-switched advertisements, all advertisement content was presented in the local language. Four versions of the advertisements were developed and

formatted identically, featuring a brand name, bold headline, product visual1, and product name in the same order. The brand name (Darlie or Colgate) was placed on the top left corner of the

advertisement. The center of the body copy featured a bold headline (“Freshness, Nature, Nice Breath”) and a subheadline (“New on the market”). The bold headline was followed by a picture of the toothpaste, which showed the brand name and the attribute “natural herb.” The bottom of the advertisement showed the product name (“Darlie toothpaste” or “Colgate toothpaste”) and a smaller product visual (the same as the visual shown in the center). The differences among the four

advertisements were the brand (Darlie vs. Colgate) and the language (Chinese vs. English) used in the brand name, bold headline, brand name on the product visual, and product name. The

advertisement content of the four advertisements is presented in Appendix 1.

The local and global identity priming tasks were designed by adapting the sentence-completion task developed by Srull and Wyer (1979) and Zhang and Khare (2009). The respondents formed meaningful sentences from sets of scrambled words. Under the global priming condition, respondents completed five sentences related to global identity (e.g., “I belong to the whole world”) and four sentences of filler questions (e.g., “The sky is blue”). By contrast, under the local priming condition, respondents completed five sentences related to local identity (e.g., “I belong to the local

community”) and the same four filler questions. In addition, we included a control priming condition. Respondents completed five sentences related to environmental identity (e.g., “I belong to the natural environment”). The detailed priming tasks are presented in Appendix 2.

Procedure. A total of 550 adults (age = 19-74 years, M = 37.81 years; men = 36.2%) were recruited from an online panel, similar to the Amazon Mechanical Turk, in Taiwan to participate in this online experiment in exchange for NT$100 (i.e., $3 USD). They were randomly assigned to 12 between-subjects conditions.

This main study was divided into two parts. In the first part, respondents completed the identity priming tasks and then indicated the focus of their thoughts as a manipulation check (Zhang and Khare 2009). We used the following three statements: “For the time being, I mainly identify myself as a (an)…,” “At this moment, I feel I am a (an)…,” and “On my mind right now are thoughts of being a (an)...” Each item was rated on a seven-point scale (1 = global citizen; 7 = local citizen). The items were averaged to generate a composite score ( = .85). A higher manipulation check composite score indicated a higher local identity.

When they began the second part of the study, respondents were informed that the study purpose was to understand consumer evaluations of a print advertisement for an existing toothpaste brand. They viewed the advertisement and indicated their attitude toward the advertisement and the

product. In addition, the respondents’ advertisement comprehension was measured on the basis of their responses to three items (described in the following subsection) related to the language used in the advertisement to eliminate the possibility that their advertisement comprehension affected the evaluation of the advertisement. These data were then included as a covariate in analysis.

Subsequently, the participant’s advertising involvement was measured. Finally, information on gender and age was collected in the end of the questionnaire.

Dependent Measures. Advertisement and product evaluation scales were adopted from Luna, Lerman, and Peracchio (2005) and Luna and Peracchio (2005a; 2005b), respectively. The

advertisements were evaluated using six 7-point semantic-differential scales (bad/good, dislike/like, uninteresting/interesting, unpleasant/pleasant, negative/positive, and poor quality/excellent quality;  = .92). The products were also evaluated using six 7-point semantic-differential scales (poor

quality/high quality, not appealing at all/very appealing, I would not buy it/I might buy it, I would not recommend it to a friend/I would recommend it to a friend, inferior/exceptional, and very bad/very good;  = .92). A higher rating indicated a more positive advertisement or product evaluation.

The effectiveness of the manipulation of the brand origin was assessed on the basis of the participant’s indication of the advertised product’s brand origin. To control for the effect of advertisement comprehension on advertisement and product evaluations, advertisement

comprehension was measured using three questions adapted from Chang (2009; 2013) and was rated on a six-point Likert scale (I understand the meaning of the words in the advertisement, I am familiar with the language used in the advertisement, and I am proficient in the language used in the

advertisement;  = .85). Advertising involvement was also rated on a three-item, six-point Likert scale, adapted from Miniard et al. (1991) (I was involved with the advertisement content, I