1 Introduction

Corporate Citizenship, a derivative of Corporate Social Responsibility (CSR), has become a fundamental business practice through increased public expectations and demands placed upon industry conduct. Poor Corporate Citizenship may create negative impacts upon stakeholders, the environment, and, eventually, the firm itself. However, executing socially responsible actions in line with society’s continuously shifting expectations has become arduous, as Corporate Citizenship composes a dynamic phenomenon (Verbeke and Tung 2013) and has no universally accepted definition (Torugsa et al. 2012). As a result, it is ambiguous what constitutes [ir]responsible business behavior. Corporate Citizenship should not favor an instrumental approach or take preference of an individual stakeholder group over other groups but should engage stakeholders in meaningful, two-way engagement. Additionally, because Corporate Citizenship consists of a social construct (Buhanita 2015), it is necessary to include a broad array of stakeholders to comprehensively conceptualize what is deemed [ir]responsible behavior. Since the corporation is a sub-system of the societal macro-system and not a self-contained structure, it is interdependent upon stakeholders (McLennan and Banks 2019), with society’s stakeholders providing the corporation with a ‘license to operate’ (Neron 2016) in exchange for a contribution to the common good through its socially responsible behavior and actions.

Evaluating Corporate Citizenship thus inherits many of these challenges outlined. One of the most widely adopted models that have been operationalized for measurement is Carroll’s Pyramid of CSR, which is despite being established solely for adoption by managers and academics (Carroll 1979, p. 502). This indicates that the model was created to fulfill a specific purpose and may not reflect the perspectives of other stakeholders (Alvarado-Herrera et al., 2017). Additionally, many stakeholders are bereft of knowledge and possess information asymmetry, thus are unable to determine whether the corporation is adequately carrying out its civic responsibilities. This information asymmetry creates a distinction between Corporate Citizenship and perceived Corporate Citizenship, in which Corporate Citizenship is a tangible, objective metric (i.e., this is); however, perceived Corporate Citizenship addresses the information asymmetry by evaluating the often-subjective perspectives of stakeholders towards Corporate Citizenship (i.e., I think). This, in turn, means perceived Corporate Citizenship is defined by stakeholders and may not accurately represent actual corporate performance. However, it is suggested that research on Corporate Citizenship perceptions can be more effectively employed than objective Corporate Citizenship since stakeholders are inclined to act more reliably on their perceptions rather than upon the objective truth and reality (Hansen et al. 2016). However, when operationalizing measurement instruments for perceived Corporate Citizenship, the measurement criteria are frequently chosen on the basis of what can be tangibly or physically measured (i.e., technical) instead of what should be assessed (i.e., normative) (Salvado et al. 2015). Thus, when developing measurement instruments, they must be built in collaboration with the intended users to ensure the normative demands are fulfilled.

The issue of measuring perceptions of Corporate Citizenship and other CSR derivatives was highlighted by Freeman (1984), as he recognized a flaw in the stakeholder concept he developed, with stakeholders only being represented by management’s ability to relate and emphasize with them. There has been conceptual work by Idemudia (2011) suggesting that using a bottom-up approach would allow for the normative stakeholder perspective, as well as McLennan and Banks (2019) proposing that examining the event of interest through a ‘reversed lens’ generates a more holistic picture (McLennan and Banks 2019); however, the top-down approach still appears dominant in many studies. This can be evidenced by Alvarado-Herrera et al., (2017) scale development where the items of the instrument were developed from the top-down dominant literature and Fatma et al. (2014), who developed items of the instrument through qualitative interviews but only incorporated managers, thus again being a top-down perspective. The information asymmetry issue is exacerbated further with items such as ‘This organization has been successful at maximizing its profits’ from Dhanesh (2014), which is clearly a top-down perspective, as stakeholders such as customers or community members are unlikely to check the company’s balance sheet before engaging with the company.

This study thus aims to build upon a prior bottom-up conceptualization of perceived Corporate Citizenship developed from a case study and explore how far this conceptualization generalizes whilst also refining and validating the operationalized measurement instrument in multiple contexts. This is completed through not only Confirmatory Composite Analysis (CCA) but also nomological and known group validation. The paper continues with a brief literature review consisting of Conceptualization of Perceived Corporate Citizenship and Measuring Perceived Corporate Citizenship. Following this, the methodology and results section is divided into the four studies that this paper consists of, with studies 1 and 2 refinings and validating the operationalized measurement instrument with the case study firm, whilst studies 3 and 4 look to generalizing and further validating the measurement instrument in a broadening context, namely national and international. The discussion section deliberates both theoretical implications and methodological implications of the study before concluding the paper and evaluating limitations of the study and suggestions for future research.

2 Literature review

2.1 Conceptualization of perceived corporate citizenship

Corporate Citizenship suggests that the corporation is a fellow citizen within society and should not be managing stakeholders but rather have obligations to them. The overarching concept of CSR generally takes the form of selected implementation with managers ‘cherry picking’ activities based upon stakeholder management (Roth et al. 2018). The ‘management’ of stakeholders suggests the firm is ‘above’ the stakeholders or fellow citizens, which is no longer acceptable. This is based upon a change in thinking in the middle of the last century, when the business went from being considered a sealed unit into a system within another system through Systems Theory (Boulding 1956), where business is but one smaller system within the larger system of society (Lozano and van Haartman 2018). The corporation is part of society, not above it, and therefore can be considered a citizen of society. Thus, Corporate Citizenship complements the stakeholder concept, where the company considers the wants and needs of the stakeholders that comprise society as a whole, where the company affects society with its business activities. Organizations are dependent on the various stakeholder groups of society for their continued existence. When Corporate Citizenship combines with the ‘stakeholder notion,‘ a good Corporate Citizen is an organization that fulfills its’ responsibilities imposed upon it by the stakeholders that have a ‘stake’ or interest in the operations of the company.

Hence, to be able to evaluate Corporate Citizenship effectively, all stakeholder groups’ perspectives need to be accounted for. This method has been acknowledged within the literature as the bottom-up approach (Idemudia 2011), with normative Corporate Citizenship coming from a specific context composed of multiple stakeholder views (Fordham and Robinson 2018). From the firms’ perspective, the expenditure of effort, time, as well as other resources in addressing stakeholder interests is a rational and justified business activity (O’Riordan and Fairbrass 2014). Thus, building this normative view of Corporate Citizenship requires ‘reversing the lens’ and viewing what is deemed as a responsible business from the target community of stakeholders (McLennan and Banks 2019). This view stems from the reverse anthropology used by Kirsch (2006) and complements the bottom-up approach proposed by Idemudia (2011). However, complexity is introduced from the subjective nature of the theories of CSR, and thus its derivative theories, including Corporate Citizenship, as they have no agreed objective definition (Dahlsrud 2008).

This subjective nature thus moves towards the stakeholder perceptions of the phenomenon, which to are subjective. Where perceived Corporate Citizenship has been defined by how a stakeholder perceives an organization as honoring its responsibilities of economic, legal, ethical, and discretionary levied by the stakeholders upon the organization (Evans and Davis 2011; Maignan and Ferrell 2000). However, this study challenges this, as Carroll’s four-factor conceptualization consisting of the economic, legal, ethical, as well as discretionary responsibilities is imposed within the definition, allowing no subjectivity from the stakeholders’ perceptive. Perceived Corporate Citizenship varies in that it contains the subjective element of the stakeholder’s information asymmetry and perceptions of a company. Therefore, this study defines perceived Corporate Citizenship as; how an individual perceives an organization as fulfilling its’ responsibilities imposed upon it by its stakeholders. This then allows each stakeholder group to perceive their own conceptualization of what Corporate Citizenship is. This idea is not new, Evans and Davis (2011, p. 458) state;

‘[Perceived Corporate Citizenship] refers to individual evaluations and interpretations of business activity, not necessarily actual firm actions. Hence, the construct represents perceived corporate citizenship; an individual’s interpretation of the degree of fulfillment and importance of organizational processes, activities, and policies concerning CSR.‘

Whilst Corporate Citizenship and perceived Corporate Citizenship differ in; Corporate Citizenship consists of a tangible objective measure, i.e., how it is, however, perceived Corporate Citizenship consists of the stakeholder perceptions, i.e., I think, about Corporate Citizenship. Both Corporate Citizenship and perceived Corporate Citizenship are needed as the subjectivity of perceived Corporate Citizenship can be influenced by ‘greenwashing’ and other means, whilst objective Corporate Citizenship measures tangible components, i.e., number of employees injured at work, financial performance, etc. Perceived Corporate Citizenship of a company is only what the stakeholder perceives and may not directly reflect the true performance. Nevertheless, it has been debated that research focused upon stakeholder perceptions of Corporate Citizenship can be more of greater use than objective Corporate Citizenship (Hansen et al. 2016); this means that the corporation should listen to its stakeholders’ perceptions. The perception of the stakeholder is also affected by the ability of individuals to transform these perceptions into categories outlined within a measurement instrument (Durham et al. 2011).

From the literature, it was deemed that prior knowledge of stakeholders drastically affected their ability to answer survey items accurately. In order to perceive Corporate Citizenship, stakeholders must be aware of Corporate Citizenship, meaning that stakeholders require some fundamental knowledge. It is suggested by Kim and Ji (2017) that socially responsible activities provide a company with little benefit if consumers are unaware of such socially responsible activities; this could be further extended to all stakeholders. Stakeholders requiring prior knowledge is evidenced by Alvarado-Herrera et al. (2017), where they propose that one of the limitations of their scale is that the consumer must have some prior knowledge or information around the company’s socially responsible activities in order to stop “erratic, unthinking responses or reactions that are marked by social desirability .“This issue of requiring prior knowledge leads towards a measurement instrument needing to originate from the stakeholder perspective in order to get a meaningful response from the participants, as the measurement instrument will accommodate the prior knowledge level.

2.2 Measuring perceived corporate citizenship

Stakeholder typologies offer an opportunity to develop a measurement instrument in a theoretically strong manner. Donaldson and Preston (1995) have interpreted three typologies from the stakeholder theory: normative, descriptive, and instrumental. Freeman (1999) defines these clearly; descriptive typology is the way that the world really is, normative typology prescribes how the world should be, whilst instrumental typology links means and ends. The dominantly used typology is instrumental, as it allows the firm to use stakeholders to achieve its own goals. However, asking stakeholders for their opinions of what the firms’ responsibilities should be, means that the stakeholder paradigm has moved to the normative typology. Whilst the use of the descriptive paradigm by stakeholders is acknowledged within Maon et al. (2010) review of the developmental stages of CSR models. However, a firm can never be fully normative, as the underlying premise is that the firm must turn a profit to exist and is fundamental. Likewise, a firm cannot be fully instrumental, as society has a minimum expectation of firms to act responsibly. This is co-creation of organizational behavior between the corporation and its’ stakeholders are the most desirable purpose to stakeholder engagement (Lane and Devin 2018). The idea of co-creating the organizational behavior supports the Corporate Citizenship theory, as co-creation is created by company-stakeholder and stakeholder-stakeholder interactions (Hur and Kim 2017), thus mimicking the societal interactions that are complex and interdependent.

Furthermore, within the literature, there are multiple conceptual and theoretical ideas that can assist with the development of a measurement instrument. Mirvis and Googins (2006) suggest that the Corporate Citizenship of a firm is developmental as it moves towards normative logic. Therefore, a combination of the developmental nature of Corporate Citizenship and stakeholder typologies can be used to aid the measurement of perceived Corporate Citizenship. Thus, the phenomenon of Corporate Citizenship can be considered a continuum, with the top of the continuum being that the firm is completely normative (should be) and the firm being completely instrumental (means-end) at the bottom of the continuum. This means that the rating a stakeholder gives will be of descriptive typology (what is); this would then interpret as to how far the firm is progressing towards being a Corporate Citizen and the normative logic.

3 Methodology and results

The methodology of this study follows the discussion within the literature review, with a measurement instrument being developed for perceived Corporate Citizenship from a case study company and then psychometric properties tested as the contextual scope is expanded. The methodology begins with the normative conceptualization of perceived Corporate Citizenship developed by concept mapping from a prior study (Homer 2021). This prior study was conducted within Malaysia and gave a list of 101 statements created by the stakeholders of the case study company, with the bottom-up approach being adopted. This means that the items that will be used within the measurement instrument originate directly from the stakeholders themselves. A case study allows for the research to delve deeply into the phenomenon allowing for greater exploration of the contextual component of the phenomenon (Baxter and Jack 2008). Furthermore, the context-dependent nature of case studies is necessary because context-independent and generalized studies have become separated from the real-world context (Joo and Shin 2018). This is particularly true within social science disciplines where theories are particularly broad and thus are detached from the real-world (Trochim 1985). The importance of retaining the conceptualizations’ statements was because they had been developed from qualitative data collected from various stakeholders and then validated by another group of stakeholders, thus not only addressing the lack of bottom-up scale development highlighted in the introduction but also giving strong face validity to these scale statements.

The methodology consists of four studies. The first two studies consist of exploratory factor analysis and then a confirmatory factor analysis to check the stability of the factors derived. This was deemed necessary as the original conceptualization was developed with only a small sample, and thus the generalization of this conceptualization to a wider number of participants was necessary before proceeding to a wider context. The second study also uses additional nomological and known group validity across three different stakeholder groups. Once the measurement instrument had been confirmed and validated in the case study company’s context, studies 3 and 4 were then used to expand the generalizability of the instrument. Both studies selected a representative and similar company from a supermarket, banking, and petrochemicals from Malaysia and the UK. Study 3 then confirms the measurement instrument in the Malaysian context with consumers, using again nomological validation but also known group validity between the three different representative firms. To further validate the measure, the same process as study 3 is then used within the UK concept. Each study is now described and results presented.

Study 1

This study used the case study firm as the subject of evaluation for study 1 from Homer (2021), as it was deemed necessary to confirm and validate the instrument within this specific context before expanding to a more general context. The review paper on the topic of using concept mapping to develop measurement instruments by Rosas and Ridings (2016) was used as a guide to the validation of the measurement instrument. The initial step entails reducing the statement list; however, consideration needs to be given to adequately cover the domain. This was done by retaining the most important statements from the concept mapping, reducing the list by half.

Rosas & Ridings (2016) propose the best way to address the dimensionality of the measurement instrument is through factor analysis, with studies using exploratory factor analysis proceeded by a confirmatory factor analysis which will determine the stability of the factors. This study has adopted this approach to strengthen the scale development process and enhance the validation procedure. The choice for the process of factor analysis is justified as concept mapping from the prior study uses Hierarchical Cluster Analysis and has not been statistically verified. The exploratory factor analysis means another reduction in the number of statements can be achieved by identifying representative variables (Hair et al. 2014), thus reducing participant fatigue within subsequent stages.

For exploratory factor analysis, the study reduced the statement list to the 50 highest rated statements. Along with this statement list of Perceived Corporate Citizenship, a quality check question was used. A Quality Check Question, as used by Gallagher et al. (2018), involves including a statement such as ‘If you are paying attention, select strongly agree on the scale .‘Thus, respondents are removed from the study if they respond in a way other than directed. A convenience sample of students was used, as they are a unique stakeholder group in the context of the study where they transverse the typical boundaries within the literature of stakeholder groups. The sample size aimed for approximately 400 participants to adequately cover 5 to 10 responses per variable (Hair et al. 2014). For this study, students were selected as participants based upon their transboundary stakeholder group, in that many students studying within the case study company private university, visited retail outlets operated by the case study company, lived in properties managed by the case study company, and also sought internship or work opportunities with the case study company. Thus, students fulfilled multiple stakeholder roles of consumers, community members, employees, and as well perspective employees. The students were recruited through convenience sampling and included a wide demographic of college, undergraduate, and postgraduates, including national and international.

Three hundred and ninety-nine respondents submitted the questionnaire; those respondents that incorrectly responded to the Quality Check Question were eliminated, reducing the usable responses to three hundred and seventy-five. The exploratory factor analysis was performed on SPSS 25. Sampling adequacy was interrogated using Kaiser-Meyer-Olkin generating a value of 0.961, which is within necessary limits. A Principal Axis exploratory factor analysis with Varimax rotation was performed. The results can be seen in Table 1, with all loadings over 0.400 being selected based on Dyer et al. (2007). Factors 6 and 7 were disregarded as no variable had a loading over 0.400. Factor 1 has 14 statements; Factor 2 has 12 statements; Factor 3 has 8 statements; Factor 4 has 9 statements, and Factor 5 has 4 statements. Statements “[firm] runs its business efficiently,“ “[firm] products & services are harmless to society” and, “[firm] practices no fraudulent business” had loadings under 0.400 and thus disregarded.

At this point, the statements which were to be retained were assigned an index; these items will be referred to for the rest of the study. For the selection criteria, a trade-off between representing the domain and reducing participant fatigue needs to be considered; based upon this, the six highest loading statements from each factor were retained. Cross loading items that were above the 0.400 cut-offs were also considered for retaining if they were within the six highest loadings of a particular factor and joined the factor that they had the highest loading in. These were retained to test the stability of the cross-loading in the proceeding study. The statement of “[firm] refrains from underpaying its employees” was retained within the 4th factor as this had the higher loading, meaning that factor 5 had only four statements whilst all other factors had six.

Table 1 Exploratory factor analysis results
Full size table

Study 2

Stage 2 was to assess the stability of the factors derived from study 1 based upon the case study firms’ context by using Confirmatory Factor Analysis (CFA). The CFA was performed upon the SmartPLS software, with the PLS-SEM variation of CFA being Confirmatory Composite Analysis (CCA) as suggested by Schuberth et al. 2018. The choice to use PLS had been made as it has been advised that researchers should particularly use PLS-SEM with CCA in the case of measurement models that are indirectly measuring conceptual concepts (Hair and Sarstedt 2019), with the CCA procedure being outlined by Hair et al. (2020).

Nomological Validity was incorporated, with a measure being nomologically valid if it behaves as expected with respect to another construct, where this construct is theoretically related to the measure (Donia et al. 2017), meaning that the measurement instrument will have predictable associations which could be found in prior literature. Academic texts were reviewed to seek empirical studies that had known relationships between perceived Corporate Citizenship and stakeholder-specific constructs. These constructs would be correlated with the perceived Corporate Citizenship instrument and the relationship compared to the literature; students, employees, and consumers were selected for validation. For students, Firm’s Attractiveness was selected from Bhattacharya and Sen (2003) and was expected to have a positive relationship with Corporate Citizenship. Organizational Commitment had been chosen for employees with an expected positive correlation (Wang et al. 2013). Whilst consumers had the construct of Trust chosen, using the 3-item instrument used by Park et al. (2017) and again an expected positive correlation.

Known Group Validity was also conducted; this involves the instruments’ ability to differentiate among groups, where the groups were expected to rate different to one another on specific traits or aspects (Netemeyer et al. 2003). The prior selected stakeholder categories of employees, customers, and students were likely to differ in their interpretation around the social responsibility of business. Known group validity can be analyzed by a t-test (Rosas and Ridings 2016); however, with three stakeholder groups, a one-way ANOVA was applied on a composite score.

Additionally, Corporate Citizenship measurements are unusually susceptible to the influence of social desirability, as highlighted by Hur et al. (2018), but also with Podsakoff and Organ (1986), suggesting social desirability can be a leading cause for shared method variance, which can then lead to common method bias. Thus, a test of social desirable responding was incorporated, with the Balanced Inventory of Desirable Responding 16-item (BIDR) being chosen (Hart et al. 2015), which composes of two factors which are Self Deceptive Enhancement (SDI) and Impression Management (IM). The test for social desirability is conducted by correlating perceived Corporate Citizenship with the BIDR, with non-significant results indicating no or little social desirability bias (Gupta and Agrawal 2018). This is important as participants need to answer from their personal perceptions and not answer how they think they should answer, i.e., socially desirable answers.

The sample consisted of 626 participants with evaluations based upon 234 students, 196 customers and, 196 employees. What became evident was that the overall performance was greatly improved by merging the 4th and 5th factors from the EFA. Table 2 shows the results from the CCA; in bold are those loading values that fell below the 0.707 thresholds. However, Hulland (1999) suggests that loadings above 0.600 can be acceptable within exploratory studies. Construct reliability and validity were within the acceptable values of Cronbach’s alpha (α) > 0.700, rho_A > 0.700, Composite Reliability (CR) > 0.700 and Average Variance Extracted (AVE) > 0.500. HTMT values were within acceptable values, with two values (in bold) being borderline as there is ambiguity about the cut-off point; Kline (2011) suggests 0.85, and Gold et al. (2001) suggest 0.90. The factors were named based upon literature and discussion between the authors; Products & Services, Environmental Oriented, Legal & Ethical Responsibilities, and Philanthropic Oriented.

Results for nomological validity have a positive and significant standardized Pearson Correlations between Perceived Corporate Citizenship and all three constructs; for Firm Attractiveness and students [n = 234] correlation of 0.546, for Organizational Trust and customer [n = 193] correlation of 0.581 and Organizational Commitment and employee [n = 196] correlation of 0.594. The customer sample group was reduced from 196 to 193; this was because three participants did not rate the scales that were to be used for the nomological test. The measurement instrument was acting as expected and can be considered nomologically valid.

The BIDR was also correlated with each group to test for socially desirable responses. For students, both factors of the BIDR had non-significant correlations suggesting the respondents were not answering in a socially desirable manner (SDE − 0.044 and IM 0.023). For customers, a small yet significant correlation was found on both components of BIDR (SDE 0.236 and IM 0.219), meaning that there is some socially desirable responding. Employees had a non-significant relationship between BIDR self-deceptive enhancement (0.131) but did have a small significant relationship with BIDR impression management (0.198). Whilst the results from testing for social desirable responses demonstrated some small correlations, which may suggest that there is a risk of shared method variance.

Next, known group validity was conducted using one-way ANOVA to test for the expected differences among the selected stakeholder groups; results can be found in Table 3. Significant differences are presented between stakeholder groups which are to be expected and validated the instrument further.

The next step of the instrument development is to validate in a wider context to devise how generalizable the instrument is.

Table 2 Study 2 confirmatory composite analysis results
Full size table
Table 3 Study 2 ANOVA results
Full size table

Study 3

The third study generalized the measurement instrument within the Malaysian context by selecting three companies; a petrochemical company, a banking company, and a supermarket store. The purpose here was not only to perform a CCA but also to run known group validity between these companies as they are from different industries. The stakeholder group remained fixed as customers, as using customers was important as they are most likely to suffer from information asymmetry; thus, if the measurement instrument was appropriate for them, it should be appropriate for the other stakeholder groups. Again, this study attempted to build nomological validity by including the measure of Customer Loyalty which from the literature has a positive correlation with PCC, the instrument used was adapted from Martinez and Bosque (2013). A target of 300 participants was adopted for this study, with a final sample consisting of 294 participants with evaluations based upon 138 supermarkets, 111 banking and, 45 petrochemicals.

In Table 5, the results from the CCA are presented; in bold are those loading values which fell below the 0.707 thresholds. However, as previously discussed, these should not be disregarded immediately. PPC10, [Firm] provides a variety of products & services, was below the tentative 0.500 limits that were discussed by Hulland (1999) and was removed for this study. Yet, this may well have been expected as two selected industries, namely banking and petrochemical, offer limited products and services whilst supermarkets have a wider range. Construct reliability and validity, and HTMT values are all within acceptable values.

Proceeding the CCA, the other forms of validity were conducted. This includes nomological validity; PCC was correlated with Customer Loyalty and resulted in a significant positive correlation of 0.547, which was expected. The next validation was known group validity, which was done using an ANOVA test again to look at the difference between the industries selected. Results from this test can be seen in Table 4, where the ANOVA and Post-Hoc results are presented. Noticeably, although there are differences, none are significant. When further exploration was done by conducting ANOVA on each factor of the instrument (not presented due to space limitations), there were significant differences in the Legal and Ethical Responsibilities component. This presents an interesting discussion point, as were participants suggesting that the only difference between the PCC for the three companies were the legal and ethical frameworks? Or was it a case of information asymmetry, and the participants were unable to answer accurately on the other factors. However, some variation does offer validation for the instrument, yet further validation was still needed in a yet broader context.

Table 4 Study 3 ANOVA results
Full size table
Table 5 Study 3 confirmatory composite analysis results
Full size table

Study 4

The fourth study aimed to validate and test the generalizability of the measurement instrument to an international context. This was done in a similar way to study 3 but was conducted with a sample of participants within the UK. The UK was selected because it was a developed nation, where much of consumer and regulatory pressures for Corporate Citizenship originate in such countries, whilst Malaysia is still ‘developing’ although it is an upper-middle-income country. Malaysia and the UK have a long relationship that dates back to the colonial period and being members of the Commonwealth post-independence, with 50 to 60% of the population speaking English to some level. It was due to these ties that the UK was selected above other Western nations to validate the instrument within a different national context. Three comparable companies to those of study 3 were selected from British companies: one petrochemical, one banking, and one supermarket. The same stakeholder group of customers was selected, and the same measure for customer loyalty was included. A larger sample was targeted this time of 500 participants, with the final sample consisting of 551 participants with evaluations based upon 383 supermarkets, 98 banking and, 70 petrochemicals. Table 7 shows the results from the CCA; in bold are those loading values that fell below the 0.707 thresholds. However, as previously discussed, these should not be disregarded immediately, one item from Philanthropic was below this threshold, and all four items which had previously composed the fifth factor were also below the 0.707 thresholds; however, all were above the 0.500 thresholds that would be a cause for concern. Construct reliability and validity, and HTMT values are all within acceptable values, with the exception of one value for HTMT, which was borderline (in bold). This is due to some elasticity around an appropriate cut-off threshold within the literature.

Known group validity was conducted using ANOVA to test the difference between the industries selected; the results are presented in Table 6. These results demonstrate multiple significant differences among the industries within the post-hoc results, which would be expected as the companies have very different business practices. This demonstrates a stark difference to study 3’s results, where possible reasons were discussed; however, it is not within the scope of the study to delve deeper into why the differences between the results of these studies exist. Next nomological validity correlated PCC with Customer Loyalty resulting in a significant positive correlation of 0.446, further validating the instrument.

Table 6 Study 4 ANOVA results
Full size table
Table 7 Study 4 confirmatory composite analysis results
Full size table

The four studies have refined and developed a measurement instrument from the original 101 item conceptualization from the Homer (2021) paper. The multiple contexts and validation stages incorporated in the study have tested the psychometric properties of the developed instrument on a very broad basis. The measurement instruments’ performance varied within the different studies, however for exploratory testing of a newly developed instrument, the results were acceptable. The paper continues with the discussion next to scrutinize the theoretical underpinning of the instrument, as well as the application.

4 Discussion

This study constructed a measurement instrument based on an existing bottom-up conceptualization and extensively examined its psychometric qualities in a variety of contexts. The results provided a four-factor solution that had been statistically confirmed and validated across stakeholder groups, industries, and cross-national. The shift from the widely used instrumental, top-down approach adopted by CSR is quite evident, exemplified by the shift to Products & Services, which demonstrates what the company has to offer society, rather than what the company can take from society (i.e., profits). This is in line with theory, as CSR tends to be from the organizational, top-down perspective; however, Corporate Citizenship revolves around the co-creation of company behavior based on interactions (Hur and Kim 2017). Thus, although differentiating between CSR and Corporate Citizenship is highly debated and contested, this study would propose that whilst similar items can be used, the difference would depend upon who’s perspective the measurement items are intended for. Either top-down management and organizational perspective, which is instrumental for CSR, or whether it is bottom-up and formed around what the company can offer society for Corporate Citizenship.

4.1 Theoretical implications

The four-part model derived statistically from these studies includes the Environmental Oriented factor, Philanthropic Oriented factor, Products & Services factor, and Legal & Ethical Responsibilities factor; this appears to complement the widely used Carroll’s Pyramid of CSR (1991b); however, there are also key differences within this paper that adapts Carroll’s model and as such addresses many of the criticisms within the literature. Carroll has consistently defended the Pyramid of CSR conceptualization as recently as 2016 with ‘Carroll’s Pyramid of CSR: Taking another look’ with the original conceptualization being published in 1979 and thus now has been the dominant model of CSR for over 40 years. Some authors have attempted to modify Carroll’s successful model, such as Masoud (2017) with the International Pyramid of CSR and Baden (2016) with the pyramid of CSR for the 21st century. However, these attempts have been based in the conceptual and theoretical domains and thus have found it difficult to be accepted as Carroll’s model has been operationalized and has proven useful for several decades. This paper has added some empirical support to the necessity for Carroll’s model to be adopted to fulfill its role from different stakeholder perspectives and societal demands. Each component of the developed instrument is now described and compared to Carroll’s model in order to establish the instrument’s theoretical foundation and discuss similarities and differences.

Carroll’s conceptualization is consistent with the Philanthropic Oriented factor developed within this study, embracing similar items to prior measurement instruments. The factor covers how businesses give back to society, not just financially, but also via the use of their workforce and by setting an example for other businesses, as all good corporate citizens should. Carroll’s conceptualization of the CSR pyramid comprises the following: It is critical to conduct business in a way congruent to society’s altruistic and philanthropic expectations. It is further vital to assist the fine and performing arts. It is critical for managers and employees to engage in volunteer and charitable events in their local communities. It is critical to support private and public educational institutions, and it is critical to contribute to projects that improve the community’s “quality of life” Carroll 1991a, p. 41). This complements the new instrument declarations, which include the following: [Firm] adds value to the community, [Firm] is accountable for its performance, and [Firm] gives back to society.

The Products & Services factor is comparable to Carroll’s economic factor in that enterprises must offer products and/or services in order to generate profit, which is necessary for their own survival, but the perspective has shifted to a stakeholder looking into the firm from the outside, rather than the management looking out of the firm. However, one criticism leveled at Carroll’s economic component is that the assessment instruments produced from his model include questions such as ‘Attempts to maximize profit from its activity’ (Salmones et al. 2005). However, given the asymmetry of information, it is unclear whether a consumer would have adequate expertise to answer this or related inquiries about economic performance, which results in the “erratic, mindless responses or reactions characterized by social desirability” described by Alvarado-Herrera et al. (2017). However, by shifting the focus to the products and services offered, statements such as [Firm] ensures customer satisfaction with its products and services, as well as [Firm] provides a pleasant consumer experience become much more tangible to stakeholders who may be unaware of the business’s day-to-day operations. Thus, this change of perspective means that more stakeholders would have sufficient information to evaluate the statements to at least some extent.

The merging of the Legal and Ethical components within the developed instrument, when compared to Carroll’s model, is unsurprising, given there has been a lessening of the gap between the legal and ethical dimensions in the literature. This has been suggested to be because what is deemed ethical has become increasingly ingrained in legislation (Visser 2005). What constitutes ethical behavior varies by culture and ethnic origin (Hur and Kim 2017). There has been an increasing understanding that religion has an impact on corporate behavior, including business ethics (Du et al. 2016). Western ethics are influenced by philosophers such as Aristotle, whilst Eastern ethics are influenced by philosophers such as Confucius (Farooq et al. 2014). While these philosophers share some concepts and values, they also have disagreements about what constitutes ethical behavior. However, it appears from this research that the fundamental principles of ethical business practices are widely acknowledged in the two countries examined. There is room for an investigation into the extent to which similar ethical behaviors extend and are sought by stakeholders in various cultures worldwide.

The Environmental Oriented element devised in this study adds a new dimension to the CSR and Corporate Citizenship models, including Carroll’s pyramid of CSR. According to Moorthy et al. (2017), numerous studies employ Carroll’s discretionary [or philanthropic] responsibilities and then separately measure the environmental component. Thus, including this feature in the model is both necessary and opportune in order to address this critique of Carroll’s widely used model. While practitioners have embraced the ESG model (Environmental, Social, and Governance), these issues have often been overlooked in academia due to their qualitative character (Halme et al. 2018). This could be due to the absence of environmental and social considerations in academic models, i.e., Carroll. Thus, the practitioner perspective validates the inclusion of this component in the proposed measurement instrument for Perceived Corporate Citizenship.

The discussion thus far suggests that the instrument developed in this study complements existing models, most notably the Pyramid of CSR developed by Carroll, but also addresses several criticisms and provides a contemporary adaption on Carroll’s aging model. However, an element of Carroll’s model that is outside the scope of this study is the pyramid shape the model consists of. The shape of a pyramid comes from the weighting assigned by Carroll, namely 4 for economic, 3 for legal, 2 for ethical, and 1 for philanthropic, with each element stacked upon the prior (Pinkston & Carroll, 1996). However, without further empirical studies, the weightings for the elements derived in the present study of philanthropic oriented, products & services, legal & ethical and, environmental oriented, are unclear. The pyramid shape may well be lost as it is unlikely stakeholder groups such as customers and community members will place such emphasis upon the economic, i.e., products & services.

The conversation proceeds with an evaluation of the measurement instrument’s applicability.

4.2 Methodological implications

While it was possible to minimize and refine the number of statements in the measurement instrument statistically, it was determined that this was not necessary. This is because the measurement instrument is meant for use by a variety of stakeholder groups and thus includes a broader range of statements, which should mitigate the risk of information asymmetry. With a higher number of statements available, some customization can be accomplished by omitting statements that particular stakeholder groups may lack the necessary expertise to answer appropriately. Thus, the probability of “erratic, mindless answers or behaviors characterized by social desirability” is diminished (Alvarado-Herrera et al., 2017). This does mean that the instrument will need to be validated in additional scenarios, such as study 2, where established group validity is utilized to determine which stakeholder groups have access to which information.

Additionally, there is a chance that information asymmetry changes by industry for different clients. This is in response to participant comments that indicated they preferred to answer questions regarding retail companies because they have more experience interacting with them. This is evident within studies 3 and 4, in both Malaysia and the United Kingdom. This also facilitates the retention of a larger number of assertions, allowing the instrument to be adapted to certain sectors. While it was outside the scope of this research to examine the disparities between industries, it does highlight how one-size-fits-all measurement tools may not be the most appropriate method of measurement.

Furthermore, the usage of BIDR is critical for the future use of not only this instrument but maybe all instruments pertaining to social responsibility. This is because socially desirable responses are indicative of information asymmetry; when stakeholders lack sufficient information to respond objectively, they begin to respond how they believe they should respond. This is demonstrated in Study 2 when customers are likely to have the least knowledge or information available to them due to their occasional interactions with the organization, and hence socially desirable responses exhibited minor but substantial connections. This section of the discussion raises more issues than it answers and necessitates additional research to determine definitely whether this proposition holds true for other measurements.

Nomological Validity is defined as the degree to which predictions from a theoretical network containing the concept under scrutiny are confirmed (Netemeyer et al. 2003). This means that a measure is nomologically valid if it behaves as expected with respect to another construct to which it is theoretically related (Donia, Tetrault Sirsly, & Ronen, 2017). Within this study, nomological validity was used within studies 2, 3, and 4. Whilst within study 2, each stakeholder group evaluated a construct that had been nomologically linked to Corporate Citizenship within the literature; this composed just a small proportion of the extensive nomological network that exists for Corporate Citizenship. Likewise, within studies 3 and 4, only one stakeholder group was evaluated upon one nomologically linked construct. So, within this study, four different constructs have been confirmed for their nomological validity and have appeared to add complexity to the study, but this is just a small part of the vast nomological network that exists. The implication for methodology here is that to fully validate a measurement instrument, it needs to be used with an extensive range of different constructs to ensure the nomological network is complete and as was expected regarding the literature. This, however, may be why Carroll’s model has been so popular for measurement, as it has been used in numerous contexts with many relationships tested and thus nomologically verified, even with the criticisms that have been offered within this paper and others.

Known group validity involves the measurement scales’ ability to distinguish between groups of individuals who should score differently on particular traits (Netemeyer et al. 2003). Whilst the study attempted to use known group validity embedded within studies 2, 3, and 4, which appeared to add a lot of complexity to the study, this actually only consisted of only a small section of comparisons that could have been made. Within study 2, only three stakeholder groups were incorporated, which does not represent the entire stakeholder base of a firm, and even then, there may also be variation within stakeholder groups which could be compared along demographical lines or other aspects, i.e., large and small shareholders of a firm. Whilst in studies 3 and 4, only three industries were represented, and only one representative firm from the industry was selected for each country, thus again, although the study appears to incorporate great depth, this is a small subsection of the entire industrial representation globally. The implications for methodology from this study would be that whilst known group validity can be incorporated easily into many studies, perhaps it should not be used simply to look for a difference for validation but can be used strategically and holistically to construct a network similar to the nomological networks. For known group validity within the Corporate Citizenship and broader CSR context, all stakeholder groups (and sub-groups), industries, countries, and size of the organization are all likely to have variations on what is deemed [un]responsible business practice and thus could be compared against one another.

5 Conclusion

The findings of this study suggest a four-factor model consisting of Products & Services, Legal & Ethical Responsibilities, Environmentally Conscious, and Philanthropically Conscious. This created a measure of Perceived Corporate Citizenship that is both contemporary and relevant to stakeholders. This builds on previous work presented by Carroll’s Pyramid of CSR and other models and addresses criticisms leveled at preceding models and measures of Corporate Citizenship and CSR. Thus, the issue of Corporate Citizenship assessment tools being constructed from an organizational perspective and not taking into account the knowledge of the various stakeholder groups asked to respond is somewhat rectified. The purpose of this study was to build a measurement instrument to evaluate Perceived Corporate Citizenship from the bottom-up before generalizing it for wider use. As a result, the instrument is adequately adapted for stakeholders who do not have access to all relevant information. While the measurement instrument requires further refinement, as it has been used in only four studies on seven different organizations, this paper shows good findings that may assist capture the subjective character of Perceived Corporate Citizenship. Measurement instruments must be regularly revised and adapted to reflect the dynamic and ever-changing nature of Corporate Citizenship and the broader CSR phenomena, reinforcing CSR as a dynamic and ever-changing phenomenon (Verbeke and Tung 2013).

5.1 Limitations

The study has a main limitation based around the novel measurement instrument’s limited contextual use. Although four investigations were undertaken with an expanding contextual scope of application, the measurement instrument has not been deployed in a large number of situations, and so its generalizability cannot be assured.

5.2 Suggestions for future research

The critical component of the future study will be to continue to employ the measurement instrument in a variety of scenarios and to assess its performance. This also applies to the phenomenon’s nomological network; the instrument should be used in conjunction with theoretically or empirically linked structures to verify that the instrument operates as predicted. Academics should validate, modify, and adapt the instrument as necessary to ensure an accurate assessment of Perceived Corporate Citizenship.