INFLUENCE OF CONSUMER ATTRIBUTIONS AND SERVICE QUALITY ON SUPPORT OF CORPORATE SOCIAL RESPONSIBILITY

. This study examines whether or not a consumer will support a firm due to its CSR activities depending on what perceived motivation (egoistic, strategic, stakeholder and value driven attributions) the consumer has towards those activities and also analyzes how the relationship between consumers’ attributions and the change in their supportive behaviour according to their perceived service quality. The survey was conducted with customers of mobile telecommunication services in Turkey. In order to collect data to test the hypotheses, in accordance with the theory of stated preferences versus theory of revealed preferences, a face-to-face survey was carried out by adopting a quota sampling technique based on the market share of each mobile service provider. The sample consisted of 400 respondents who were drawn from the Istanbul metropolitan area. The results show that value driven, strategic and stakeholder driven motives have a direct and positive effect. Egoistic motive was found to be significant and had a negative effect. Service quality moderated only the relationship between strategy driven motives and supportive behaviour, and value driven motives and supportive behaviour, but in an unexpected direction.


Introduction
In today's competitive market environment, companies are looking for different strategic implications with the aim of distinguishing themselves from their competitors and, more importantly, enhancing their brand image. The corporate social responsibility (CSR) initiatives are one of the most common ways which firms use to influence consumers, differentiate product offerings and have higher brand equity. For instance, companies in the tobacco and oil industries (e.g., BP, Exxon, Shell and Monsanto) at-tempted to enhance their image by stressing their environmental and social initiatives (Arnold, 2001;Yoon et al., 2006).
Even though CSR is used greatly by companies, its common and accepted definition is not available. One of the first definitions of CSR was given by Bowen (1953cited in Türker, 2009, who said that CSR is the obligations of businessmen "to pursue those policies, to make those choices, to make those decisions, or to follow those lines of action which are desirable in terms of the objectives and values of our society" (p. 6). Since the time this definition was formulated, many different definitions have occurred. The most extensively accepted conceptualization of CSR was given by Carroll (1979), and according to him; "CSR's driving force is society's expectations". His famous pyramid of responsibilities has four dimensions: economic and legal responsibilities that are "required", whereas ethical responsibilities are "expected", and philanthropic responsibilities are "desired" by society.
Most organizations have the supposition that consumers will reward firms which support social programs, and for that most firms engage in CSR activities (Levy, 1999). However, prior research (Barone et al., 2000;Brown & Dacin, 1997;Creyer, 1997;Ellen et al., 2000;Sen & Bhattacharya, 2001) showed that consumers do not always accept these social initiatives as sincere actions, and they question why those firms are engaging in social programs and whether it is for the interest of the community or to the benefit of themselves. On the other hand, studies examining the impact of consumer suspicion on consumer behaviour related to CSR efforts are limited. Therefore, the main aim of this study is to understand whether or not a consumer will support a firm due to its CSR activities depending on what perceived motivation (egoistic, strategic, stakeholder and value driven attributions) the consumer has towards those activities. The other purpose of the study is to verify how the relationship between consumers' attributions and their supportive behaviour change according to their perceived service quality. Quality is accepted as one of the most important contributors to consumer's preference for service providers, and it is an important input to customer satisfaction (Caruana, 2002). Nevertheless, investing only in quality does not guarantee positive outcomes such as satisfaction, retention and loyalty. As Zeithaml et al. (1996) suggested, service quality is just one of many variables like pricing, advertising and image that simultaneously influence profits, so interactions among the variables and strategy must also be considered. These implications guide this study toward regarding perceived quality as a moderator variable to understand whether higher (lower) quality perception changes the impact of perceived motivation on support.
The structure of this paper is as follows: next section reviews CSR in Turkey, the conceptualization of consumer attributions, service quality, supportive behaviour and hypotheses formulation. It is followed by a section which outlines the research methodology employed, validity issues, moderating effects and discusses key findings from the empirical analysis. Finally, the paper offers conclusions, considers managerial implications and future directions.

CSR in Turkey
Even though much less is known about CSR in the developing economies than in the developed ones, studies which have been carried out so far emphasized two major points (Li et al., 2010, p. 636): "First, firms in developing countries adopt CSR less than their counterparts in the developed world (e.g., Welford, 2004), and second, the main reason for this gap is due to their low economic development levels (e.g., Baughn et al., 2007)". However, there is not enough study about consumer attributions regarding CSR in the context of emerging economies. For this reason, managers of firms operating in emerging economies all need to gain a better understanding of CSR attributions. Taking all the above into consideration, this study was conducted in Turkey. Turkey's philanthropic history has been rich since the Ottoman era, (Ararat, 2004, p. 255). Also, there is a strong philanthropic tradition in the business community. However, CSR efforts have not been chiefly various (Türker, 2009, p.412). This is not surprising because Turkish society primarily supposes economic performance (jobs) from the companies (Ararat, 2008, p.5). On the other hand, why Turkey has been so much conscious of CSR is related with the country's increased incorporation with the international bodies, developments, events and campaigns. The United Nations Conference on Human Settlements (Habitat II), which took place in Istanbul in 1996, gave impetus to starting discussions about eradication of poverty and discrimination, promotion and protection of all human rights and fundamental freedoms for all, and provision of basic needs, such as education, nutrition and life-span health care services. When compared to governmental agencies and regulations, corporate governance structure or culture, multinational companies (MNCs) have more influence on CSR in Turkey and they positively affect CSR activities. Their branches and suppliers are forced in a positive way to run CSR campaigns, and this has become a movement in Turkish companies (kssd. org) (Turkey Corporate Social Responsibility Baseline Report, 2008, p. 43).

Consumer Attributions
Although CSR was considered to be simply support for charity by corporations in the past, recent studies attribute a strategic importance to CSR, as it allows firms to benefit financially through positive publicity and goodwill and by creating societal value (Lee et al., 2014). Following this mind change, CSR activities are increasing in profit oriented industries, and it becomes vital to know whether consumers perceive these activities as profit oriented or value oriented. One way of considering how consumers view a company's CSR activities is through attribution theory. Attribution theory proposes that individuals aim to predict and control what occurs around them (Heider, 1958cited in Marin et al., 2016. As Gilbert and Malone (1995) said, "…people care less about what others do than about why they do it" (p. 21). Based on attribution theory ( Jones & Davis, 1965;Kelley, 1972cited in Groza et al., 2011 in the context of CSR, a consumer's favourable attitude towards a social program is likely dependent on certain attributions he/she makes about the reasons behind an organization's actions (Du et al., 2007;Ellen et al., 2006;Klein & Dawar, 2004;Menon & Kahn, 2003;Sen et al., 2006;Webb & Mohr, 1998). An attribution (or corporate association) is what an individual knows or feels about a particular organization (Brown & Dacin, 1997). These associations "… serve as the 'reality' of the organization for an individual" (Brown et al., 2006, p.105), and what individuals accept as true can in the end affect various organizational responses (Walker et al., 2010, p. 663).
How to classify consumers based on their attributions is still ambiguous. One of the studies was conducted by Swanson (1995), who suggested a three-dimensional model for CSR. The dimensions were a utilitarian perspective (CSR is taken as a strategic approach to improve the bottom line), a negative duty perspective (CSR is taken to conform to stakeholders' values), and a positive duty perspective (CSR is undertaken as a right thing to do). Later, Maignan and Ralston (2002) proposed another three dimensional approach (values-driven, performance-driven, and stakeholder-driven). A recent study by Ellen et al. (2006) showed that consumer responses to CSR actions are more complex than traditionally viewed. According to the researchers, consumers attribute multiple and specific corporate motives, which are basically profit-centered (egoistic and strategy driven) or benevolence-centered (values and stakeholder driven) attributions. When consumers think that the company's CSR activities are profit-centered, it means firms are performing these actions for their own well-being rather than for society. For example, egoistic driven motivations mean that the first aim of firms engaging in CSR activities is to use the cause for their own benefit instead of helping society. In a similar vein, strategy-driven motives are self-centered attributions implying that firms are supporting CSR events as they want to achieve some business goals (e.g., increase market share, increase brand awareness) and use these social causes as tools to accomplish their goals. On the other hand, benevolence-centered motivations are related to helping society's needs first and not using those social causes for the company's own interest. In stakeholder motivation, consumers believe that firms perform CSR actions because they want to please their stakeholders. Value-driven motives are related to benevolent intentions. In this situation, consumers believe that firms care about society and behave in a moral way.
In this study both types of motivation, profit-motivated giving and benevolence-motivated giving, were investigated. Previous studies show that both types of attributions are likely to influence, directly or indirectly, consumers' behavioural intentions and attitudes towards firms. For instance, purchase intention (Becker-Olsen et al., 2006;Ellen et al., 2006;Groza et al., 2011), repeat patronage (Vlachos et al., 2009), consumers' brand and product evaluations (Klein & Dawar, 2004) and recommendation intentions (Ellen et al., 2006;Vlachos et al., 2009;Walker et al., 2010) were the outcomes that were affected by consumers' perceived motivations. Considering the importance of these attributions to consumer behaviour, the objective of the present research is to identify whether the attributed motives a consumer assigns to CSR affect consumers' support to socially responsible actions. Maignan (2001) stated that the French and German cultures are communitarian, so consumers in those two countries are expected to take into account society's well-being in their shopping decisions. Similar situation is expected in this study, as according to Hofstede (2012), Turkey is high on collectivism and it is likely that Turkish consumers may consider corporate social responsibility as an important purchasing criterion. They may even be ready to make specific efforts -for instance, paying more to buy products from a socially responsible company, considering the ethical reputation of businesses when preferring companies, avoiding to buy products from companies that have engaged in immoral actions, buying from the firm that has an ethical and socially responsible reputation even though the price and quality of the two products is the same.
As mentioned above, when consumers attribute egoistic motivations to firms, they view them as unethical and believe that firms are trying to gain profit, increase their market share but are actually doing nothing to benefit society; also Rawwas et al. (2005) have found that Turkish consumers tend to be sensitive to unethical practices, so it is hypothesized that: Hy pothesis 1: Egoistic-driven attributions have a negative effect on consumers' supportive behaviour. According to Ellen et al. (2006), strategy-driven attributions would positively affect purchase intentions. The reason for this prediction was that consumers largely accept CSR initiatives that are enacted to support the strategic goals of the firm. (Groza et al., 2011). Also, as mentioned above, Turkish consumers primarily expect economic performance (jobs) from the companies (Ararat, 2008, p. 275). In this study, the same assumption was taken into account and it is hypothesized that: Hy pothesis 2: Strategy-driven attributions have a positive effect on consumers' supportive behaviour. Stakeholder-driven motives relate to the support of social causes due to stakeholder pressures. Vlachos et al. (2009) mentioned in their study that "…consumers are likely to view attributions of stakeholder-driven motives negatively, as they believe the company is acting to avoid retribution from stakeholders and fear that a company's worthy programs may disappear in the next downturn (Franklin, 2008)" (Vlachos et al., 2009, p. 172). In this study, contrary to previous approaches, stakeholder-driven attributions are thought to be positively perceived by GSM consumers as firms have a responsibility to their shareholders to make a profit (Carrol, 1999). So it is hypothesized that: Hy pothesis 3: Stakeholder-driven attributions have a positive effect on consumers' supportive behaviour. When consumers think that firms care about society and they are not exploiting the cause for their own gain, they attribute a firm's CSR actions as value-driven and support the firm. So it is hypothesized that: Hy pothesis 4: Value-driven attributions have a positive effect on consumers' supportive behaviour.

Service Quality Perceptions
According to Keaveney (1995), service quality is one of the most important standards consumers use to assess, select and change service providers. Although having a high level of quality perception is critical for service providers, high quality itself is not the only driver of consumer behaviour. In highly competitive markets, firms are competing on many cognitive aspects like quality, reliability, durability, etc., and affective aspects like emotions, experiences, image, etc. Besides, social norms and situational factors are also making an impact on consumers' reactions (Dick & Basu, 1994). So the firms must take into account the interaction of all these factors when determining their strategy. For instance, firms should know whether CSR activities (social norm) pay off low quality (cognitive aspect) or whether consumers prefer firms that are engaging in CSR activities with egoistic motivations (situational factors) because of lower price (cognitive aspect) without taking into account quality of the service. This study treated perceived quality as a moderator variable that changes the impact of CSR attributions on supportive behaviour referring to the studies of Vlachos et al. (2009) and He and Le (2011). In their study, Vlachos et al. (2009, p. 173) stated that "…service quality provision is more personally relevant than CSR, since the former relates to satisfaction of lower-order needs (physiological needs), whereas the latter relates to satisfaction of higher order needs (self-enhancement needs). In times of economic uncertainty and mistrust, consumers may give priority to satisfying needs placed lower in the needs hierarchy pyramid (Herzberg 1966 cited in Vlachos et al., 2009, p.173)". In addition, Luo and Bhattacharya (2006), Barone et al. (2000) and Berens et al. (2007) indicated that consumers are likely to close the eyes to poor CSR attributions and may carry on to support firms due to their high service quality. Based on these statements, it is logical to expect that even though people think firms are involved in CSR activities for their own benefit and for economic purposes rather than society's well-being, if the firm has a high level of perceived quality, they will support these companies. Additionally, negative behaviour is likely to occur due to the effect of low service quality. In contrast, in the high service quality situation, consumers are more likely to support a firm with value-driven, strategy-driven and stakeholder-driven motivations, but consumers are less likely to support firms with low levels of service quality even though they engage in CSR actions with value-driven, strategic and stakeholder motivations. In other words, when the perceived service quality is low, value-driven, strategic and stakeholder-driven motivations underlying CSR cannot compensate for poor service quality, this results in no support or weak support for firms engaging in CSR actions. So it is hypothesized that: Hy pothesis 5: Perceived service quality moderates the relationship between egoistic -driven and consumer support behaviour. The negative relationship will be weakened or null for high perceived service quality, and the negative relationship will be strengthened for low perceived service quality. Hy pothesis 6: Perceived service quality moderates the relationship between value-driven, strategic and stakeholder driven motivations and consumer support behaviour. The positive relationship will be strengthened for high perceived service quality, but will be weakened or null for low perceived service quality.

Research Design
This study explored consumers' attributions related to CSR efforts of GSM (Global System for Mobile Communications) operators since the mobile service providers are relatively well known for their CSR activities in Turkey. Since 2005, one of the business magazines, Capital, has conducted an annual the "leaders of corporate social responsibility" research in cooperation with GFK Turkey. The aim of this research is to find out the most successful companies and projects in terms of CSR performance in the eyes of public and white-collar managers. In 2014, 305 white-collar managers and 1238 participants from 15 cities in Turkey answered the questionnaire. The results of this study showed that Turkcell, the leading GSM operator, was also evaluated as the leader in CSR performance in the eyes of the public and ranked second in the eyes of business managers. Vodafone, the second biggest GSM operator in Turkey, was evaluated as the ninth successful company in the eyes of business managers, and Avea, the third biggest GSM operator in Turkey, was evaluated as the seventh successful company in the eyes of the public. In addition, the most successful ten CSR projects, according to the public, include three projects executed by Turkcell called respectively "Kardelenler", "Gönül Köprüsü" and "Engel Tanımayanlar". In this study, three leading GSM operators were chosen on the basis of their marketshare (51.33%, 28.45%, 20.22%), and respondents answered questions about the GSM operator that they most frequently used. The primary data were collected through a personal survey, in accordance with the theory of stated preferences.
This study was applied to customers of mobile communications services, since the mobile service providers are relatively well known companies investing in CSR activities in Turkey. Besides, the GSM market has a really high penetration rate among the population (the population in Turkey aged 15 and above was 58.8 million in 2013, according to the address based registration system of Turkish Statistical Institution (TÜİK). At the end of the third quarter of 2015, the number of mobile telephone subscribers in Turkey had reached 73.2 million (http://www.ttinvestorrelations.com/ turk-telekom-group/investing-in-turk-telekom/turkey-telecom-sector.aspx). The market is growing enormously, and (as mentioned above) there are three providers of GSM services (Turkcell, Vodafone and Avea). Accordingly, the market can be considered highly competitive and with a typical oligopoly structure. In that kind of competitive environment firms need to differentiate their offering in terms of augmented character-istics rather than the basic service offerings. CSR activities can be thought of as a tool that helps enrich the offered services and, in Turkey, all of the mobile communication service providers support many causes that affect society.

Data Collection and Characteristics of the Sample
In order to collect data to test the hypotheses, a face-to-face survey was carried out by adopting a quota sampling technique based on the market share of each mobile service providers. The sample was drawn from the Istanbul metropolitan Area in between the 16th and 28th of February, 2016 through convenience sampling. The questionnaire began with an introductory statement that asked respondents to administer their own responses, and assured them of confidentiality. This was followed by a filter question of whether or not they are a registered customer of one of the GSM operators; also, they had to indicate the brand name of their most frequently used GSM operator. After that, the respondents were asked the third filter question of whether they know the CSR campaigns of the service provider they are mostly using. If the answer was 'no' , they did not answer the following questions. In other words, the survey continued with participants who are a registered customer of one of the GSM operators and aware of the specific campaigns of their mostly used GSM services provider. The remaining questions of the survey were answered based on respondents' most frequently used GSM operator.
Following the three filtering questions, respondents indicated the length of time they had been using the services of the GSM operator they mentioned. This introductory question was followed by main questions that measured CSR attributions, five items to measure supportive behavior and six items to measure perceived quality. The Likert-type scale questions were not randomized on the survey form but asked in a randomized order during the face-to-face questioning in order to avoid biased responses and increase validity. At the end of the survey form, questions related to demographics were answered.
In total, 600 mobile phone subscribers were asked to participate in the survey, but 200 of them refused to answer the questionnaire form. As a result, a total of 400 questionnaires were obtained, with about 33.3% rejection rate of the targeted sample size. Frequencies related to the number of users from each GSM operator show that 49.3% of the respondents are customers of Turkcell, 31.7% of the respondents are customers of Vodafone and 19.0% of the respondents are customers of Avea. Among these respondents, 36.3% have been using the services of that GSM operator for more than five years, 29.7% have been using the services for 3-4 years and 34.0% have been using it for less than 3 years.
Sample characteristics also appear to be representative of adult mobile phone users in Turkey. A total of 52% of the sample is female. In terms of age, 52.7% were 18-25 years old, 19.0% 26-33 years old and 28.3% were over 34 years old. Occupation frequencies show that 38.8% of the sample were students, 28.3% were working in the private sector and 9.0% were working in the public sector. The remaining 11.6% were employed as tradesmen or workmen. Participants running their own business and in other professions constitute 12.3% of the sample. With regard to education, a total of 53% had a high school degree or less, while 47% had at least a university degree.

Common method bias
As mentioned above, in this study, self-report questionnaires were used to collect data answering the Likert-type scaled questions like CSR attributions, supportive behaviour and perceived quality, at the same time from the same respondent. According to Podsakoff and Organ (1986), this situation causes common method variance (CMV) problem which is defined as "amount of spurious covariance shared among variables because of the common method used in collecting data" (Buckley et al., 1990cited in Malhotra et al., 2006, p.1865. Evidence of CMV is likely to affect the internal consistency of the scale since the correlation among variables will be high due to the fact that the same person answers the questions at the same time in a single questionnaire (Chang et al., 2010). For this reason, in order to reduce CMV in the study, during the questionnaire construction, the order of the questions in the survey was randomized as recommended by Podsakoff and Organ (1986). Afterwards, ensuing the data collection, Harman's single-factor test was used as a statistical procedure to control CMV. The results of the unrotated factor showed that no single factor had emerged and also none of the factors had majority of the variance, showing that CMV is not a problem. In the exploratory factor analyses, six factors with eigen values greater than one came out, the larger variance accounted by "one general factor" was 24.5% of the variance in the data. Therefore, this study has no CMV problem.

Research Framework and Measures
The instruments used to measure the constructs involved in this study were adopted from existing literature. All scaled items were measured on five-point Likert-type scales with 1 = strongly disagree and 5 = strongly agree. CSR attributions measures were drawn from Ellen et al. (2006), perceived quality measure from Washburn and Plank, (2002), Aaker (1991Aaker ( , 1996; Yoo et al. (2000) and adopted according to service sector and finally, consumer support to socially responsible actions measure from Maignan (2001) (Figure No. 1).

Analysis and Findings
Research data were analyzed with structural equation modelling using LISREL 8.8. Descriptive statistics related to demographics and usage related issues were calculated using SPSS 17.0. Reliability of the constructs employed in this study was evaluated with Cronbach's alpha, and the validity of each construct was tested with CFA (confirmato-ry factor analysis). In order to test the hypothesized direct effects, structural equation modelling was used in line with Anderson and Gerbing's (1988) two-step approach, and in order to test the moderating effect of perceived quality, multi-group analysis was conducted.

Measurement Model
Measurement model was tested by using CFA in order to assess discriminant and convergent validity as well as construct reliability. Based on the CFA results, the measurement model was revised prior to estimating the structural model. All the variables in the measurement model, except one variable to measure the stakeholder driven motive, have significant factor loadings. Because the t value of the variable measuring stakeholder driven motive was insignificant, it ("They feel their employees expect it") was deleted from the measurement model and then CFA was run again for the revised model. Final measurement variables, their standardized regression weights and significant loadings are shown in Table 1. Fit indices of a 23 item and six-dimensional confirmatory factor model were also above or near acceptable thresholds: χ2 = 735.76; df = 213; p = .000; GFI = .97; AGFI = .96; CFI = .96; IFI = .96; RMSEA = .078 (90% confidence interval for RMSEA = 0.072 -0.085) and SRMR= .014.
The construct validity of the measurement model was estimated by convergent and discriminant validity. Three measures commonly used to test convergent validity are standardized factor loadings, average variance extracted (AVE) and composite reliability (CR). Each item in the measurement model has standardized factor loading higher than the recommended 0.40 cutoff (Nunnally & Bernstein, 1994), an AVE for each dimension is higher than 0.5, and CR was above 0.70 across the constructs, exceeding the recommended threshold value (Hair et al., 2010).
Discriminant validity was tested in line with Fornell and Larcker's (1981) suggestion. According to this test, discriminant validity was achieved if the average variance extracted by a latent variable is greater than the shared variance (i.e., the squared correlation) with other latent variables. As Table 2 shows, this criterion was met by all the variables in the study, as no squared correlation exceeds the average variance extracted.

Results for Direct Effects Hypotheses
Hypotheses H 1 to H 4 were tested with structural equation modelling. The standardized parameter estimates for the direct effects model are shown in Table 3.

Results for Moderated Effects Hypotheses
Moderated relationships implied by H 5 and H 6 for perceived quality (PQ) were tested through multiple group confirmatory factor analysis (MGCFA). Two subgroups, one with low PQ and one with high PQ, were created based on perceived quality evaluations of respondents via the median split.
Before analyzing the model in each subgroup, the measurement model was tested for the measurement invariance. In this study, measurement invariance is restricted to metric invariance (i.e. invariance of factor loadings). Metric invariance means that participants in different groups understand and react to measures in an identical manner (Meredith, 1993). The main hypothesis of measurement invariance is that low PQ and high PQ groups will hold the same measurement models. In searching for the metric invariance, a baseline model which is a constrained model (Model A) was initially tested and then the unconstrained model (Model B) was analyzed. The chi-square value of 677.09 with 262 degrees of freedom for the baseline model (Model A) that specified invariance for the number of factors and factor loadings of items defining each factor was compared against the chi-square value of 640.15 with 245 degrees of freedom for Model B. The chi-square test between Model A and Model B was significant, X 2 (∆df= 17) = 36. 94. However, given that there was no substantial difference in CFI (.89 vs. .90) and GFI (.94 vs. .95), it was concluded that there is no acceptable difference between the two groups on the metric invariance.
After support for the metric invariance had been confirmed, as a next step, an overall Chi-square difference was computed for the moderator variable. Theoretically, a model with equality constraints was compared with a model that allows the parameters to vary. This test imposes the null hypothesis that the moderator variable does not have any effect on the parameters. As can be seen from Table 4, perceived service quality does not moderate the hypothesized negative effect of egoistic-driven motive on supportive behavior. This result confirms H 5 . The multigroup analysis for perceived quality shows only two moderating effects which are for strategic motives and value driven motives. However, the coefficients for the paths between strategic motives and supportive behaviour were statistically insignificant in both high perceived quality group and low service quality groups. The coefficients for the paths between value-driven motives and supportive behaviour were statistically significant in both high perceived quality group and low service quality groups, but the strength of the association is lower in high PQ group. These results partially support H 6 .

Discussion
This study provides a deeper exploration about whether consumers' attributions regarding firms' CSR activities influence their support towards the firm, in which perceived service quality is viewed as an important process affecting the impact of CSR attributions on supportive behaviour.
CSR has received increasing concern in recent years among both academics and in practice. A growing number of organizations are planning and involving in CSR practices due to the pressure from various stakeholders, such as customers, employees, community groups, non-governmental organizations and government (McWilliams et al., 2006).
Integrating corporate social responsibility (CSR) initiatives into business is challenging but beneficial for firms. Literature reveals that the more companies engage in CSR activities, the more positive attitudes stakeholders have toward that company and these positive attitudes positively influence their purchase intention towards the products of these firms (Brown & Dacin, 1997;Creyer, 1997;Ellen et al., 2006;Lichtenstein et al., 2004;Luo & Bhattacharya, 2006;Sen & Bhattacharya, 2001), loyalty (Du et al., 2007) and recommendation intentions (Vlachos et al., 2009).
Most of the CSR studies have investigated the strategic role of CSR in organizations, importance of CSR awareness, the antecedents and consequences of CSR and the effects of CSR on stakeholders (such as consumers and employees) of the organization (e.g. Sen et al., 2006). While considerable amount of research has been done related to how consumers perceive firms' CSR activities and CSR's impact on consumer attitudes and behaviour, several questions still remain unanswered, i.e., "how consumers perceive the underlying motives of firms when performing CSR activities" and "what is the role of attributions of motive in explaining consumers' support towards these firms?" This study contributes and extends the understanding on how consumers' perceptions of corporate motive on CSR efforts influence their support towards the firm on the basis of perceived service quality. Thus, it helps elaborate how the path between CSR efforts with different motives and support may change among individuals with high and low service quality perceptions.

Summary of Findings
The findings of this study are as follows. The results of the study indicated that value driven, strategy driven and stakeholder driven motives influence support positively, whereas egoistic driven motives were found to influence support negatively, as expected. Among the most interesting findings was the fact that strategic and stakeholder driven motives had a greater effect (γ= .38, γ= .52) than value driven motives (γ= .16). In many studies these two motives (strategic and stakeholder driven) were hypothesised as negative attributions (Geue & Plewa, 2010;Walker et al., 2010;Marin et al., 2016). This implies that informing consumers of how CSR initiatives support the strategic goals of the firm (strategy driven motive) and of firms' duties towards their stakeholders to make a profit (stakeholder driven motive) has a strong impact on their supportive behaviour. It is good for companies to act in a socially responsible manner when consumers perceive their efforts as providing value to society. On the other hand, when consumers perceive the underlying motivation as self-centred, they question the CSR activity and whether or not it benefits the firm or the society, and as a result they do not support the firm.
Following the investigation of direct effects, moderating relationships were tested based on the service quality perceptions of respondents. This study interestingly showed that service quality does not moderate the negative relationship between egoistic driven motive and supportive behavior, so Hypothesis 5 is rejected. Moreover, it was found that perceived quality moderates the relationship in strategy driven motives and value driven motives to support the behaviour path. For consumers with both high quality and low quality perception, value driven motives have a significant and positive effect. Although service quality moderates the relationship between strategic motives and supportive behaviour, the relationships in both high and low quality groups were negative and insignificant. Results also showed that service quality does not moderate the relationship between stakeholder motives and supportive behaviour, so Hypothesis 6 is partially accepted.
The main reason for these results may be the difficulty of differentiating the services offered in the GSM market. The brands in this market always follow the similar campaigns. New regulations also make it really easy to transport the GSM number among the operators without changing the phone number, and the three GSM service providers always encourage number transportation by offering advantageous campaigns. As a result, consumers do not necessarily agree to pay more if companies act with strategic motives in their CSR activities rather than value driven motives. In addition, another explanation to this result may also be related with difficulties in recognizing the quality differences among the three rival GSM operators, so the consumers may not want to be viewed as foolish by others. These results may change according to the industry or product category being investigated. Also, the other reason could be related with the support scale utilized in the study as it measures consumers' buying tendency from socially responsible companies, not future buying intentions. In future studies this point should also be tested by using a purchase intention scale.

Theoretical contribution
This paper makes a significant academic contribution. The study broadens the theoretical research related to consumer attributions. First, this study validates the measurements for consumer attributions and consumers' support for corporate social responsibility for GSM operators in Turkey. This study shows that the dimensions of consumer attributions in Turkey are similar to those in other countries. Second, the study provides empirical evidence suggesting that consumer attributions have influences on consumers' supportive behaviour, and these influences are moderated by service quality perceptions. Third, the results of the existing studies related to CSR are generally based on products; the current study examines the role of perceived service quality in affecting the impact of perceived CSR motives on consumers' support in the context of mobile communication service industry so as to fill the gap in the literature. Finally, by getting data from Turkey, this study advances existing theory related to consumer attributions and outcomes in the context of a non-western country. This is a significant contribution since previous research mostly focuses on the USA, Europe and other major industrialized markets. By looking at Turkey, this study not only provides further efficacy to attribution theory, but also it has been done in one of the most important markets in the world today.

Managerial implications
In addition to its theoretical contribution, this study may be of value to managers who may draw important conclusions from the results of the study. They need to understand how their consumers attribute the CSR activities that firms conduct,and as Marin et al. (2016: 256) mentioned in their study, "to secure consumers' trust; marketing communications could provide details about how a firm's CSR programs have helped the company to solve social (and local) problems by emphasizing results and sustainability rather than merely introducing the form and input of their CSR activities". To sum up, it is not enough for companies just to perform CSR activities and wait for the customers to appreciate them, they need to ensure that their customers believe them and understand the real reason behind the firm's CSR activities. They must understand which attribution has a significant and unique influence on supportive behaviour and then may take specific actions. The study showed that when consumers perceive CSR to be driven by stakeholder motives, its effect on their supportive behaviour was the highest one. If GSM firms know the positive attributions which are specific to them, they are more likely to run their positioning campaigns towards their stakeholders. Furthermore, as mentioned before, even though quality is considered one of the most vital determinants to consumers for choosing service providers (Caruana, 2002), it may not guarantee retention and purchase intention. The findings of the study confirmed this thought. As Dick and Basu (1994) mentioned, social norms may also have impact on consumers' reactions. For this reason, considering the competitive nature of the Turkish GSM industry, the findings of this study may contribute to practitioners by highlighting the importance of designing and implementing CSR activities at strategic level along with branding and positioning. Understanding the role of attribution in affecting customers' support towards the firms involved in CSR actions may help managers to monitor and enhance consumer outcomes by managing CSR-based attribution processes. Therefore, the effectiveness of CSR can be improved and, consequently, exhibit more positive outcomes toward that firm.

Limitations and Further Research
There are several limitations of this study, which present opportunities for future research. This study tested a model with mobile communication service customers based on a convenience sample from Turkey. Future research should test this model with larger random samples or samples in other contexts, for instance, in different service businesses or product categories and cultures. The present study tested the moderation effect of perceived quality, future studies should take into account brand image, corporate reputation, product involvement, ethical sensitivity of consumers as moderators. Additionally, besides moderation effects, the mediating role of consumer trust, corporate reputation and company identification should also be investigated. This study investigated only the effect of CSR motives on customers' supportive behaviour. The effects of other CSR features, such as CSR target, CSR fitness, CSR media channel, should also be tested. Furthermore, the dependent variable in this study should also be extended in future research. Satisfaction, loyalty and future patronage intentions might be incorporated.