The Effects of Religiosity and Attitude on Consumer Boycotts

Boycott is one of the manners in which consumers portray their hatred or disagreement with companies’ corporate social responsibility (CSR) policies and practices and thus is unwelcome by companies. In spite of the massive rise in consumer boycotts, marketing pays little attention to this issue. Addressing this deficiency, this study uses the value-attitudebehaviour model to examine how religiosity and attitude affect consumer boycotts. The sample used in this study is the Malaysian consumer, with a total of 330 respondents. Data for this study was collected using an online survey and analysed using partial least square structural equation modelling. The findings discover that consumer’s religiosity and attitude positively influence consumer boycotts. The study also revealed that attitude plays a mediating role between religiosity and consumer boycotts. This study has made significant contributions to both theory and practice. From a theoretical perspective, this study fills in the literature gaps and enhances understanding on how religiosity and attitude affect consumer boycotts. This article provides insights to improve companies’ CSR practices and policies to avoid consumer boycotts in the future.


Introduction
There are studies show that consumers are increasingly concerned about the environment, for example, with a large number of consumers in the United Kingdom agreeing to reject brands that affect the environment in order to achieve long-term sustainability (Kantar, 2019). It is not impossible for consumers to expect more from the brand, and at the level of individual action, people are more willing to boycott the brand. In addition, Kantar (2019) argues that 924 (77 %) out of 1,200 consumers are willing or likely to consider switching brands, avoiding or boycotting certain products on the basis of brand sustainability and environmental policies. Due to growing pressure from consumers on companies to embrace corporate social responsibility (CSR), companies worldwide have adopted CSR to stay competitive in business. The right CSR initiatives are able to provide sustainable competitive practices. Despite a growing interest on boycott's and its critical importance in today's competitive business environment, consumer boycott is still underexplored (Cromie & Ewing, 2009;Lee et al., 2009) and thus research that focused on this issue remains scarce (Zeng et al., 2020;Hoffmann & Műller, 2009). In addition, evidence of religiosity and attitude influence on boycott participation have not been fully addressed, especially from the standpoint of Malaysian consumers. Based on the value-attitude-behaviour (VAB) model (Homer & Kahle, 1988), this study tries to answer the following research questions: (a) How religiosity and attitude affect consumer boycotts?, (b) Does attitude mediates the relationship between religiosity and consumer boycotts? This study seeks to fill in the literature gap by examining how religiosity and attitude affect consumer boycotts as well as to assess the mediating role of attitude between religiosity and consumer boycotts. This study provides significance contributions to both theory and practice. For theoretical significance, this study enhances understanding on the effect of religiosity and attitude on consumer boycotts. From a managerial significance, the findings provide guidelines to help managers and policy makers to improve their CSR implementations that would help their companies to gain sustainable competitive advantage.

Determinants of Consumer Boycotts
Consumer boycotts can be defined as an act in which one or more parties want to achieve certain objectives by urging consumers to refrain from making selected purchases in the marketplace (Friedman, 1985). Thus, it can be said that consumer boycotts are a form of modest, nonviolent action against economics, political, and social injustices that are conducted in the marketplace (Sasson, 2016). They can be seen as collective actions to confront the corporation to coerce corporate change (Yang & Rhee, 2020). Consumer boycotts express severe consumer dissatisfaction with a firm (Shaw & Newholm, 2002) and are recognized as catalyst for social and economic change. Research has shown that they can have dramatic consequences for companies, such as negative effects on stock price and impacting company's image (Luo, Zhang & Marquis, 2016;Omar et al., 2019). Boycotting is also known as anti-consumption behaviour (Sen, Gurhan-Canli & Morwitz, 2001;Yuksel, 2013;Yuksel & Mryteza, 2009). Consumer boycott could be defined as an expression of animosity that results in an unwillingness to make selected purchases in the market due to various reasons such as military violence, political injustice, economic inequality practices or a firm's irresponsible behaviour (Abd-Razak & Abdul-Talib, 2012). There are two categories of consumer boycott and these include micro boycott that refers to boycott towards a firm and macro boycott that describes a boycott towards a country (Friedman, 1999;Klein, et al., 2004). Boycott is a consumer's volunteer action of not purchasing a certain product from a certain company or country altogether. Boycott is also the individual result of boycott campaigns that can target either a specific company or all companies of a given country (Abosag, 2010).

Religiosity
Religion is an organized system of beliefs, practices, rituals and symbols, which allows one to be closer to the sacred or the transcendent (God, a supreme power, ultimate reality or truth), and to understand one's own responsibility and relation to others in a community (Koening et al., 2000). More simple and precise religiosity is defined as "individual preferences, emotions, beliefs, and actions that refer to an existing (or self-made) religion" (Stolz, 2009, p. 347). The definition of religiosity from consumers' perspective, is "the degree of being religious" (Cyril De Run et al., 2010). Religion is a part of a culture, and significantly influence the peoples' systems of values and habit. On the other hand, religion also influences their lifestyle such as dictates their buying decision and consumption behavior (Dekhil, Jridi & Farhat, 2017). Furthermore, they found that highly religious people are more likely to be involved in boycotting a brand, and this result is consistent with the findings of Al-Hyari et al (2012). Furthermore, Giorgi and Marsh (1990) found that religion and the degree of religiosity of individuals have a positive effect on their moral behaviors and this is in line with Farah (2014) who showed that religiosity significantly motivates boycott and the way consumers perceive the brand image, which will then affect the way consumers judge these products. Abou-Youssef et al., (2015) stated that religion shapes attitudes and how religiosity grows are of inquiry. In addition, numerous social studies have shown that religion has a positive effect on people's lives (Abou-Youssef et al., 2015) and several studies have also shown that religion affect consumer attitudes and behaviour (Armstrong, 2001;Arnould, Price & Zikhan, 2004;Burkett, 1980;Cochran and Cohen, 1983;Delener, 1994;Hawks and Bahr, 1992;Matitila et al., 2001;Pettinger et al., 2004).

Attitude
In this study, attitude refers to attitude towards responsible companies. One of the essential aspects of attitudes is that they have been learned (Abou-Yousseff et al., 2015). In the marketing context, the attitudes which have a link with purchasing behaviours, actual experience and product testing. Attitudes have a motivational factor, as they may lead consumers to certain actions or may jeopardize certain behaviours (Fazio, 1986;Schiffman and Kanuk, 2000). In addition, consistent with VAB model, attitude is projected to mediate the relationship between religiosity value and boycott participation.

Theoretical Framework and Hypotheses Development
A deeper understanding of the consumer decision-making process in the context of boycotts is crucial. Among the various expressions of socially responsible consumption, boycotting has been considered to be "the most deliberate form of ethical purchase behaviour" (Delistavrou et al., 2019;Smith, 1987). Following this meaningful insight, it is proposed that boycotting is a form of socially responsible consumption. In the case of boycotting, a consumer who participates in the boycotting campaign is involved in a problem-solving process. This behaviour is, by nature, a case of value-attitude-behaviour cognitive hierarchy. Therefore, the VAB model (Homer & Kahle, 1988) was considered as the appropriate theoretical framework to study consumer boycott. In social psychology and marketing studies, the VAB model has been widely used to understand behaviour (e.g., Kang, Jun & Arendt, 2015;Honkanen et al., 2006;Vaske & Donnelly, 1999). Based on the VAB model, value-attitude-behaviour are organized in cognitive hierarchy sequence, in which values influence behaviour indirectly through attitudes. Homer and Kahle (1988) found causal sequence from abstract cognitions (i.e., values) to mid-range cognitions (i.e., attitudes) to specific behaviours. Applying the VAB model in the context of this study, it is proposed that religiosity (value) influences boycotts (behaviour) through attitude towards responsible companies (attitude). With reference to the VAB model, the following theoretical framework and hypotheses were proposed for this study: Therefore, based on the VAB model and past literature, the following hypotheses were developed: H1 Religiosity has a positive influence on consumer boycotts H2 Religiosity has a positive influence on attitude H3 Attitude has a positive influence on consumer boycotts H4 Attitude mediates the relationship between religiosity and consumer boycotts

Methodology
Sampling technique and sample size determination A quantitative research with a convenience sampling method was used in this study and this sampling is highly recommended due to the target population being unknown (Hulland et al., 2017). The survey data was collected through the online survey questionnaire that was presented in the Google Forms format and made available via an internet connection within a month. This study utilized G*Power 3.1.9.2 software to reach the minimum sample size needed. The sample size necessary for one independent variable and two dependent variables using effect size (f 2 ) of 0.15 (medium) is 107 respondents. However, a total of 330 questionnaires were collected from Malaysian consumers.

Measurement Items
The distributed questionnaire consists of four sections. The instruments involving 14 items related to the research framework were adopted from existing literature and refined based on the context of this study. In section A, the respondents were asked about boycott participation that was adopted from Farah (2014), boycotting unsafe products, unhealthy products from companies or certain countries that support inhuman activities and the way to express his/her anger to them with five items by using seven-point Likert-scale (1= strongly disagree and 7= strongly agree). In section B, the religiosity (value) and attitude were measured through a seven-point Likert-scale (1= strongly disagree and 7= strongly agree) with three items for religiosity (value) adopted from Bakar, Lee and Hashim (2013) and six items for attitude adopted from Chen and Kong (2009). In section C, the respondents were asked about their gender, age, marital status, occupation, monthly income, educational level, race, religion, family size, work experience and residence. Then, by using purposive sampling technique a pre-test acceptable for the quotation of comments and recommendations from respondents regarding potential defects including the questionnaire format, design, and wording.

Data Analysis
The data does not have the normality requirement when it is analyzed with Smart PLS. However, as recommended by Hair et al. (2017), this study was tested for the multivariate normality by looking at the skewness and kurtosis using the software available at: https://webpower.psychstat.org/models/kurtosis/results.php?url=5520c7e5d0ef1285e1b17 1728b1b6641. If the distribution is normal, the values of skewness and kurtosis are zero. Kurtosis values that are more than zero display a distribution that is too peaked with short, thick tails, and kurtosis values that are low than zero display a distribution that is too flat (also with too many cases in the tails). Non-normal kurtosis demonstrates an underestimate of the variance of a variable (Tabachnick & Fidell, 2007). The result exhibits that the Mardia"s multivariate skewness (β = 5.799, p < 0.01) and Mardia"s multivariate kurtosis (β = 22.903, p < 0.01). Hence, the data was slightly not normal and it is appropriate to apply the Smart PLS software in this study. The Partial Least Squares (PLS) analysis (using the SmartPLS 3.2.9 software) was used to analyze the measurement model (Ringle et al., 2015). There are two suggested stage for analytical procedures. First is the measurement model (validity and reliability of boycott (BCT), religiosity (RLG) and attitude (ATT)) (Anderson & Gerbing, 1988). Second, the examination of the structural model (testing the hypothesized relationship) (Hair et al., 2014). Next, is to test the common method variance (CMV), which disclosed that it should be a problematic issue for the study if the data were collected from a single source (MacKenzie et al., 2011). To overcome the CMV, the study applied the statistical method. For the statistical method, a full collinearity analysis was done. The full collinearity analysis is considered to be a comprehensive procedure for the simultaneous assessment of both vertical and lateral collinearity (Kock & Lynn, 2012;. Table 3 presents the resulting VIF values for all latent variables (boycott participation, religiosity and attitude) in the model that were less than 3.3. Hence, this result confirms that the CMV is not a serious threat to the study.

Demographic Profile
As can be seen in Table 1, 238 respondents (72.1%) were female and the rest (92, 27.9%) were male. In regards to age, the largest proportion (161) of the respondents (48.3%) were aged between 20 to 29 years old and this is followed by 64 respondents below 20 years old and only one oldest respondent that aged between 60 years old and above. As indicated in

Measurement Model
Under the reflective model, 2 types of validity are assessed: 1) Convergent validity and 2) Discriminant Validity. Convergent validity is the degree to which indicators of a specific constructs converge or shares a high proportion of variance that are in common (Hair et al., 2010, p.771). As suggested by Hair et al. (2017), factor loadings, and AVE are used to assess convergent validity. The indicator loadings, CR and AVE of the reflective constructs are shown in Table 2. All the loadings which exceed the recommended value of 0.5 (Hair et al., 2017) are retained. Moreover, all three constructs meet the threshold values/minimum cut-off values for CR and AVE, where all CRs are greater than 0.7 and all AVEs are greater than 0.5 (Hair et al., 2017). Item that did not fulfil the criteria outlined above were deleted from the model such as RLG3. It is concluded that the constructs meet reliability and convergent validity requirement at this stage. Subsequently, discriminant validity of the model is assessed. Indicators should load more strongly on their own constructs than on others in the model, and the average variance shared between the construct and its measure should be greater than the variance. According to Franke and Sarstedt (2018), the heterotrait-monotrait (HTMT) ratio is a stringent method to confirm the discriminant validity of a study if it made use of Smart PLS. Referring to Franke and Sarstedt (2018), the HTMT ratio is confirmed if the values of all variables in the study are <0.85. Since all the HTMT values were lower than 0.85, it was an indication that the  Table 3 illustrates the results of the HTMT ratio analysis. Prior to evaluating the structural model, it is vital to confirm that there is no lateral collinearity issue in the structural model. All the inner VIF values for all the variables (religiosity, attitude and boycotts) that need to be examined for multicollinearity are less than 3.3, indicating that the multicollinearity is not a concern in this study (Diamantopoulos & Siguaw, 2006). In order to analyze the hypotheses of this study, a bootstrapping technique with a resampling of 5000 was applied as proposed by Hair, Sarstedt and Ringle (2019). Based on the assessment of the path coefficient as shown in Table 4, all three relationships are found to have a t-value >1.645, thus significant at 0.05 level of significance. Specifically, the predictors of religiosity (β = 0.336, p<0.01) and attitude (β = 0.334, p<0.01) are positively related on boycott, which explains 33.2% of variances in consumer boycotts. Thus, H1 and H3 are supported. The R 2 value of 0.332 is above the 0.26 value, that indicates that is a substantial model as suggested by Cohen (1998). Next, the result of the effect of attitude on consumer boycotts indicates that attitude (β = 0.438, p<0.01) is positively related to boycott participation, explaining 19.2% of the variance in the boycott. This result supports the H2 of this study. The R 2 value of 0.192 is above the 0.13 value that indicates the moderate model as suggested by Cohen (1988). Next, the effect sizes (f 2 ) are assessed. To measure the effect size, the Cohen (1988) guideline is used. The values of 0.02, 0.15 and 0.35 represent small, medium and large effects respectively (Cohen, 1988). From Table 4, it can be observed that religiosity has a medium effect in producing the R 2 for attitude (0.237). Furthermore, the result indicates that religiosity (0.136) and attitude (0.143) have small effects in producing the R 2 for boycott.  Shmueli et al (2019) proposed that PLS predict is a hold-out sample-based procedure that generates case-level predictions on an item or a constructs level by using the PLS predict with a 10-fold procedure to check for predictive relevance. Shmueli et al (2019) also suggested that if all the item differences (PLS-LM) were lower than LM, then there was a strong predictive power. If all the item differences are higher than LM, then the predictive relevance is not confirmed. If the majority of item differences are lower than LM, then there is a moderate predictive power. If the minority of item differences are lower than LM, then there is a low predictive power. As can be seen in Table 5, the results of the analysis are mixed. Based on guidelines by Shmueli et al. (2019), for attitude, the result of this study showed that the predictive power is moderate, whereas for boycott, the predictive relevance is low.  Preacher and Hayes (2008) which proposed bootstrapping in the indirect effect. The result for this study found that attitude mediates the relationship between religiosity and boycott (β = 0.151, p<0.01), hence H4 supported.

Discussion
In today's market, consumer boycott refers to something that is related to the belief that consumers will use their purchasing power to boycott the target's behaviour as well as the belief that the boycott will succeed in forcing the target to discontinue its egregious behaviour. This study has three main findings. Firstly, the results showed that consumers are willing to boycott when it concerns with the misbehavior of offering the unsafe or unhealthy product. Secondly, the results for this study demonstrated that religiosity positively influences the attitude and boycott. Thirdly, attitude mediates the relationship between religiosity and boycott participation. One possible reason the result showed that religiosity positively influences the boycott participation could be due to the fact that the majority of the respondents in this study are Malays and Muslim. In Malaysia, Malays are usually born as Muslims. Furthermore, the striking use of religiosity appeals in a boycott campaign, and the rising Islamic conservatism among Muslim consumers would place religiosity at the core of the group's motivation and participation in a boycott (Muhammad, Kamarudin & Fauzi, 2018).
Another key objective of this study was to examine whether attitude mediates the relationship between religiosity and boycott participation and it was examined within the framework of the VAB model. Therefore, the hypothesis in the study is consistent with Homer and Kahle (1988), which posit that from religiosity (value) affects attitude, which in turn to specific behaviors (boycott). This sequence can be described the value attitude behaviour hierarchy.

Conclusions and Recommendations
Consumer boycotts of brands and manufacturers have become commonplace in today's competitive business environment. Effective boycotts are considered to be one of the most important means by which consumers can induce socially responsible conducts/ethical business practices and put pressure on companies to introduce sustainable developmental actions (Delistavrou et al., 2019;Lavorata, 2014). Yet, CSR and consumer behaviour literature have not paid attention on the rising severity of consumer boycotts. Thus, the literature on this issue is rather limited. Understanding the effect of religiosity and attitude on consumer boycotts would therefore enhances understanding on factors impacting consumer boycotts as wells as how these factors (i.e., religiosity and attitude) impact consumer boycotts. This study establishes that religiosity and attitude have positive impacts on consumer boycotts. In addition, this study confirms the mediating role of attitude in the relationship between religiosity and consumer boycotts. The findings provide support for the VAB model.
This study provides several theoretical and managerial contributions. For theoretical implications, this study contributes to the literature on consumer boycotts in the context of developing countries. Moreover, this study has shown the significant influence of religiosity and attitudes on consumer boycotts among consumers in Malaysia. This study extends the study conducted by Farah (2014) in that this study examined boycott behaviour instead of boycott intention. In addition, this study also extends the study conducted by Dekhil et al. (2015) in such a way that this study examines the mediating role of attitudes on the relationship between religiosity and consumer boycotts. From a managerial perspective, this study provides CSR managers, boycott organizers as well as policy makers with information on how to manage consumer boycotts. Moreover, this study proposes a tool for measuring ability to participate in boycotts for sustainable development as perceived by consumers. This will allow companies to determine what action they need to take to incorporate sustainable development into their practice. While it seems exciting for companies to know determinants of consumer boycotts, the findings suggest that religiosity and attitudes play important roles in consumer boycotts. First, managers need to take note that although some consumer boycotts are punitive in nature, many are not. Therefore, companies need to rethink and transform their business model towards more sustainable business models that are free from unethical business practices such as producing unsafe and unhealthy products, exploitations of human and animal rights, environmental degradation, etc. This will enhance consumers' attitude towards companies' products and services. Second, companies should improve the overall marketing mix to enhance consumers' attitudes, satisfaction and loyalty towards companies' products and services. This will diminish the impact of religiosity and attitude on boycotts.
This study is not immune to some limitations. First, although the study establishes that religiosity and attitude have positive impacts on consumer boycotts. There are other factors that also contribute towards consumer boycotts. For instance, the effect of social media and social norms on consumer boycotts warrants further investigation. Additionally, whilst this study establishes the mediating role of attitude, there remain other mediating factors that may also affect the relationship between religiosity and consumer boycotts. Second, this study uses an online survey. Future study may employ in-depth interviews in order to gain detailed information regarding the reason why consumers are willing to boycott certain products from certain countries and/or companies. This sampling technique is also more appropriate to be used in identifying the religiosity factor in an in-depth manner because this factor is quite sensitive and the respondents may want to share and explain in detail why they get involved with boycott. Moreover, using an in-depth interview provides a richer and detail responses on how other factors affect consumer boycotts would also be further revealed, it may be suitable to.