Consumers Become More or Less Ethnocentric? A Meta-Analysis on Level of Consumer Ethnocentrism

A huge body of research on consumer ethnocentrism has occurred in cross-cultural consumer behavior research area since the seminal work of Shimp and Sharma (1987). There is, however, a research gap on meta-analysis of the level of consumer ethnocentrism. This study seeks to address this gap by employing, as far as we are aware, the first meta-analysis on level of consumer ethnocentrism. we draw several conclusions with meta-analytical data of 153 mean values in 87 articles during the period of 1987 to 2013 (N = 42840): (1) The average score of consumer ethnocentrism is 3.58 (7 in total); (2) General consumers are more ethnocentric than student consumers; (3) Consumers in developing countries are more ethnocentric than consumers in developed countries.


Introduction
Image following scenario: If you are going to buy a TV in Wal-Mart, there are many of potential brands that you can choose, such as Sony from Japan, Samsung from South Korea or TCL from China. While you will find there are no essential differences in technological advancement and external design for same type among these brands. How would you make your purchase decision in this case? Ah ha! Of course! You probably buy a TV made in your home country! With the rapid development of technology in the world today, manufactures are confronted with many challenges such as products homogeneity (Cheng et al., 2013;Gassmann & Keupp, 2007). Consumers buy something not only depending on the price, quality or design, but also depending on social norm that whether it is appropriateness and morality to buy foreign-made products (Supphellen & Rittenburg, 2001;Vida & Reardon, 2008). This belief is well known as consumer ethnocentrism, which is becoming more and more important in consumer purchase decision for transnational consumption (Cleveland et al., 2009;Shimp & Sharma, 1987).
Since the seminal work of Shimp and Sharma (1987) on concept definition and measurement scale development of consumer ethnocentrism, a growing body of research on this theme has fostered in cross-cultural consumer behavior research area (see a review of Shankarmahesh, 2006). There is, however, a research gap on meta-analysis of the level of consumer ethnocentrism. Consumers become more or less ethnocentric? Which country has the high level ethnocentric consumers and which country is relatively low? How about the differences for level of ethnocentrism between consumers in developing countries and consumers in developed countries? Are there any differences for level of ethnocentrism among general consumers and student consumers? All of which still demand an explicit answer.
This study seeks to address these gaps by employing, as far as we are aware, the first meta-analysis on level of consumer ethnocentrism. The first contribution of this study is achieving the mean value of consumer ethnocentrism from 87 articles during the period of 1987 to 2013. Second, this study compares the level of consumer ethnocentrism horizontally among five most popular countries and shows the longitudinally trend of these five countries. Last but not the least, this study investigates the differences of mean consumer ethnocentrism in different respondents and countries with different economic development levels.

Literature Review and Hypotheses
When facing with purchase choices between domestic and foreign products, consumer ethnocentrism represents consumers' tendency that whether it is appropriateness to buy foreign-made rather than home country made products (Balabanis et al., 2001;Deb & Roy Chaudhuri, 2012;Sharma, 2011;Shimp & Sharma, 1987). This tendency is derived from consumers' feelings of ethnocentric insecurity (Siamagka & Balabanis, 2015); they think buying foreign products may trigger unemployment and other social problems in their home country (Shimp & Sharma, 1987). With such ethnocentric insecurity perception, they view other countries based on their ethnic group interests and, therefore, hold distorted cognition towards products from these countries (Sharma, 2014;Siamagka & Balabanis, 2015). As a result, they prefer products made in their home country to other countries, even regardless of the quality and price (Klein, 2002;Shoham & Brenčič, 2003). In other words, consumer ethnocentrism represents an overall favorable attitude towards domestic products instead of foreign products, which consisting of an affinity for domestic products in affective aspect, an evaluation bias in favor of domestic products in cognitive aspect and, as a result, a preference for domestic rather than foreign products in behavioral aspect (Sharma, 2014;Shimp & Sharma, 1987;Siamagka & Balabanis, 2015).
As a form of prosocial behavior (Siamagka & Balabanis, 2015), consumer ethnocentrism can be learned from an early age and thus, become a kind of enduring general traits (Sharma et al., 1995;Shimp & Sharma, 1987;Siamagka & Balabanis, 2015). While the level of consumer ethnocentrism is not always stable but may change over time (Good & Huddleston, 1995). When consumers at a relatively young age, especially for these young students who are receiving education, they are intensively curious about the world outside and thus, they are less likely to devote themselves to definite ways of life and certain domestic products (Cleveland et al., 2009). As they are getting older, they are aware the adverse consequences that may be caused by import products (Javalgi et al., 2005). The best way to avoid these unfavorable things occurring is buying local and domestic products rather than foreign products (Huddleston et al., 2001;Javalgi et al., 2005). Therefore, compared to younger student consumers, older general consumers are more likely to buy products made in their home country and, certainly, are hold higher level of ethnocentrism (Huang et al., 2010;Javalgi et al., 2005). In light of these arguments, we propose the following: H 1 : General consumers are more ethnocentric than student consumers.
In addition to time's effect on ethnocentrism, surrounding conditions of consumers will also have an effect on their ethnocentrism (Shankarmahesh, 2006). Consumers will show different ethnocentrism level with changes of surrounding economic conditions of their home country (Siamagka & Balabanis, 2015). For consumers in less developed countries, their feelings of ethnocentric insecurity probably higher than consumers in developed countries (Kucukemiroglu, 1999;Sharma et al., 1995). They may hold the belief that these developed countries are doing business unfair with their home country (Klein et al., 1998). Such insecurity perception can prompt consumers to take some actions so that other countries cannot get rich of them. Thus consumers in a developing country are more likely to be influenced by some patriotic events such as "buy local" promotional campaigns (Hamin & Elliott, 2006), as a result, they may hold high level of ethnocentrism (Bandyopadhyay et al., 2011). In light of these arguments, we propose the following: H 2 : Consumers in developing countries are more ethnocentric than consumers in developed countries.

Method
We choose articles according to following criteria: (1) the articles must be published in English journals, (2) the research must be quantitative study, (3) mean value of consumer ethnocentrism has been reported, (4) sample size has been reported in corresponding study, (5) each study has exclusive respondent sample.
By searching of databases EMERALD, ELSEVIER, ABI/INFORM and searching on "Web of science" "Google scholar", we finally achieve 153 mean values (effect size) in 87 articles (Note 1) during the period of 1987 to 2013 (N = 42840). Specifically, respondents among these studies are mainly from US (17 studies), China (13 studies), India (13 studies), South Korea (11 studies) and Russia (8 studies).
The instrument for measuring consumer ethnocentrism is the CETSCALE that contains multi items with Likert-type statements (Shimp & Sharma, 1987). To make the mean value of consumer ethnocentrism in different studies comparable, we conduct mathematic conversion as follows: (a) If the reported value is an average score with a seven-point Likert scale, we select it directly; (b) If the reported value is an average score with a five-point Likert scale (Note 2), we transfer it by following formula: The mean value we need = (the reported average score) * (7/5).
(c) If the reported value is a total score with a seven-point Likert scale, we transfer it by following formula: The mean vale we need = (the reported total score) / (the number of items).
ijbm.ccsene (d) If the r and then fo

The Le
The averag in 87 artic countries (

The Le
Among th extracted a countries a Indian con consumers only one l there is si significant

The Tr
We then in the same y new mean shows the American in India.

Control variables
Published time: a year names a number. Due to only two articles were published before 1995, one was in 1987 and another was in 1991, these two name 1. From 1995 to 2013, we give the number 2 to 20 for each year successively.
Journal influence: we use the Article Influence scores in http://www.eigenfactor.org, which is a widely recognized journal evaluation method.

Independent variables:
Sample type: we name general respondents and student respondents as 0 and 1 respectively.
Development level of consumers' home country: we name developed markets as 1, developing markets as 2.

Dependent variable:
Consumer ethnocentrism: the mean value.
We use SPSS 21 to test our hypotheses. The results can be found in table 1. The results show that sample type is negatively related with consumer ethnocentrism (M 2 , β = -0.16, p < 0.10). It indicates that compared to sample of young students, general consumers are more ethnocentrism. Thus, hypothesis 1 is supported. Furthermore, result of T-test shows the mean of consumer ethnocentrism for general respondents is significantly larger than mean for student respondents (M general = 3.69, SD = 0.85; M student = 3.38, SD = 1.02; t (151) = 1.97, p < 0.05), further supporting hypothesis 1.
Moreover, development level is positively related with consumer ethnocentrism (M 3 , β = 0.22, p < 0.01). It means consumers in developing market are more ethnocentrism than those in developed market. Thus, hypothesis 2 is supported. Furthermore, result of T-test shows the mean of consumer ethnocentrism of developing market is significantly larger than mean of developed market (M developing = 3.78, SD = 0.88; M developed = 3.35, SD = 0.92; t (151) = -2.96, p < 0.01), further supporting hypothesis 2.

Conclusion
The research note of meta-analysis on level of consumer ethnocentrism can be mainly concluded that: (1) Mean of consumer ethnocentrism is 3.58 with 7-point Likert-type scale based on 153 effect size in 87 articles during the period of 1987 to 2013; (2) General consumers are more ethnocentric than student consumers.
(3) Consumers in developing countries are more ethnocentric than consumers in developed countries.
(5) During the period of 1987 to 2013, the levels of ethnocentrism for American and India consumers have ijbm.ccsenet.org International Journal of Business and Management Vol. 12, No. 2;2017 experienced slight increase, while the trend of ethnocentrism likes an inverted "V" for Chinese consumers, and shows large fluctuation in the first half and then keeps stable for South Korean consumers. Russian consumers, however, keep steady trend of ethnocentrism at a relative low level.

Limitations and Future Research
There are few limitations in this research.
(1) Although the samples are different in studies that we selected for meta-analyzing, some general consumer samples may contain students due to the situation that general consumers and student consumers in extant research are not strictly mutual independent. Future research should pay more attention to distinguish the two samples and to uncover the different effect of age diversity on consumer ethnocentrism.
(2) This research focus on articles that published from 1987 to 2013, during the period of collecting data, we had tried our best to collected relevant studies. Some studies, however, may still be ignored and are not included in this meta-analytical note. Future research can include and examine those unheeded studies on consumer ethnocentrism.
(3) In this article, we mainly focused on the level of consumer ethnocentrism among different countries. There are more issues of consumer ethnocentrism need to further explore. For example, a comprehensive meta-analysis on relationships of consumer ethnocentrism and its antecedents and consequences is a meaningful future research avenue.