Effects of Corporate Social Responsibility perception on consumer satisfaction with the brand

Abstract

Using two different studies we assess under what conditions Corporate Social Responsibility (CSR) affects consumer satisfaction. In the first study, based on the American Customer Satisfaction Index, and a content analysis of CSR initiatives reported by 65 US companies, findings show a positive direct relationship between CSR and customer satisfaction with CSR training and environmental initiatives, but a negative direct relationship between CSR corporate communication initiatives and customer satisfaction. The second study, through an empirical study based on an online survey of consumers, points to a positive relationship between CSR and customer satisfaction. Our findings suggest a direct and an indirect route for this relationship through brand attitudes. These apparently divergent, but complementary results, between the two studies are explained by the different conceptual approaches that underline the CSR communication processes and the relationship between consumers and brand. Our findings also suggest that public policies on CSR visibility must be reviewed.

Resumen

Utilizando dos estudios diferentes, evaluamos bajo qué condiciones afecta la Responsabilidad Social Corporativa (RSC) a la satisfacción del consumidor. En el primer estudio, basado en el Índice de Satisfacción del Cliente Americano (American Customer Satisfaction Index), y un análisis del contenido de las iniciativas de RSC reportado por sesenta y cinco empresas americanas, los hallazgos reflejan una relación positiva directa entre la RSC y la satisfacción del cliente, y la formación en RSC y las iniciativas ambientales, y una relación negativa directa entre las iniciativas de comunicación corporativa de la RSC y la satisfacción del cliente. El segundo estudio, a través de un análisis empírico basado en una encuesta online a consumidores, apunta hacia una relación positiva entre la RSC y la satisfacción del cliente. Nuestros hallazgos sugieren una vía directa y una indirecta para esta relación, a través de las actitudes de la marca. Estos resultados aparentemente divergentes aunque complementarios, entre los dos estudios, se explican mediante los diferentes enfoques conceptuales que recalcan los procesos de comunicación de la RSC y la relación entre los consumidores y la marca. Nuestros hallazgos sugieren también que deben revisarse las políticas públicas sobre RSC.

Introduction

Corporate Social Responsibility (CSR) is nowadays part of the recent agenda for (i) public policy makers, like the European Commission (EU COM, 2011), (ii) marketers through CSR indexes and social campaigns worldwide (e.g., Johnson and Johnson or Ikea) and (iii) academics with an increasing number of papers and topics. Despite criticism (Karnani, 2011) CSR is becoming part of the social agenda in magazines and newspapers (Epstein-Reeves, 2012).

Among the different stakeholders addressed by CSR, consumers stand for the large economic impact and diversity of assessing effects. Prior literature on CSR and consumers tested the effect of CSR on corporate reputation (Brammer & Millington, 2005; Lichtenstein, Drumwright, & Braig, 2004), emphasizing that one key element in any CSR research should be how corporations report the CSR and consumer perception. Otherwise no matter what kind of CSR practice is implemented by companies or how CSR initiatives engage in, the impact of CSR on consumer behavior may be null or even negative (Du, Bhattacharya, & Sen, 2010). Satisfaction is considered a relevant construct for evaluating customer appraisal. Recent marketing literature is increasingly devoting attention to analysis of customer satisfaction and exogenous variables of company performance such as stock prices (Fornell, Mithas, Forrest, Morgeson, & Krishnan, 2006), employee satisfaction and retention (Frey, Bayón, & Totzek, 2013), and consumer spending growth (Fornell, Rust, & Dekimpe, 2010).

Satisfaction with CSR initiatives therefore becomes a key metric of the successful development and implementation of a CSR policy. Consumers may wish to reward companies with good CSR records positively or even with non-negative information in the form of resilience to negative information about the company (Bhattacharya & Sen, 2004). Recent literature has not only delineated how to achieve this relationship between CSR initiatives and positive attitudes, but also how to offer to business practitioners a comprehensive framework for identifying appropriate responses to consumer demands, based on three different motivations: instrumental, relational and moral (Caruana & Chatzidakis, 2014). However, the relationship between CSR initiatives and customer satisfaction has been scarcely researched and consequently, our understanding of the disaggregated drivers and causes may be incomplete. We also argue that positive attitudes may be expected to trigger satisfaction and hence CSR must have a positive influence on satisfaction.

Thus, our aim in this research is twofold: first, to assess what kinds of CSR initiatives significantly influence consumer satisfaction with the company; second, to assess the mediating role exerted by brand attitude in the relationship between CSR and satisfaction. These two main research objectives are accomplished through two different studies in two different countries, USA and Spain that reflect two cultural settings and also depict different cultural patterns in each of the five current six cultural dimensions suggested by Hofstede and Hofstede (2001).

The first one aims to examine the relationships between a set of different disaggregated CSR initiatives from 65 US firms analyzed through content analysis with consumer satisfaction. Consumer satisfaction data are gathered from the American Customer Satisfaction Index (i.e., ACSI). The second study addresses the direct and indirect relationship between perceived CSR and satisfaction using a different methodological approach based on primary data from 351 Spanish consumers.

With this research we contribute to the existing literature in the following ways: First, by extending the conceptual approach to the effect of disaggregated CSR initiatives on customer satisfaction and providing evidence to support this relationship. In doing this, we reinforce the role of satisfaction as the core metric for managers, as pointed out above. Second, by identifying the mechanisms of the effect of CSR on customer satisfaction with the brand, we show two complementary routes.

The rest of the manuscript is organized as follows. First, the relationship between CSR and customer satisfaction is examined, deriving some research questions. Then the two studies in two different countries are presented and the research questions and consistent hypotheses are tested. The manuscript ends with some discussion and implications for firms and public policies and outlines further research.

CSR and consumer satisfaction

Following Beckmann (2007), it can be said that CSR has a wide variety of effects on consumers that can only be tested in a diffuse rather than a compact way. In fact, consumers react to some CSR initiatives but not others due to their level of knowledge, congruence fit between consumer profile and company products, and brand characteristics (Bigné, Currás, Ruiz, & Sanz, 2012); their reaction also varies in relation to the relationships between corporate skills and CSR initiatives (Du, Bhattacharya, & Sen, 2007; Lichtenstein et al., 2004); and perceived trustworthiness of the information source (Mohr, Webb, & Harris, 2001; Robinson, Irmark, & Jayachandran, 2012). In addition to these consumer-related effects, cultural context, technological, economic, political and social factors have been shown to influence CSR assessments (Dahlsrud, 2008). Although the literature is abundant in terms of CSR and its effects, there are still gaps in the understanding of other variables that may assess the performance of a brand in adopting CSR initiatives. Brand loyalty has been suggested as a recurring theme in the CSR literature showing that higher levels of CSR associations are linked to stronger loyalty behavior because consumers develop a more positive, stronger evaluation of the firm (Brown & Dacin, 1997; Sen & Bhattacharya, 2001). Recent studies have focused on novel effects of CSR, like brand equity (Torres, Bijmolt, Tribó, & Verhoef, 2012) or behavioral effect, such as repeat patronage intentions (Auger, Devinney, Louviere, & Burke, 2010; Vlachos, Tsamakos, Vrechopoulos, & Avramidis, 2009).

In this paper we attempt to argue the role of satisfaction as a key metric in evaluating CSR performance. In the marketing domain, satisfaction has been used typically as a measure of overall performance of a brand. Furthermore, it has proven its relevance for non-marketing effects, such as stock prices (Anderson, Fornell, & Mazvancheryl, 2004; Fornell et al., 2006), employee satisfaction (Frey et al., 2013) and consumer spending growth (Fornell et al., 2010). Given the broader nature of CSR for firms, satisfaction might explain the adequacy in monitoring company performance. Satisfaction is considered to be a judgment that is cognitive-affective and relative, that is, the result of comparison between a subjective experience and a previous standard Oliver (1997, p.13) that may be used to evaluate companies’ CSR initiatives. Customer satisfaction, based on customer relations with the brand and its CSR initiatives can be approached from the perspective of overall or specific issues. In this study we examine overall issues because specific issues tend to be related to other constructs such as cause-brand fit (Bigné et al., 2012), brand credibility (Bigné, Chumpitaz, & Currás, 2010) or altruistic attributions (Ellen, Webb, & Mohr, 2006) deriving in context related findings. Although those factors play a role, they might be considered mediators in a specific initiative rather than overall and superior explanatory issues.

Our first goal therefore is to assess what kinds of CSR initiatives developed by companies have more impact on consumer satisfaction. It might argue that the different nature of CSR initiatives undertaken by a company will have a specific influence on the CSR consumer response mechanism (Carroll, 1991; Robinson et al., 2012). Some studies have analyzed the differences in the social commitment of the company perceived by consumers depending on the type of CSR initiative used (Dean, 2003; File & Prince, 1998; Polonsky & Speed, 2001). It is reasonable to expect that these differences are also visible in the influence of CSR on consumer satisfaction. Assuming the different scope of CSR initiatives, we propose a first research question that seeks to highlight the different influence of CSR initiatives:RQ1

What kinds of CSR initiatives significantly influence consumer satisfaction?

Through CSR, firms try to show a brand personality characterized, to some extent, by altruistic values. Therefore, consumers begin a process of cognitive elaboration albeit in a very simple fashion, with one main goal: to acquire guarantees in relation to the firm's good faith in its social commitment, guarantees that the way the firm is presenting itself though its CSR program is consistent with the firm's real corporate values (Forehand & Grier, 2003). This cognitive process is based on a series of judgments about the organization's credibility, its reputation or congruence between the CRS programs, the firm's main activity and its brand positioning which have been pointed out in CSR literature. Furthermore, consumers may be generally satisfied with CSR activity in the following ways: (i) interacting with the brand that carries out CSR initiatives is a way to satisfy consumer interest in participating in the social welfare of the community (Bhattacharya & Sen, 2004; Bhattacharya, Korschun, & Sen, 2009), unless they perceive opportunism (Bigné et al., 2010) or similar unfair issues; (ii) consumers reward CSR initiatives as a trigger for future initiatives in the same directions, since companies will read customer satisfaction as a path to develop more CSR initiatives or increase current ones. Following the above reasoning, we argue that there are two mechanisms for obtaining consumer satisfaction. The first route is a direct one consisting of a cognitive process triggered by perception of the CSR activity (García de los Salmones, Herrero, & Rodríguez del Bosque, 2005). Essentially this means that consumers must be made aware and convinced of CSR initiatives and sincerity through communication tools (e.g., social reporting). The second is an indirect route through attitudes (He & Li, 2010), where social consumer identification with the company engaging in CSR initiatives creates an identification that must have a direct, positive influence on consumer attitudes toward the brand (Currás, Bigné, & Alvarado, 2009). And following Du et al. (2007) it can be suggested that consumers tend to have more positive perceptions of CSR and “reward” CSR actions in terms of attitude. Reinforcing the indirect route through attitudes, it seems that satisfaction requires a previous acknowledgment effort after which, once consumers are aware of CSR activity, they encode positively. When any stimuli affect consumers, for instance shopping, evaluative processes, or repeated exposure to the CSR activity, consumers trigger an overall brand evaluation process that determines satisfaction, unless a negative disconfirmation process creates dissatisfaction. In sum, our second aim is to assess how customer satisfaction is reinforced if a company adopts a CSR policy and whether this effect can be direct or indirect through attitudes. Therefore we propose a second research question:RQ2

Is the effect of customer CSR associations on consumer satisfaction a direct effect and/or an indirect effect, through brand attitude?

Empirical research

Study one

Background

The main aim of the first study is to analyze the relationship between CSR reported initiatives and customer satisfaction; and in particular, attempt to identify which CSR initiatives have more impact on customer satisfaction. In pursuing this goal, we might support decision making in firms’ CSR policies. First, CSR is becoming a global issue. Second, data on satisfaction are available from an established study and on a continuous basis at company level, namely from the American Customer Satisfaction Index, ACSI (2016).

Methodology

This first study gathers data from two independent datasets: (i) American companies who formally declare their CSR policy in their annual reports on sustainability or CSR and devote space on their websites dedicated exclusively to their CSR initiatives. A list of the latest available data on the CSR initiatives from 2010 to 2011 of 65 American companies from different industries were compiled for this study; (ii) consumer satisfaction index borrowed from ACSI. This dataset is the most consolidated and cited customer satisfaction index in the marketing literature (Fornell, Johnson, Anderson, Cha, & Bryant, 1996) and shows growing interest in non-marketing effects. The ACSI sample of companies is representative of the USA market and main sectors, as supported by ACSI methodology.¹

The firms were chosen on the following criteria: (i) must state an explicit CSR policy on its website; (ii) provide online information on their specific disaggregated CSR activities, following Maignan and Ralston's (2002) suggestion; (iii) CSR initiatives should be developed in 2010 and 2011; (iv) be part of the ACSI dataset. Finally, 65 firms met the above criteria from the list of 230 firms that are included in the ACSI.

The information provided from their websites was coded as follows. First, all of them were coded; then a factor analysis was developed, without significant correlations between the initiatives; then the CSR initiatives were firstly sorted by the researcher according to the CSR literature in three different categories: corporate, internal and external CSR initiatives (Maignan & Ferrell, 2001). A valued data set was obtained, considering the diverse kind of CSR initiatives and the number of actions – frequency-, carried out by each of the 65 companies inside each type of communicated CSR initiatives. Finally, after the content analysis, an inventory for the 14 types of different CSR initiatives emerged, as shown in Fig. 1.

Figure 1.

Inventory of CSR actions.

In order to assess which initiatives may impact on customer satisfaction, the fourteen types of CSR initiatives were regressed using linear and non-linear ordinary least squares regression methods to the ACSI index, as follows:

Results

We performed a linear regression model, since the non-linear regression did not support more explanatory capacity. Table 1 shows the results for the best model explaining the relation between ACSI and the fourteen CSR initiatives, with a 25% ability to explain the dependent variable. We tested other quadratic multiple regressions, always with a worse fit:

Table 1.

Results of the CSR initiatives inventory vs. ACSI linear regression model and model fit.

	β coefficient	Collinearity statistics
		Tolerance	VIF
Constant	75.862^a
Train.	15.668^a	0.775	1.291
Environ.	8.774^a	0.817	1.224
Str.Com.	−5.598^a	0.843	1.186
R^2	0.250
Snedecor F	6.783^a

a

p < .01.

In relation to RQ1, it can be seen that only two types of CSR initiatives are positively associated with consumer satisfaction measured through the ACSI index, that is, (i) CSR and professional training actions for employees and (ii) environmental actions. Surprisingly, corporate communication actions relate negatively and significantly to consumer satisfaction. These findings posit two interesting issues: (i) reinforce some previous research indicating that when a firm communicates itself as socially responsible it modifies the common reference framework of maximizing profit, used to evaluate firm behavior. And although consumers tend to positively accept firms’ CSR initiatives, they might be initially skeptical of them (Ellen et al., 2006; Forehand & Grier, 2003), so brands are increasing “pro-social communications”, even though their effectiveness is uncertain (Pomering & Dolnicar, 2009), due to that skepticism; (ii) the negative relation between corporate communication of CSR initiatives and customer satisfaction is consistent with recent research by Korschun, Bhattacharya, and Swain (2014) who point out that companies can increase the impact of their CSR initiatives by increasing awareness and credibility of their CSR activities among employees and customers, and not just generally communicating CSR corporate policy. This idea is also supported by López and Smith (2014) who suggest that effective CSR communication should include detailed information of CSR initiatives and be aligned with consumer perceptions.

Study two

Background

Whereas the first study attempted to approximate the study of the relationship between CSR initiatives and customer satisfaction using secondary data on CSR published by the companies themselves and the ACSI index, Study 2 is a complementary work, based on a self-reported questionnaire on customer perception of the CSR of a specific brand and satisfaction linked to a specific consumption experience. Given that it is naïve to relate CSR perception and customer satisfaction in isolation and taking into account the limited finding from Study 1, depending on the CSR initiatives, which shows that the simple dissemination of corporate information on CSR does not ensure greater customer satisfaction, it was decided to extend the research by examining the mediating role of brand attitude in the CSR-satisfaction relationship. The model tested in Study 2 is shown in Fig. 2.

Figure 2.

Theoretical model in Study 2.

The proposed model explains how consumer perception of brand associations with the CSR initiatives, is a type of brand image capable of generating satisfaction directly and influencing it indirectly through the effect on brand attitude. Prior literature supports our hypotheses. Firstly, according to Sen and Bhattacharya (2001), CSR initiatives are not only particularly good at revealing a brand's character, but also this value-based positioning is valued by the consumer because it is perceived as central, distinctive and enduring (Albert & Whetten, 1985). Thus, CSR generates a favorable context around the brand that, through the halo effect, provokes favorable judgments on the service experience (Brown & Dacin, 1997; Luo & Bhattacharya, 2006) and a more sympathetic response to any shortfalls in service provision (Klein & Dawar, 2004), thereby improving customer satisfaction (H1). Secondly, prior studies suggest that CSR is a particular dimension of brand personality (Madrigal & Boush, 2008). Thus social responsibility initiatives can be expected to create strong associations that differentiate the brand (Hoeffler & Keller, 2002). Different studies (Bhattacharya et al., 2009; He & Li, 2010) show that consumers tend to have more positive perceptions of the CSR and “reward” companies that practice CSR initiatives in terms of increased attitude to the brand. According to Pomering and Dolnicar (2009), CSR may be effective in eliciting favorable consumer attitudes, while recent research, like the study by López and Smith (2014), underlines that knowing which CSR domains are relevant to consumers is key for the formation of consumer preferences and attitudes, and consequently for maximizing the impact of CSR initiatives (H2a). Finally, the marketing literature (Cronin, Brady, & Hult, 2000; Szymanski & Henard, 2001) suggests that improvements in brand attitude will lead to improvements in brand satisfaction. According to Keller (1993) the strength of brand associations is a function of the quantity and quality of brand information in consumers’ memories. When consumers process and actively elaborate brand significance, they consolidate stronger brand attitudes thereby increasing the likelihood that brand attitude will be activated after evaluating their consumption experiences with that brand, such as satisfaction, (H2b). Therefore we propose two main hypotheses as follows:H1

Consumer CSR associations have a positive direct impact on consumer satisfaction.

H2

Consumer CSR associations have a positive indirect impact on consumer satisfaction, mediated by brand attitude.

H2 comprising:H2a

Consumer CSR associations have a positive impact on consumer brand attitude.

H2b

Consumer brand attitude to CSR associations has a positive direct impact on consumer satisfaction.

Methodology

A quantitative empirical study based on an online structured questionnaire was implemented to test the model. In order to test the proposed model the sportswear sector was chosen, for the following reasons: (i) this sector is dominated by global brands that practice CSR with positive and negative reactions from consumers; (ii) the potential for CSR in the sector of global sportswear brands is high, mainly due to the characteristics of the industry, where there is high risk of violating basic CSR standards (Lim & Phillips, 2008; Torres et al., 2012). The Spanish market was chosen because of its economic relevance (accounting for 13.4% of the European market, World Bank, 2016), and because CSR is achieving public relevance in this country, as far as more Spanish consumers are including social and environmental criteria in their consumption decisions (Valor & Carrero, 2010).

351 valid answers from real sportswear Spanish consumers using a professional online panel were collected in June 2012, which represent a sample error of 5.2% (p = q = 0.5).

In the questionnaire, respondents were asked to identify their preferred sportswear brand from ten already selected brands chosen for their popularity. Then participants selected their preferred brand by highest number of purchases over the last two years. Later, respondents were exposed to 2 of the CSR actions developed by their preferred sportswear brand through detailed descriptive texts of the CSR initiatives developed by the brand sorted in three CSR categories following the sustainable development approach: economic, social and environmental initiatives. Those real initiatives were obtained from the sustainability and CSR reports from the ten sportswear brands. In this way, respondents did not evaluate their general CSR perceptions, but rather gave their evaluation of the CSR actions the companies really undertook, in accordance with the CSR disseminated information. Based on consumer reactions to the visualized CSR initiatives, participants completed a questionnaire that measured the relations between the CSR actions of each consumer's preferred sportswear brand and the variables of interest in the model to be tested: consumer satisfaction and brand attitude.

Brand associations with the CSR initiatives were measured with a scale of six items based on the works by Brown and Dacin (1997) and Sen and Bhattacharya (2001). Consumer satisfaction was operationalized following the four-item scale from Cronin et al. (2000). Finally, brand attitude was measured on a five-item scale based on Dabholkar and Bagozzi (2002). All scale items were 7 point Likert-type scales. Scale items are shown in Table 2.

Table 2.

Items of the scales in Study 2.

	The brand of your preference…
CSR associations
csr1	Tries to manage economic resources well
csr2	Tries to improve the working conditions of collaborators
csr3	Tries to contribute to the improvement of the communities in which it works
csr4	Tries to make contributions to social causes
csr5	Tries to promote the environment sustainability
csr6	Tries to have behave in an ethically responsible manner
Customer satisfaction
sat1	The brand meets my expectations
sat2	The brand has the expected quality
sat3	I am satisfied to buy this brand
sat4	I have done the right thing by buying this brand
Brand attitude
att1	I think it is a brand that offers products with good performance
att2	It is a brand I can rely on
att3	I find it is a brand that offers a good value for money
att4	I think it is a brand with a good image
att5	I think it is a suitable brand for sportswear

Measurement reliability and validity was tested by running confirmatory factor analysis (CFA) of all the multi-item constructs in our framework using EQS 6.1 software. Raw data screening showed evidence of non-normal distribution (Mardia's coefficient normalized estimate = 76.90). Although other estimation methods have been developed for use when the normality assumption does not hold, the method of correcting the statistics rather than using different estimation methods has been followed to provide robust statistics (Satorra & Bentler, 1994). Table 3 shows the main goodness of fit indicators for the measurement model and the values of the indicators calculated to examine the model's psychometric properties. The values for BBNFI = .924, BBNNFI = 9.33, CFI = 9.46, IFI = .946 and RMSEA = .080 show that the model offers good global fit as the corresponding critical values are exceeded (Hair, Black, Babin, Anderson, & Tatham, 2010).

Table 3.

Confirmatory factor analysis: measurement of psychometric properties.

Factor	Item	Convergent validity		Reliability
		Factor loading (robust t-value)	Loading average	α	CR	AVE
CSR associations (CSR)	csr1	Eliminated	.83	.92	.92	.69
	csr2	.77 (13.44)^a
	csr3	.84 (16.50)^a
	csr4	.86 (15.42)^a
	csr5	.86 (14.35)^a
	csr6	.82 (9.00)^a
Satisfaction (SAT)	sat1	.90 (2.24)^a	.88	.93	.93	.76
	sat2	.85 (2.25)^a
	sat3	.89 (2.23)^a
	sat4	.84 (2.23)^a
Brand attitude (ATT)	att1	.82 (2.24)^a	.83	.90	.90	.69
	att2	.92 (2.25)^a
	att3	Eliminated
	att4	.84 (2.23)^a
	att5	.75 (2.23)^a

	BBNFI	BBNNFI	CFI	IFI	RMSEA
Goodness of fit indexes
S-B χ^2 (109) = 355.39 (p = 0.00)	.924	.933	.946	.946	.080

Note: α = Cronbach's alpha; CR = composite reliability; AVE = Average Variance Extracted.

a

p < .01.

Table 3 also demonstrates the high internal consistency of the constructs. In each case, reliability indicators were higher than their corresponding desirable values; Cronbach's alpha exceeded the literature recommendation of .70, composite reliability was higher than .60 (Bagozzi & Yi, 1988) and calculation of Average Variance Extracted (AVE) resulted in values greater than .50. As evidence of convergent validity, the CFA results indicate that all items are significantly (p < .01) related to their hypothesized factors and all standardized loadings are higher than .60 (Bagozzi & Yi, 1988) and the averages of the item-to-factor loadings are higher than .70 (Hair et al., 2010).

Finally, the measurement model was checked to ensure discriminant validity. Firstly, it was found that inter-factor correlations were significantly below one, through calculation of the corresponding confidence intervals (Φ-value ± two standard errors, see Table 4) (Anderson & Gerbing, 1988); secondly, for each pair of factors, it was verified that the difference of χ² between the proposed measurement model and a restricted model where the correlation between said factors was set at 1 was significant. Finally, the Variance Extracted test showed that AVE for each factor was higher than the square of the correlation coefficients with each of the other factors (this condition was fulfilled for every factor, except SAT and ATT; see Table 3). Thus, the positive results of the three tests make it possible to confirm the measurement model's discriminant validity.

Table 4.

Discriminant validity.

	CSR	SAT	ATT
CSR	.69	.26	.25
SAT	[.41;.61]	.69	.70
ATT	[.40;.60]	[.75;.92]	.76

Note: Diagonal represents AVE; above the diagonal is the shared variance (squared correlations); below the diagonal is the 95% confidence interval for the estimated factors correlations.

Results

Table 5 shows the standardized coefficients of the structural relations contrasted with their associated t value and the verification of the corresponding hypotheses.

Table 5.

Structural equation modeling: causal relations analysis.

Hypotheses	Structural relation	β	Robust t	Contrast
H1	CSR → satisfaction	.12	2.44^a	Accepted
H2a	CSR → brand attitude	.52	7.66^b	Accepted
H2b	Brand attitude → satisfaction	.77	12.62^b	Accepted

	BBNFI	BBNNFI	CFI	IFI	RMSEA
Goodness of fit indexes
S-B χ^2 (114) = 441.75 (p = 0.00)	.964	.986	.989	.989	.034

a

p < .05.

b

p < .01.

The goodness of fit measurements for the structural model show good global fit (BBNFI = .960; BBNNFI = .914; CFI = .928; IFI = .929; RMSEA = .091). The Lagrange multipliers test did not suggest the inclusion of any new structural variable between the latent variables which would improve model fit and so the proposed theoretical model was regarded as valid. As the results show, CSR brand associations have a direct significant influence on satisfaction (β = .12; p < .05; H1 accepted) and an indirect influence through their effect on attitude (β = .52; p < .01; H2a accepted); and attitude affects satisfaction (β = .77; p < .01; H2b accepted). To ensure the robustness of our findings about the role exerted by attitude, a mediation test was conducted to analyze the role of attitude in the relationship between CSR and satisfaction.

The classical mediation test of Baron and Kenny (1986) was adopted to ensure that brand attitude exerts a mediation effect on the influence of an independent variable (i.e., CSR associations) on a dependent one (i.e., satisfaction). The existence of a mediation effect will be supported if these conditions are fulfilled (Baron and Kenney 1986, p. 1176): ‘(a) variations in levels of the independent variable significantly account for variations in the presumed mediator, (b) variations in the mediator significantly account for variations in the dependent variable, and (c) when a and b are controlled, a previously significant relation between the independent and dependent variables is no longer significant [full mediation] or it is significantly decreased [partial mediation]’.

Although Baron and Kenny's test is very simple and intuitive, it has an important shortcoming: it involves regression equations and, thus, manifest variables have to be used instead of latent variables measured through indicators (i.e., Structural Equation Modeling) (Holbert & Stephenson, 2003). As Baron and Kenny (1986) recognize, like any regression, their basic approach makes no particular allowance for measurement error, which is simply subsumed into the overall error term. Because of this shortcoming and consistent with the approach adopted in our research we have preferred to use Structural Equation Modeling to assess the classical mediation test of Baron and Kenny (1986), following the recommendations by Iacobucci, Saldanha, and Deng (2007), because multi-item scales are preferable according to classical test theory and notions of reliability that more items comprise a stronger measurement instrument (Iacobucci et al., 2007).

We proceeded as follows: we estimated an additional model to test the direct effect of CSR associations on satisfaction, reporting the standardized β of the direct relationship and the R² of the dependent variable. Then, the direct effect model was compared with the original mediation model, providing the effect size (f2) of the R² variation (Cohen, 1988); levels higher than .35 denote a large effect size and if ‘full’, ‘partial’ or ‘no’ mediation exists. Results of these estimations are shown in Table 6. As hypothesized in H2, mediation test results confirm a partial mediation effect of brand attitude in the relationship between CSR and satisfaction.

Table 6.

Mediation test.

	Direct model		Mediation model
	β	R^2	β	β ind.	R^2	f^2	Mediation
CSR → satisfaction	.51^a	.26	.12^a	.39^a	.71	.63	Partial
CSR → attitude			.52^a
Attitude → satisfaction			.77^a

a

p < 0.05.

β are standardized coefficients.

β ind.: indirect effect.

Conclusions, implications, limitations and new lines of research

According to Study 1, results suggest, firstly, that disaggregating the CSR policy into different initiatives such as employee education and training actions has a positive effect on customer satisfaction, and in turn the CSR practiced by employees favors the development of their social and professional skills. Secondly, environment-related initiatives also have a positive impact on customer satisfaction. This finding is in line with Miles and Covin (2000) who justify the relation between environmental responsibility and economic impact based on the effects of environmental social initiatives on reputation, as the sum of credibility, trust, reliability and responsibility. These conclusions reinforce the results of Bird, Hall, Momente, and Reggiani (2007) who show that consumers and the market value the fact that firms undertake a minimum number of environmentally friendly initiatives and that they are proactive in involving employees in CSR practices. Finally, based on Study 1, CSR communication initiatives also influence consumer satisfaction, but negatively so. This finding might suggest that the origins of the negative effect of corporate communication are based (i) on consumer perception of the CSR initiatives as not brand coherent (communication of CSR initiatives is not consistent with company values), and (ii) on CSR initiatives not reported comprehensively, since there is no integration with the external brand communication strategy. This counterintuitive finding may be explained by previous studies which show that (i) the CSR initiatives communicated, under the form of advertising, negatively affects the CSR – value relation if there is an inconsistency between the firm's CSR and the company's overall reputation (Servaes & Tamayo, 2013), and (ii) that the negative effect on consumer behavior of excessive CSR communication which is perceived negatively reduces the perceived sincerity of the firm's motives for CSR initiatives (Van de Ven, 2008). Indeed, Du et al. (2007) also show a gap between CSR initiatives communicated at corporate level by firms, initiatives effectively carried out and perceptions of genuine brand involvement with CSR. This gap can have a negative impact on consumer satisfaction.

Study 2, based on a goal driven survey supplemented Study 1 and shows two mechanisms that influence CSR consumer satisfaction, one directly and the other indirectly through the mediation of brand attitude. The results confirm the direct influence of CSR on customer satisfaction, and Baron and Kenny's (1986) mediation test confirms the mediator role of brand attitude in CSR influence on satisfaction. Consumers react positively through specific relations in their brand relationships to Corporate Social Responsibility initiatives, but consumers will only be affected by concrete CSR initiatives as tested in Study 1. Therefore consumers are connected to the social dimension of the brand through different complementary variables: increased positive reactions that reinforce brand attitude and increased social satisfaction.

Managerial implications

Study 1 may provide firms with a CSR inventory pattern disaggregated into 14 initiatives that can be grouped in three categories anchored in the literature: corporate CSR, inward CSR and outward CSR. The study findings suggest that firms should pay particular attention in their budget and management to environmental initiatives as part of their CSR strategy. It is also recommended that firms inform, train and motivate their employees in CSR practices, making human capital part of the firm's social assets. Finally, business owners and CSR managers should invest in integrated (value coherent) CSR corporate communication able to provide internal and external visibility to CSR initiatives (rather than leaving CSR strategy to be communicated only at corporate level).

The results of the Study 2 model of CSR and consumer behavior suggest that the role of CSR as a source of competitive advantage (Vilanova, Lozano, & Arenas, 2009) is confirmed by brand differential positioning through greater consumer recognition of the firm's social actions and the potential repercussion on purchase behavior. Furthermore, the results of this work therefore suggest reinforcing CSR initiatives which improve the impact on brand attitude and from attitude to consumer satisfaction; so it is important to build up marketing strategies that integrate CSR with consumer orientation from the double dimension of the model validated in the work: satisfaction and brand attitude. Compliance with this objective would be aided by developing relational marketing tools to segment consumers based on their social profiles which integrate the social dimension of their CSR consumer brand relationships.

Finally, as a common managerial implication for Study 1 and Study 2, it can be added that companies may be encouraged from a public policy perspective to develop corporate social initiatives that share active social responsibilities with consumers. In this way, common social values for companies and consumers, would be aligned and practiced, while optimizing the effectiveness of CSR initiatives and increasing consumer “social satisfaction”.

Limitations and future research

As limitations of Study 1, the conclusions are based on only 65 firms in the United States and so the findings are not directly generalizable to other markets because the social demand for CSR initiatives may be different. As we discussed earlier the context may act as moderator variable. This assumption might be easily overcome using ACSI database in other countries with different intensities of CSR initiatives. Study 1 also has the limitation of a certain lack of uniformity as regards the description of CSR initiatives and their quantification, as each firm provides different levels of detail on its CSR activity. Study 2 has limitations, such as non-integration in the model of other antecedents of attitude like trust and customer identification with the brand; and not exploring consumer responses to different levels of a wider disaggregation of CSR initiatives.

This research also suggests the need to address new lines of study from the CSR demand side. In particular, individual consumer preferences for CSR initiatives might be included as a moderator variable. Capturing preferences allows in-depth estimation of the relationship between each CSR initiative and satisfaction with the brand; and from the CSR supply side, how brands may shape consumer demands, and implement a response to them, while assuming consumer co-creation and participation in CSR initiatives. Finally a further line of research would be analyzing the influence of the credibility, using source credibility, of the CSR reports issued by firms. Adding credibility and trust in the firm may shed more accurate light on consumers’ assessments of each initiative. Although literature shows evidence of the relationship between satisfaction and firm's performance through market share, the productivity of the firm is gaining attention as Rust, Ambler, Kumar, and Srivastava (2004) suggested. In the CSR domain this is of special interest due to the long term commitment of such initiatives and the corporate level adopted in many companies.

Conflict of interest

None declared.