The Relationship of Corporate Social Responsibility on Digital Platforms, Electronic Word-of-Mouth, and Consumer-Company Identification: An Application of Social Identity Theory

Abstract

The concept of corporate social responsibility (CSR) is an ever-evolving concept in the field of business management. Even in 2021, its boundaries are evolving and researchers are linking the concept of CSR to different variables to achieve different outcomes. However, the concept of CSR in the healthcare sector is not well-explored in prior literature. The current study is an application of social identity theory to induce electronic word-of-mouth (eWOM) from consumers for a specific brand, through its CSR engagement on social media (CSRS) and consumer-company identification (CCI) in the healthcare sector of an emerging economy. The data of the current survey were collected from different patients of four large hospitals in a large city through a self-administered questionnaire (paper-pencil technique). To validate different hypotheses of the current study, the authors employed the structural-equation-modeling (SEM) technique using AMOS software. The output of SEM analysis confirmed that CSRS positively influences eWOM, and CCI mediates this relationship. The findings of the current study will be helpful for policymakers in the healthcare industry to improve their understanding of CSRS and CCI, inducing eWOM through the lens of social identity theory.

1. Introduction

Corporate social responsibility (CSR) comprises all the varied societal practices of an organization to boost the congruence between the societal expectancy of an enterprise and stakeholders’ behavior [1]. In response, various researchers have demonstrated that a good relationship exists between an organization’s CSR engagement and its economic performance [2,3,4] and social performance [5,6]. In this regard, the effectiveness of communication is a matter of great concern for all businesses in the recent era, because effective communication can improve the evaluation of an organization on the part of consumers [7]. Different scholars have acknowledged the importance of effective communication to induce consumers’ identification with a brand [8,9]. The advent of social media and their corresponding platforms (e.g., Facebook, Twitter, Instagram, etc.) offers organizations new opportunities of communication to remain in touch with consumers for the sake of conveying the organization’s activities to them [10].

CSR is a well-known term of management that combines business goals with social norms. It is used as a strategy to ensure competitive differentiation, both in terms of business dynamics and community-oriented goals [11]. Organizations believe that the enthusiasm for CSR activities allows for strategic competition and sustainable development [12]. Firms often incorporate CSR strategies into their branding infrastructure for building and creating a strong brand [13]. CSR has been ubiquitous in different countries and in different industries for brand building [14]. Past researchers have accepted CSR as a key enabler for building strong brands [15] and concluded that CSR leads to brand authenticity [16], positive brand relations [17,18], perceived quality [19], and building strong brand equity [20]. Other scholars like Nadanyiova and Durana [21], have argued that CSR is a tool for brand enhancement. Similarly in a study, Baskentli, Sen, Du and Bhattacharya [8] confirmed that when consumers’ moral foundations are congruent with CSR domains, positive pro-company behaviors increase.

Businesses in the recent era have comprehended the importance of being engaged in corporate social responsibility (CSR) dialogues and efforts as a way to socialize with the stakeholders, including consumers [22]. Effective communication of an organization’s CSR activities absolutely affects its stakeholders’ engagement with the organization [23]. It is well-established in the literature that through their engagement in CSR, organizations do not merely expect better stakeholder’s attitude (consumers in this case) but build strong stakeholder–business relationships [24]. Businesses have been increasingly depending upon digital platforms to communicate their CSR activities, especially through interpersonal networking websites on social media [22]. This shift from conventional media to social media platforms has been acknowledged after realizing the fact that the consumers’ involvement might be enhanced using virtual platforms because virtual platforms are flexible and interactive in comparison to conventional media [25]. Accordingly, businesses feel that social media is an effective communication medium for serving different communication objectives, including CSR communication [17]. It is contended by different extant scholars that the interactive capability of social media can boost the efficacy of CSR communication with stakeholders, including consumers [26,27]. For that reason, social media websites have come to be a useful platform for organizations to share their CSR engagement using these digital platforms [28]. Furthermore, organizations with high CSR engagement have a tendency to create a bigger online presence (e.g., number of followers) when compared to organizations with lesser involvement in CSR [29].

The extant literature shows that consumers consider and search for the electronic word of mouth (eWOM) for a product while they make a purchasing decision [30,31]. However, the exploration of extant literature regarding CSR communication on social media (CSRS) has revealed that it is not known how CSRS can induce eWOM on the part of consumers for a specific brand. There have been some studies highlighting the importance of CSRS and eWOM [32,33], but most of these studies are inconclusive. Hence, there is a dire need to conduct more research in this area. Thus, the basic aim of the current study is to explore the relationship between CSRS and eWOM by applying social identity theory [34]. The current study also proposes consumer–company identification (CCI) as a mediator in this relationship.

This study has selected the healthcare sector of Pakistan (HCS) to test the proposed relationships. There are some specific logics to choose HCS to serve the objectives of the current survey. The first logic lies behind the argument that HCS is a kind of service sector in which humans (the patients) are the direct receivers of the service delivery by a specific hospital. Hence, the consumers of health-related services are well-concerned while choosing a specific hospital as compared to other service sectors. Due to the patients (consumers in this case) being direct receivers of health delivery services by a hospital, they are less likely to take the risk of experiencing a hospital of which they are not sure. Thus, patients search for eWOM about a hospital in order to assess the quality of delivery of healthcare service. Hence, possessing a good eWOM for a hospital is of utmost importance. The importance of eWOM is also acknowledged by different extant researchers in the healthcare context [35,36,37]. The second logic for choosing HCS is that most of the hospitals in Pakistan are engaged in CSR activities related to philanthropic orientation; for instance, treating the poor free of cost or providing medicines to needy patients, etc. However, this sector is ignoring the importance of CSRS to induce eWOM, which is very important for every organization. Findings of the current study will be helpful for this sector in improving the understanding that through well-planned CSR strategies and communicating them with the stakeholders like patients, eWOM can be induced. Hence, based on these arguments and support from extant researchers, the selection of this sector is not without logic.

The current study significantly adds to the previous literature from the domain of business management. In this regard, the first contribution of this study is that it supplements extant literature from organizational related attributes such as CSR-related attributes of an organization to induce eWOM. In this context, the majority of the literature from the domain of business studies has stressed product/service-related attributes to enhance eWOM on the part of consumers. For example, the relationship of product/service quality and eWOM [38,39], product features and eWOM [40], or brand reputation and eWOM [41]. The authors’ argument here is that although product-related attributes of an organization are important to induce eWOM; however, organizational-attributes such as CSR activities can produce far better consequences for an organization to earn eWOM from the consumers’ side because the organizational-related attributes of an organization make it clear to consumers that the organization is not only concerned with products, but with society and the environment through CSR engagement. Another contribution of the current study is that it highlights the relevance of HCS with CSRS to earn a positive eWOM from healthcare consumers (patients). Although the relationship of CSRS and eWOM is established in the existing literature [32,42,43], the healthcare industry received little attention from contemporary scholars. Likewise, this study uses the phenomenon of CCI as a mediating variable in the proposed relationship of CSRS and eWOM from the perspective of social identity theory. In this context, the authors’ argue that following social identity theory, the patients proudly identify with a hospital that is socially responsible, and thus they put forth effort to support that hospital as a member of the group (the hospital in this case) through spreading positive WOM with others. Last but not least, this study attempts to explore the relationship of CSRS, CCI, and eWOM under a single research model (Figure 1) that is not well-established in prior literature. The remainder of this article is composed of the following sections: the next section deals with literature review and theoretical framework to formulate different hypotheses of the current study. Following, there is the methodology section in which the authors have discussed the sampling procedure, instrument, and social-desirability-related issues. The last two sections of the current study include the result and analysis section in which the authors have drawn different empirical results to validate the hypotheses and the discussion part, which deals with the discussion of empirical results, along with implications.

2. Theory and Hypotheses

The current study is underpinned in the theory of social identity by Tajfel [34] to seek support for hypotheses formulation. According to this theory, the individuals identify themselves with a social group due to belongingness for that group which is based on their emotional attachment to that group. Moreover, the theory of social identity contends that an organization is likely to change the behavior of the individuals through modifying their self-concept towards the organization. The theory of social identity has been extensively used by various scholars to explain how consumers are urged to develop their identities and belongingness with an organization that is involved in CSR activities. The organization’s engagement in CSR motivates the consumers to identify themselves with a socially responsible organization [44,45]. As an instance, Schlosser, et al. [46], built a framework for CCI and argued that CSR activities of an organization encourage the consumers to identify with a specific organization as a socially responsible organization is expected to fulfill their need for self-definition. This is why, when consumers’ self-concept is in line with different traits of CSR activities of an organization, they build a stronger identification with the organization which stimulates them to evaluate a brand positively and enhance their positive behavioral intentions [47,48].

In the context of the current study, this theory is relevant because, when the patients observe that a hospital has some serious concerns for society and the environment by engaging itself in different CSR activities, it helps in changing their self-concept for the hospital. The patients thus begin to identify themselves with that hospital as a member and support such socially responsible hospitals everywhere. Thus, the patients use positive WOM for a hospital on different digital platforms due to its engagement in different CSR activities. Social identity theory has been extensively used by different scholars to explain the behavior of consumers [49,50,51].

CSR engagement of an organization is regarded as how organizations communicate their CSR-related activities with stakeholders for commercial reasons. This study uses the definition of Sampaio, et al. [52] for defining CSR engagement as “ it is referred to as an organization’s response and seriousness towards society and the environment-related issues”. The importance of CSR engagement for businesses has gained a lot of momentum recently because stakeholders such as consumers are showing a serious concern to preserve society and nature, and expect businesses to adopt sustainability practices [53]. Furthermore, CSR engagement of an organization promotes transparency among all units of an organization for creating value for their consumers [54]. Thus, the importance of CSR engagement for an organization is obvious.

Social media has emerged as an important part of everyone’s daily life, which has greatly improved people’s awareness of their choices. This media facilitates the user with a friendly environment [55]. It is also used by organizations to share information regarding products, consumer-feedback, and communicate with individuals and the public [56]. Specifically, this sort of networking is used by a number of businesses of diverse sizes and types to better convey their societal participation [57]. Using social networking can improve the efficacy of CSR communication because the content on social media can readily be shared with others. For example, it is helpful to decrease information asymmetries by creating awareness among individuals regarding organizational activities [23]. Regardless of the significance of CSRS, the extant literature did not explore this relationship well enough as most of the studies have only explored the impact of social media on organization–stakeholder relationships [58,59]. However, the impact of CSRS on eWOM still lacks in the existing literature.

Consumers trust the opinion of their friends and peers, and perhaps this is the reason why WOM has become one of the most effective sources of promotion for organizations in the current digital age. Based on a study conducted by Nielsen [60], more than 90% of consumers believe in the suggestions from their friends in comparison to advertisements. Moreover, WOM does not stop after a single discussion between friends, peers, and family members as once starts, the discussion about a specific product or brand continuous as different people continue commenting and sharing about that brand [61]. Thus, the importance of WOM to induce consumers’ preferences cannot be denied by any means. Positive WOM is associated with positive communication for an organization that the consumers willfully share with others [62]. The relevant role of WOM is well-explained in literature for gaining new consumers [63,64]. In particular, WOM has been shown to have an impact on attitudes, knowledge, understanding, expectations, and behaviors [65]. Therefore, WOM is very important for organizations, which can use WOM constructively to increase business success.

With the advent of digital technologies, social media has facilitated the concept of WOM that is regarded as, eWOM. WOM in the online context is defined as “any good or bad statement about a brand or product by the consumers that is available to the public online is called eWOM” [66]. Organizations in the current digital age extensively use digital platforms (like Facebook, Instagram, Twitter, etc.) to serve different communication purposes. One of these purposes is the CSR engagement of an organization with stakeholders on various digital platforms [24]. These sorts of online relationships with stakeholders are developed using a network of people who are interested in receiving what is shared by the organization regarding social responsibility and others [28]. Thus, building social relationships and peer-to-peer communication is of paramount importance among social media users. In this regard, comments posted on a digital platform via eWOM are considered trustworthy by users because they are based on the similarity of the group [67].

Furthermore, companies are involved in various CSR activities for a number of reasons. One such reason is to gain consumers’ attraction for the brand. The organization’s CSR engagement leads consumers to talk about the responsibilities of an organization in the community to their friends and family. Therefore, involvement in CSR helps an organization to build its socially responsible image [68]. There are different studies that have established that an organization’s CSR engagement on social media can induce eWOM on the part of consumers [32,42,43]. Therefore, in line with the theory of social identity, CSR engagement of an organization develops positive feelings among consumers. Thus, they want to identify themselves with such organizations that are socially responsible. Hence, as a member of the social group (organization in this case), the consumers support the organization and promote it among their friends, peers, and family members by using positive eWOM. Based on the above literary discussion and support from social identity theory, the authors propose the following hypothesis.

Hypothesis 1 (H1).

CSRS engagement of an organization directly relates to electronic word-of-mouth.

The current study defines CCI in line with the definition of Bhattacharya and Sen [69], who proposed that “CCI refers to a social relationship between an organization and its consumers”. This is critical for an organization to note that while considering CCI, it is essential to be mindful that consumers recognize not just the organizations’ goods and services but also the organizations [70]. The consumer in our current age is concerned about what the organization is doing in other areas not related to product performance; for example, the contribution of an organization towards society and the environment [17]. In this scenario, CSR participation of an organization has a significant function in the organizational identity which may induce consumers to spot and produce feelings of connectedness with an organization that is socially responsible [71]. The consumers get inspired by an organization that is engaged in CSR activities and they establish this belief that the organization is not only concerned with profit-maximization, but also has a concern for the betterment of society and the environment. Hence, CSR engagement of an organization boost CCI on the part of consumers. This argument has also been supported by extant researchers [72,73,74,75].

The capability to establish congruence between consumers’ anticipation and CSR actions depends upon the way the CSR is communicated with various stakeholders, including consumers [76]. Together with the diffusion of social networking webpages, like Facebook or Twitter, individuals are no longer passive recipients of the communication. Instead, they actively take part in the analysis and production of different content on social media, which contributes to an increased consciousness about firms’ CSR participation and, consequently, greater CCI [58]. Due to this, consumers that know an organization’s CSR involvement are likely to produce a stronger sense of identification in contrast to individuals that are unaware of such actions [57]. Furthermore, people are more inclined to identify with socially accountable actions to state moral or ethical values which improve their positive feelings for an organization [77].

Following the social identity theory, consumers with a stable CCI are more likely to take part in discretionary role behavior, like a positive brand evaluation [78], encouraging the products of a sustainable brand [79], and boosting their level of brand loyalty [70]. In like manner, when consumers start identifying themselves with a particular brand they are expected to promote that brand as they strongly feel as a member of the social group (the organization). Thus they recommend the brand to others on different social media platforms and use positive eWOM [80]. It is probable that part of consumers’ willingness to discuss businesses they consider to be socially responsible relies on their identification with the organization since it satisfies their self-esteem being the consumer of a socially responsible organization. An organization’s CSRS engagement supplies an encouraging perspective for consumers to rate the organization and participate in conversations with other people that might result in favorable eWOM [43]. Social media platforms permit people to take part in collaborative and co-creative surroundings, in addition to providing numerous means for organizations to get in touch with their consumers and build trust and relationships with them. Thus, a stable relationship of consumers with a socially responsible brand is developed, and hence the following set of hypotheses is proposed.

Hypothesis 2 (H2).

Consumer-company identification directly relates to electronic word-of-mouth.

Hypothesis 3 (H3).

Consumer-company identification mediates between CSRS and electronic word-of-mouth.

3. Methodology

The current research study selected the healthcare sector of Pakistan to test the proposed relationships. To achieve this, the authors have chosen four large hospitals from the city of Lahore in Pakistan. These four hospitals included Hijaz hospital (HH), Pakistan Kidney and Liver Institute and Research Centre (PKLI), Shaukat Khanum Memorial Cancer Hospital and Research Centre (SKMCH&RC), and Iqraa Medical Complex. There were specific reasons for choosing these hospitals, such as that all these hospitals are engaged in different CSR-related activities. Moreover, these are large, famous, and state-of-the-art hospitals of Lahore that treat a large number of patients on daily basis. Hence, these selected hospitals have their recognition among patients. Likewise, these hospitals have also established their separate pages on different social media platforms from where they communicate with different stakeholders including the consumers. It is also to be noted that there is reasonable fan-following of these hospitals on different social media webpages. Thus, the selection of these hospitals is obvious to serve the purpose of the current study.

This study considered the patients as respondents who visited the selected hospital for some treatment. Before starting the actual data analysis phase, the authors contacted spokespersons of the selected hospitals to seek their support and permission to collect the data from their staff. The authors also signed an agreement with the ethical bodies of these hospitals to maintain ethical standards in the process of data collection. Further, the authors got informed consent from every respondent to participate in the survey voluntarily. The respondents were also given a choice to leave the survey at any stage if they feel uncomfortable. After seeking formal approval from the officials of each hospital, the authors arranged for the data collection process. The wake of Covid-19 posed serious challenges for the authors during the process of data collection and hence, the authors had to arrange for special protocols in this regard. Thus, the authors had to stay for long hours in hospitals for the sake of data collection due to strict protocols in place for Covid-19. The data were collected from the patients who visited one of the sampled hospitals during January 2021 and February 2021. The authors asked two screening questions to assess the suitability of the respondents. The first question was asked to assess if the respondent has basic knowledge of CSR and the second question was asked to know whether the respondent is an active social media user or not. The respondents who qualified for both screening questions were then served with the print version of the survey (paper-pencil technique). The authors distributed a total of 800 surveys among the respondents of these four hospitals and received back 409 surveys from different respondents. Hence, the response rate of the current survey was 51.12%. The authors followed the guidelines of Roscoe [81] in deciding the sample size for the current survey. He suggested that the sample size should be larger than 10 times the number of items and hence the sample size of 409 is justified. This approach is also supported by Sekaran and Bougie [82], Hair, et al. [83], Hulland, et al. [84] and Kline [85].

This study followed the ethical guidelines given in Helsinki Declaration. The authors also got approval from the ethical committee of PKLI before starting the actual phase of data collection.

Measures and Handling of Social Desirability

The current survey used already-established scales to measures the constructs. Thus the issue of validity and reliability was non-existent here because the adapted scales have their pre-established validity and reliability. The scale of CSRS was taken from the studies of Klein and Dawar [86], and [87] that consisted of five-items. Four-items of CCI were adapted from Mael and Ashforth [88] and three items of eWOM were taken from the study of Zeithaml, et al. [89]. The authors used a five-point Likert scale for the current survey.

There are various issues attached with self-reported surveys. For instance, one of the disadvantages of self-reported surveys is the way the respondents behave in response to the survey [90]. Likewise, another issue with self-reporting surveys is that responses may be exaggerated by the respondents [91]. Another issue that is associated with self-reported surveys includes different biases that can influence the results; for example, social desirability bias is one of such biases that can produce inadequate results. Similarly, sometimes the respondents are smart enough to judge the hypotheses of the study, and based on their judgment, they may provide biased responses [92]. In order to address the issue of social desirability, the authors took several measures. For example, the survey items were randomly scattered throughout the questionnaire. The authors did this in order to break any sequence of answering the responses by the respondents. This step is also helpful in dealing with the likelihood of any liking and disliking for a particular construct. Likewise, the instrument was checked for accuracy and suitability by experts in the field. This step is necessary in order to address any ambiguity or confusion in any item statement due to complex or dual-meaning words. And likewise, the authors requested the respondents for their true response so that the findings generated by their input may reflect the reality. Different scholars also recommend these steps to mitigate the level of social desirability [73,93,94,95]. The demographic-related information of the sample is presented in

Table 1. Demographic information of the sample.

Demographic	Frequency	%
Gender
Male	237	57.95
Female	172	42.05
Age-group (Year)
20–25	48	11.73
26–30	109	26.65
31–35	136	33.25
36–40	73	17.85
Above 40	43	10.51
Education Level
12 Years	42	10.27
14 Years	128	31.29
16 Years	157	38.39
Higher	82	20.05
Total	409	100

.

4. Results

4.1. Common Method Bias

The current study obtained all the information of the survey from the same individual and hence there may be the existence of common method bias (CMB). The authors took two steps to detect whether the issue of CMB exists in the data or not. In this regard, first of all, the authors applied Harman’s single-factor analysis (HSFA) [96]. Following the standard guidelines of HSFA [97] if CMB exists, then a single factor will explain the majority of variance in predictor and criterion variables. In the current scenario, the results of HSFA disconfirmed the existence of any of such factors because the largest variance shared by single-factor was 34.83 percent. Hence, the authors established that there is an absence of CMB for the current study. Yet, to further validate the absence of CMB, the authors applied one-factor confirmatory factor analysis (CFA) in which all the variables were allowed to load on one other variable. The result of model-fit-indices (MFI) validated the absence of CMB because all MFI were poor (CFI = 0.493, GFI = 0.511, NFI = 0.592, RMSEA = 0.161). Thus, based on these findings, the authors confirmed the absence of CMB in the dataset of the sample of the current study.

4.2. Convergent Validity, Factor Loadings, and the Reliability Analyses

The results of HSFA and CFA proved that there is no issue of CMB and hence the authors took the data analysis to the next level and performed different tests of validity, reliability, and factor-loadings. These results have been reported in

Table 2. Item loadings, convergent validity and reliability results.

Item	λ	λ2	∑λ2	Items	AVE	C.R
CSRS1	0.77	0.59
CSRS2	0.73	0.53
CSRS3	0.70	0.49
CSRS4	0.83	0.69
CSRS5	0.79	0.62	2.929	5	0.586	0.876
CCI1	0.68	0.46
CCI2	0.76	0.58
CCI3	0.70	0.49
CCI4	0.79	0.62	2.154	4	0.539	0.823
eWOM1	0.84	0.71
eWOM2	0.71	0.50
eWOM3	0.73	0.53	1.742	3	0.581	0.805

Notes: λ = Item loadings, C.R = composite reliability, ∑λ² = sum of square of item loadings.

for the readers. One can see from Table 2 that the factor loading results for each construct were sufficient (every-item showing a loading greater than 0.5). Thus, there is no issue with the item’s factor-loading. Next, the authors tested for the results of average-variance-extracted (AVE) to establish convergent validity and composite reliability (C.R) to establish inter-item consistency. To validate convergent validity, the general criterion is that if AVE for a construct is beyond 0.5 then the criterion for convergent validity is satisfied [98]. To measure AVE for a construct the authors first of all calculated item-loadings (λ) for each item and then took the sum of the square of these loadings (∑λ²) dividing by the number of items. As an example, there were five items of CSRS and the sum of the loading square was 2.929 which was divided by 5 (number of items) that resulted in 0.586 as AVE value. After validating convergent validity, the authors calculated C.R by using the formula C.R = (∑λ)²/[((∑λ)² + ∑(1 − λ²)], for each construct. Following the guidelines of Fornell and Larcker [98], the author checked each construct’s C.R value (C.R should be greater than 0.6) and revealed that each construct meets the criterion of C.R. Thus, there is no issue of C.R in scaled items of the current survey. These results have been reported in Table 2 in detail.

Table 3. Correlation, discriminant validity and MFI.

Construct	Mean	S.D	CSRS	CCI	eWOM
CSRS	4.16	0.54	0.765	0.19 **	0.23 **
CCI	3.88	0.47		0.734	0.29 **
eWOM	4.23	0.61			0.762
Model fit indices	Range	Obtained	Model fit indices	Range	Obtained
χ2/df	5.00	4.183	IFI	0.90	0.911
RMSEA	0.08	0.068	TLI	0.95	0.955
NFI	0.95	0.958	GFI	0.90	0.908
CFI	0.90	0.913

Notes: S.D = standard deviation, ** = significant values of correlation, bold diagonal = discriminant validity results.

presents the results of correlation, discriminant validity, and model-fit-indices. In this aspect, the correlation values for all constructs were positive and significant that means each construct is positively correlated with the other construct. As a case, one can see that the correlation value between CSRS and CCI is 0.19 which is positive and significant. To assess discriminant validity (DSv), the authors took the square-root of AVE (SqAVE) for a construct and compared it with correlation values. The general rule here is that if the value(s) of correlation is less than the value of SqAVE, then it is confirmed that the criterion of DSv is satisfied and the items of one construct are dissimilar with the items of other constructs (s) in comparison. To explain further, the value of correlation between CSRS and CCI is 0.19 which is far less compared to the value of SqAVE (0.734) and hence, discriminant validity is maintained in the current scenario. Lastly, the results of MFI are also presented in Table 3 to assess if there is a fit between theory and the data. In this regard, the authors checked the values of different MFIs against their acceptable range (given in Table 3) and revealed that there is no issue in any MFI value (χ²/df = 4.183, RMSEA = 0.068, NFI = 0.958, CFI = 0.913, IFI = 0.911, TLI = 0.955, GFI = 0.908).

4.3. Hypotheses Testing

To validate different hypotheses of the current study, the authors used structural-equation-modeling (SEM) in AMOS software. SEM analysis is a co-variance-based analysis approach that is very popular among contemporary researchers as it has advanced-level tools to deal with complex research models [99,100,101]. SEM has three major advantages over traditional multivariate techniques: (1) explicit assessment of measurement error; (2) estimation of latent (unobserved) variables via observed variables; and (3) model testing where a structure can be imposed and assessed as to fit of the data [102]. Some limitations are also attached with SEM including the problem of omitted variables, the importance of lower-order model components, and the inaccuracy of some commonly used rules of thumb [103]. However, without a doubt, the use of SEM to evaluate complex models is increasing.

The hypotheses testing was done in two parts using SEM. In the first part, the authors analyzed the direct effect model in which no mediator was introduced in the model. The authors conducted the direct effect model to validate Hypotheses 1 and 2 (H1, H2). The results of structural model confirmed that both H1 and H2 are accepted (β1 = 0.26, β2 = 0.34, p < 0.05). Hence, based on these results (

Table 4. The results for Hypotheses 1 and 2.

Path	Estimates	S.E	CR	p-Value	ULCI	LLCI	Decision
CSRS → eWOM	(β1) 0.26	0.039	6.67	***	0.316	0.941	Approved
CCI → eWOM	(β2) 0.34	0.048	7.08	***	0.298	0.729	Approved
Model fit indices	Criteria	Obtained	Model fit indices	Range	Obtained
χ2/df	5.00	3.691	IFI	0.90	0.922
RMSEA	0.08	0.060	TLI	0.95	0.959
NFI	0.95	0.969	GFI	0.90	0.924
CFI	0.90	0.927

Notes:ULCI = upper-limit confidence interval, LLCI = lower-limit confidence interval, *** = significant values.

), H1 and H2 of the current study are proved and accepted. The first stage of SEM started with checking the direct effect analysis in which there was no intervention of any mediator in the model. The results of the direct effect model are shown in Table 4. As per these results, the direct effect model produced significant results. These results confirmed that the first two Hypotheses H1 and H2 of the current survey are supported. These outcomes were declared based on beta estimates and p-values (β1 = 0.31, β2 = 0.38, p < 0.05). The results further validated that the effect of SL on EIB is stronger as compared to the effect of CSR-E on EIB. Moreover, the model-fit-indices were also significant in this regard (χ²/df = 3.691, RMSEA = 0.060, NFI = 0.969, CFI = 0.927, IFI = 0.922, TLI = 0.959, GFI = 0.924).

The second part of SEM analysis was conducted by introducing CCI as the mediating variable. To do this, the bootstrapping technique was applied which is more advanced and powerful as compared to the traditional Baron and Kenny [104] technique for mediation analysis. During the bootstrapping process, the authors used a large bootstrapping sample of 2000 and analyzed the mediation results. The output of the mediation analysis revealed that there is a partial mediation effect of CCI between CSRS and eWOM. It is worth noting that the partiality of mediation was established based on the reduced beta value which was originally β1 = 0.26; however, after inclusion of the mediator (CCI) it is dropped to β3 = 0.028, confirming that there is a partial mediation effect of CCI between CSRS and eWOM (χ²/df = 3.092, RMSEA = 0.044, NFI = 0.973, CFI = 0.942, IFI = 0.938, TLI = 0.970, GFI = 0.939). Thus, all three hypotheses of the current survey are approved in

Table 5. Mediation and moderation results for H3 and H4.

Path	Estimates	S.E	Z-Score	p-Value	ULCI	LLCI	Decision
CSRS → CCI→ eWOM	(β3) 0.028 **	0.011	2.54	***	0.193	0.311	Approved
Model fit indices	Criteria	Obtained	Model fit indices	Range	Obtained
χ2/df	5.00	3.092	IFI	0.90	0.938
RMSEA	0.08	0.044	TLI	0.95	0.970
NFI	0.95	0.973	GFI	0.90	0.939
CFI	0.90	0.942

Notes:ULCI = upper-limit confidence interval, LLCI = lower-limit confidence interval, **, ***, = significant values, S.E = standard error.

.

5. Discussion and Implications

The current survey was conducted to serve two important purposes. The first purpose of the current study was to test the relationship of CSRS and eWOM in the context of social media, which is not well-explored in the healthcare industry. To this end, the findings of the current study proved that there is a direct relationship between CSRS and eWOM. The patients as the respondents of the current study gave their opinion that CSR activities of a hospital develop positive thinking in them for a hospital that is socially responsible. The CSR engagement of a hospital leaves an impression of honesty and transparency on patients and they feel a socially responsible hospital will be fair and honest enough to treat them with due care. Hence, they make a positive evaluation about such hospitals and feel comfortable getting treatment from such hospitals. Furthermore, the respondents of the current survey also validated that when they get information about CSR engagement of a hospital on different digital platforms, they feel positive about that hospital and recommend such hospitals to others by spreading positive WOM. The social identity theory is also helpful in explaining this whole process. In line with this theory, the CSR engagement of a hospital is a kind of engagement that is well appreciated by the patients and they willingly identify themselves with a hospital that operates under the philosophy of CSR. Hence, a strong and stable relationship is developed between patients and a socially responsible hospital. This relationship is also supported by extant researchers who proposed that CSR engagement of a hospital is helpful in inducing WOM by the consumers [32,42,43].

The second purpose of the current study was to test the mediating effect of CCI between CSRS and eWOM in the online context. In this regard, the empirical results of the current survey validated that CCI mediates the relationship of CSRS and eWOM in the healthcare sector of Pakistan. The respondents of the survey confirmed that CSR engagement of a hospital provides them with a solid base to identify themselves with a social group (the hospital) and they consider themselves a proud member of such hospital. This feeling of identification with a hospital encourages them to put every effort for the success of a hospital for which they experience strong feelings of identification. Hence, they use positive WOM for such hospitals with their friends, peers, and family members on different digital platforms (Facebook, Instagram, Twitter, etc.). Following social identity theory, the current study contends that the socially responsible behavior of a hospital is an organizational attribute that is considered a discretionary attribute of an organization for the betterment of society and the environment. Consumers in the current digital age have great concern not only for the product-related attribute (product quality, performance, and etc.), but they also consider organizational attributes such as CSR engagement of an organization while they make a purchase decision. Hence, in line with the theory of social identity, this study has established that CSRS, directly and indirectly, influences eWOM in the healthcare sector of Pakistan. The above argument of the current study is also supported by previous researchers [71,75,76,105,106].

5.1. Theoretical Implications

There are some important implications of the current study, including implications for theory and practice. The authors, first of all, discuss important implications of the current study for theory. In this aspect, the first implication of the current study for theory is that it enriches the existing CSR literature from the perspective of social media to induce eWOM in the healthcare sector of Pakistan. It is worth mentioning here that most of the previous studies have explored this relationship in the offline domain [107,108,109]; however, it is quite recent that researchers have acknowledged the importance of social media to improve WOM in an online context. Similarly, the healthcare sector in the field of CSR did not receive due attention from the viewpoint of consumers. There have been different studies highlighting the importance of CSR to achieve consumer-related outcomes [17,18,43], but these studies did not consider the healthcare sector. Another important implication of the current study for theory is that it attempts to explain CSRS, CCI, and eWOM in a unified model, whereas most of the previous studies have investigated the separate impact of CSR on WOM [32,106], or CCI on WOM [73,110]. Therefore, the consideration of these three variables under a single model is also an important theoretical implication.

5.2. Practical/Managerial Implications

The practical implications of the current study are also important especially from the viewpoint of the healthcare sector. In this regard, the first practical implication of the current survey is that it highlights the importance of a hospital’s CSR engagement to earn positive word-of-mouth from patients. This finding has special importance for this sector because it has been well-established that in comparison to advertising, the opinion of friends and family members about a product or brand is heavy-weighted. The reason for this argument lies in the logic that advertisement is considered by the consumer an effort to induce sales from the sponsor (the organization). Hence, the information given in advertisements is not taken for granted by the consumers, but the consumers consider the opinion of their colleagues and family members as a genuine evaluation of a brand and hence they prefer WOM over the advertisement. Thus, the healthcare policymakers can improve their understanding of CSR from the findings of the current study and can craft their business strategies around the philosophy of CSR to create a solid base of relationship with consumers. Likewise, the current study has another important practical implication from the viewpoint of CCI for the healthcare sector.

The healthcare sector needs to realize that WOM marketing, if done accurately, can eliminate the need to place any paid advertisements. when patients have a good experience at a specific hospital, they share their experience on digital forums with their peers and friends, and this is more valuable than any paid advertising. Just one recommendation from patients can have a bigger impact than paid advertising. Many marketing experts promote WOM as they know it is more important than an advertisement. WOM marketing in healthcare is becoming an increasingly essential part for building a strong brand. In this regard, CSR is one such tool that can induce positive WOM. Thus, policymakers from the healthcare sector are encouraged to consider CSR activities close to their business operations as a way to induce WOM on the part of consumers. Likewise, the phenomenon of CCI is also helpful to influence eWOM for a hospital that is socially responsible as CSR activities of a hospital provide a base to the patients to strongly identify with a socially responsible hospital. The authors’ argument here is that CSR engagement of a hospital is appreciated by the patients on all grounds, and due to CSR practices of a hospital they strongly identify themselves with such a hospital, making every attempt to promote it among their peers and colleagues. Hence, it is suggested that the healthcare sector stress more and more the CSR philosophy if they want to establish long-term sustainable relationships with their consumers. Lastly, the policymakers from the healthcare sector are encouraged to realize the power of digital platforms to induce eWOM among consumers. The current state of involvement of the majority of the hospitals on social media is limited to the extent of the philanthropic orientation of CSR, such as in informing the community that a hospital treats the patients free of cost or provides medicines freely to the poor. The healthcare policymakers need to realize that they need a special effort on different digital platforms to effectively communicate with consumers and motivate them to spread eWOM for their hospital.

5.3. Limitations and Future Research Perspectives

The current survey may be regarded as a stepping stone to induce eWOM in the healthcare sector of Pakistan through CSRS and CCI in an online context. However, there are some limitations that need to be addressed by future researchers. In this regard, the first limitation of the current study is that it only collected the data from the hospital located in one city. Hence, the geographical concentration of this study may reduce the generalizability. Therefore, future researchers are encouraged to collect the data by including hospitals from other cities like Karachi, Islamabad, Multan, etc. Another limitation of the current study is that it only considered a positive aspect of WOM; however, neglecting negative WOM is not wise for any organization on social media, so it will be interesting for future researchers to investigate CSR engagement of an organization, which may be helpful in mitigating negative WOM for an organization. Another important limitation of the current survey is that it used cross-sectional data, which is weak in establishing a causal relationship, and hence for future studies, using longitudinal data design is advised. Another limitation of the current study is that it attempts to explain consumer behavior such as positive WOM through CSR and CCI, it is worth noting that individual behavior is a complex phenomenon that is affected by several variables. So the authors suggest adding more variables in the model of the current study in order to better explain the behavior of consumers. In this vein, the authors suggest incorporating “identity salience”; consumers’ preference for CSR may be an interesting mediator in the proposed research model of the current study. Lastly, the present study deals with social media in a positive context, but the reality is that the emergence of social media also brings some challenges, such as information privacy being an issue in social media. Additionally, other challenges include information credibility and ethical issues. Future researchers need to deal with these types of issues in order to find solutions.