Introduction

Greenwashing, as an antagonistic idea of the true meaning of corporate social responsibility (CSR), has become an emerging topic whose means and ends have captivated academics and professionals who seek to understand how it works and at the same time mitigate its spread (^{Lyon & Montgomery, 2015}). One of the many ways to do this could be through an analysis of the effect of its adoption on employees and their behaviors in the workplace (^{Donia & Tretault-Sirsly, 2016}). Mainly, organizational identification (OID) has been consolidated in the literature as a construct that allows the connection between a series of organizational sti-muli with some behavioral results at the individual level (^{Lee, Park & Koo, 2015}). Therefore, considering that this behavior has been widely analyzed in different manifestations at the organizational level, an evaluation of the way in which it could be affected by the adoption (or not) of practices that coincide with greenwashing could be pertinent. For the purpose of this work, it is determined by the perception of the authenticity (or not) of CSR.

In this way, on the one hand, it is clear that the theoretical conception of greenwashing, interpreted as an exercise of conveying untrue messages regarding CSR, can be a valid antecedent of adverse effects both for organizations and for societies in general (^{Lyon & Montgomery, 2015}). On the other hand, some organizational results such as more significant commitment and motivation, less absenteeism and higher job performance are in turn driven by adequate levels of employee OID (^{Jones & Volpe, 2011}). Nevertheless, further and more refined knowledge is still needed regarding the connection that may exist between these two concepts (^{Glavas & Godwin, 2013}). In particular, this need is tangible within contexts where aspects such as the environmental and social impact generated by the dynamics of organizations are critical to the construction of the future development of their societies (^{Jamali & Karam, 2016}; ^{Rodrigo, Duran & Arenas, 2016}).

Thus, understanding greenwashing as the difference between false and authentic actions related to CSR (^{Walker & Wan, 2012}), and OID as the self-perception of an employee expressed as an approximation of their sense of belonging to the organization of which they are a member (^{Jones & Volpe, 2011}), this work addresses the impact that the former can have on the latter. In other words, the present paper intends to identify how much the authentic or false nature of the CSR of an organization influences the sense of belonging of its employees. In this vein, the framework used integrates some contributions of the theory of organizational legitimacy (represented by the notion of greenwashing) with the foundation on CSR at the individual level, or micro CSR (represented by the OID).

Methodologically, an exploratory cross-sectional research is carried out to empirically evaluate a theoretical model in which the OID takes on the role of a dependent variable and interrelates with CSR. In turn, this latter variable is analyzed regarding its dual lens; this is through both substantial (practical and authentic) and symbolic (rhetorical or false) perceptions. Specifically, we develop a multivariate study covering a sample of two hundred and twenty professional employees with jobs of responsibility in Colombia. Thus, three hypotheses are tested using a model of structural equations based on exploratory and confirmatory factor analyses.

The structure of the manuscript consists of five additional chapters. In the first instance, a theoretical framework in which the constructs to analyzed are presented together with scientific antecedents where we demonstrated a more profound relevance of their theoretical association. Subsequently, there is address the methodology, where we develop the working hypotheses, the design of the instrument, and the respective structural equation model formulation. The fourth chapter includes the results obtained in its various stages. The fifth chapter contains the discussion generated from the respective interpretation of the results, their implications, and possible future lines of research. Finally, the paper closes with the conclusion reached.

Theoretical framework

Within the body of knowledge of organizational theory, there has been an exciting development with respect to the effects generated by the different ways in which CSR and its multiple associated concepts impact employees at the level of their perceptions and behaviors (such as motivation, commitment, satisfaction, intention to resign and identification, among others). This field of study has been called Micro Corporate Social Responsibility (Micro CSR), which is a derivation of the field of study of organizational behavior at the individual level (^{Rupp & Mallory, 2015}). In the same manner, organizational legitimacy has played an essential role within this discipline, and despite a series of fascinating contributions that shed light on this relation, there is a continuous warning of the existence of countless opportunities for the creation of more and better knowledge in this regard (^{Aguinis & Glavas, 2012}; ^{Delmas & Pekovic, 2013}; ^{Evans & Davis, 2014}; ^{Evans, Davis & Frink, 2011}; ^{Lamm, Tosti-Kharas & King, 2015}). The following theoretical framework is structured from the general to the particular, starting with CSR as a macro-field of study linked to the concept of legitimacy and ending with some research background that has somehow involved them with the OID. Finally, we will note the specific literature gap that this work intends to address.

Methodology

Some mentions founded concerning the involvement of the authenticity of CSR in its specific relation to individual employee behaviors and which call for further research development in this regard (^{Brammer, He & Mellahi, 2015}; ^{Glavas & Godwin, 2013}; ^{Rupp, Skarlicki & Shao, 2013}). Similarly, it is also stated that studies are needed to help examine the dynamics of stakeholder perceptions of the authentic and inauthentic nature of CSR. However, it is necessary to recognize that the work of ^{Donia et al. (2017)} has been the only work to make a concrete proposal to date on the subject. In fact, in their manuscripts, they explicitly invite the continuation of the particular line of research based on the empirical testing of their pro-posals, and the possibility of their study in new contexts.

Specifically, considering the impact of those concepts on organizational behavior, the theoretical propositions of these authors are two. First, the attributions of substantial CSR will be positively related to favorable behaviors on the part of employees; and Second, the attributions of symbolic CSR will have no-relation, or this relation will be negative with favorable behaviors on the part of employees (^{Donia et al., 2017}; ^{Donia & Tretault-Sirsly, 2016}). In this order of ideas, and taking into account the predictive character of modeling through structural equations developed in this study, it could be said that taking into account gender, age and profession as control variables, the first hypotheses that adjust to the problem posed here would be as follows:

H1: The attributions of substantial CSR are positively associated with the level of employee OID.

H2: The attributions of symbolic CSR are not associated, or their association is negative on the level of employee OID.

Likewise, the interaction between the independent variables of the model and the-what we could call, logical-the result of the null or negative interdependence between them found in ^{Donia et al. (2017)} is taken into consideration. In this manner, the aim is to validate that the concepts of symbolic CSR and substantial CSR are indeed contrary to, or at least independent of, each other, meaning that authentic CSR practices are incompatible with greenwashing practices. Consequently, this article tests the following hypothesis:

H3: The correlation between symbolic CSR and substantial CSR is negative or non-existent. Based on the above hypotheses, we applied a cross-sectional study in order to carry out the present investigation using the application of a self-administered physical questionnaire (applied in September 2016). Later, we analyzed the data collected using structural equation modeling (^{Wright, 1921}) carried out in R language version 3.2.2 (a free and open-source lan-guage), which allows evaluating the relationship between one or more independent variables.

The Structural Equation Model (SEM) is a statistical methodology that departs from the proposal of the theoretical interaction between variables from a path diagram, finally determining the existence of this relationship through its robustness and eventual level of significance. Specifically, the SEM uses a confirmatory factorial analysis or multiple linear regression approach, which involves determining the linear relation between latent (e.g., theoretical constructs) and observable variables (e.g., each component or item of a theoretical construct), in conjunction with dependency or covariance relations among themselves (^{Byrne, 1998}). The method starts from the development of a measurement model in order to identify the observable variables that affect the latent variables; in such a way that-and after cleansing-it is possible to proceed with the determination of their interaction and the subsequent obtaining of results (^{Manzanares & Pérez, 2015}). It is worth highlighting that the main reasons for justifying the selection of the methodology used are the capacity of the SEM to support a possible non-normality in the data. As well, the methodology can estimate through a principal components analysis loads of model (always guaranteeing the same load value), where there are as many latent variables as observed. (^{Kline, 1994}).

Sample and procedure

The corresponding instrument originally was applied to two hundred and seventy-six professionals working in Colombia, particularly in the cities of Bogotá, Bucaramanga, Cúcuta, and Barrancabermeja. The data collection was physically useful, taking advantage of the fact that these professionals were also participants in different graduate programs in management areas of the School of Industrial and Business Studies of the Industrial University of Santander. Once the surveys were completed and collected; we carry out a primary data cleaning, eliminating forty-three answers due that they were incomplete or belonged to individuals who had no direct employment relationship with any company (i.e., many of them are independent professionals and (or) contractors), and the respective foliation of the others was carried out. Subsequently, after the data tabulation, we conduct a second cleansing with the elimination of three surveys that presented inconsistent information (such as more than one answer in the same question or a high number of blank answers).

Finally, we consolidated a homogeneous sample of two hundred and twenty professionals (they work in different types of companies with diverse responsibility roles). All of them were in the preparation stage for the management exercise. Therefore, taking into account their own academic or experiential perspective, it is assumed that they have at least a basic knowledge of issues related to organizational dynamics such as CSR and human talent management. A brief description of the sample is a group of young adults with a marked engineering profile, where the majority are under 32 years of age (64.5%), without dissimilarly among genders. Figure 1 shows the main characteristics of the sample based on gender, age, and profession; variables of interest for the formation of socio-demographic profiles.

Source: Own elaboration.

Figure 1 Distribution of the main control variables 

Results

We have done the EFA choosing three factors using the ^{Kaiser criterion (1960)} through the principal components method, without finding a significant statistical difference between the control variables, thus explaining 60.7% of the data variance. From the original dataset, questions 2.2.3, 2.2.6, and 3.6 (two questions related to symbolic CSR and one to OID) are not considered because their loads did not fit the proposed three-factor model (their value was not high enough: to influence the assigned factor), decision supported by a confirmatory factor analysis or structural measurement model. Table 3 shows the relations between observed variables and constructs (loads). As a highlight, we noted that depending on the value of the loads, factor 2 (OID) presents negative but high values (absolute value greater than 0.5).

Later, we evaluated the proposed structural equation through the sem and lavaan statistical packages of the R software. Getting a high statistical significance (goodness of fit index of 0.929), thereupon, the organizational identification depends on the substantial perception of CSR developed in the companies (the value of the slope of the Substantial CSR is 8.805). In addition, although the true and false CSR perceived by professionals varies (the p-value of the influence of Symbolic CSR is significant, and there is a correlation between the variables in Annex 7.1), there is not enough statistical information to affirm a significant degree of relation (neither positive or negative) between the latter and OID. In the same way, there is a direct and statistically significant correlation between substantial CSR and symbolic CSR (estimated value of the correlation between the constructs is 0.178), which contradicts H3. Table 4 summarizes the main results.

Similarly, the model represented in Figure 2 presents a graphic approximation of the findings.

Source: Own elaboration.

Figure 2 Representation of the prooposed structural model 

Table 4 records the fitness indices or parameters of the corresponding model, finding a value that falls in the lower setting range only in the Chi-square test; however, this behavior may be due to the fact that said test is sensitive to the sample size, therefore, when increasing the latter, the probability of rejection of the fitness hypothesis is increased (^{Bagozzi & Yi, 1988}). Therefore, other estimation indices are analyzed (considering degrees of freedom and complexity of the model) that show an adequate adjustment of the proposed model. Besides, we calculate the Root Mean Square Error of Approximation (RMSEA), with the interpretation as the mean approximation error per degree of freedom. Therefore, we conclude a proper fit of the model.

Discussion

Previous studies have focused on the relation between CSR (and its different manifestations) with different employee work behaviors. Predominately, they have found some positive relations between the constructs analyzed. In fact, some scholars consolidated the idea of a positive effect generated by CSR on OID in the work environment (^{Glavas & Godwin, 2013}; ^{Jones & Rupp, 2017}). The literature has been commissioned, then, to present sufficient evidence to affirm that CSR practices effectively improve the way an employee identifies with their organization. However, these nuances have not taken into account the search for the legitimacy of such organizational practices. This study was concerned with the latter questioning, in such a way that it assessed both the authentic (i.e., substantial) and false (i.e., symbolic) nature of CSR and the effect they have on the OIDs of the employees.

We obtained results varied in nature. On the one hand, there was a strong and significant positive dependence between substantial CSR and OID, as well as a causal disconnection between symbolic CSR and OID (supporting hypotheses 1 and 2). On the other hand, we found a weak positive correlation between substantial CSR and symbolic CSR (rejecting hypothesis 3). The analysis has two fundamental assertions. The first is the initial ratification of two of the proposals to the theoretical model proposed by ^{Donia & Tretault-Sirsly (2016)}, which later served as input to the validation and operationalization of constructs (^{Donia et al., 2017}). The second is an unexpected contradiction observed, with two opposite concepts having a direct correlation, especially when in the initial validation tests of these authors, the correlation found between them was- logically in appearance-of an inverse or null nature. As for the former, it coincides with the theory. Therefore, it could be thought that an employee is indeed capable of perceiving and processing signals derived from CSR practices emitted by their organization and differentiating them between authentic (substantial) and false (symbolic) practices, as well as independently appropriating them for shaping or not shaping their behaviors (in this case, their OID). About the later, we observed a sort of consent to the organizational adoption of false CSR practices by employees in a context of self-definition as members of an organization.

In consequence, the theoretical and practical implications of the study are diverse. In terms of theory, we made progress in the knowledge of both the field of study of micro CSR and the conceptual aspects surrounding the theory of legitimacy over employees, especially when this legitimacy is defended through deception (which is part of greenwashing). Additionally, according to the knowledge and understanding of the authors, this paper is the first empirical partial validation of a recently proposed theoretical model. The fact that the research proposed by ^{Donia & Tretault-Sirsly (2016)} was carried out in the context of a Latin American country (with non-conventional results partly different to the original theoretical proposal), opens the door to identifying nuances in the formulated theory that might make one think of context as an explanatory element of the results (^{Jamali & Karam, 2016}; ^{Rodrigo et al., 2016}). However, it is essential to acknowledge that the latter requires further research to arrive at more definite conclusions.

From a practical standpoint, it is possible to capitalize on these results in the light of organizational decisions related to CSR, together with their dissemination to different sectors of society. The evidence obtained suggests that employees appear to be no strangers to these signals and can process them to shape their behaviors (perceptions, intentions, and behaviors) in the work environment. Therefore, if the decision-makers can manipulate the genuine concept of CSR, the consequences may not be so convenient for their workforce and, therefore, have repercussions in terms of the sustainability of society in general, and organizations in particular.

One of the apparent limitations of this study relates to the almost exclusive focus of the instrument used on the social issues of CSR, which could leave aside the environmental aspects of CSR. The preceding, (despite to the fact that we no included the contributions obtained), is previously justified by ^{Donia et al. (2016)}, since they explicitly propose the equivalence of social consideration with environmental protection on the part of organizations, in light of the perceptions of their employees. However, an exercise that could project more specific research fronts would be the clear delineation of the social, environmental, and even economic dimensions of CSR. Likewise, as previously done by ^{Amato, Busraschi & Peretti (2016)}, the fact that we differently approached CSR and sustainability is something that could add more value to the discussion in the present study, given that the latter concept is sometimes perceived as more critical than CSR itself.

Likewise, another point of discussion relates with the cross-sectional nature of our data, which would not allow affirming to suggest causal relations. However, we mitigated it with the use of structural equation modeling as well as strong theoretical support that in part allows predicting such relations in the conceptual model proposed. In the same way, we recognized that it is impossible to guarantee the generalization (or external validity) of the results obtained, taking into account the characteristics of the sample. Nevertheless, the balance generated by the use of similar samples in previous studies (^{Brammer et al., 2015}; ^{Carmeli et al., 2007}; ^{Kim et al., 2010}) together with the homogeneity in the profile of the members of the sample examined here, are elements that provide some reassurance in order to reduce potential biases in the study.

Despite the above, some controversy may also arise regarding the context in which it was presented. In this sense, having collected data in Colombia may make the results relevant in different ways in other cultural contexts. However, if that is the case, in addition to contributing to a better understanding of the specific process in that country, this study also opens the door to future contrast and comparison of its results in other contexts-similar or not-in order to produce a more precise and comprehensive knowledge on the subject. Both CSR and the nature of the individuals who make up the organizations are concepts with territorial nuances, and their analysis cannot ignore those variables that separate them from the rest. For example, Colombia, like many other Latin American countries, is a biodiverse and developing country. However, it presents profound social problems that make the results of this type of study a real challenge in terms of its interpretation and implementation. In this sense, corporate deception about CSR practices is something that should not be tolerated in these types of regions, as the implications of doing so can be costly for their present and future well-being.

In addition to the refinement of the above limitations, future research could focus on the validation of the relations between substantial CSR and symbolic CSR with other behavior constructs specific to the field of organizational behavior at the individual level, including, if deemed appropriate, OID through its mediatory character. Likewise, the possibility of involving moderating variables and (or) more mediating variables is an opportunity that will always be present to achieve more progress in the knowledge gathered. It also opens the door for qualitative research projects, which could provide more profound clues that add to the definition of greenwashing in a context similar to those typically studied in the literature. In this sense, case studies, content analysis, or ethnographic work in emerging countries could represent exceptional examples that contribute even more to the comprehensive understanding of employee behavior in organizational environments.

Conclusions

The evidence gathered in this study suggests that employees may perceive substantial CSR practices positively and be indifferent to symbolic CSR practices. We observed this evidence in the evaluation of their effect on the identification perceived by them as members of their organizations. However, considering previous studies, it could also be inferred that other behavioral outcomes such as job satisfaction, commitment, and even effectiveness may depend on such perceptions.

On the other hand, a disturbing finding that challenges the initial assumptions of the theoretical propositions used (^{Donia & Tretault-Sirsly, 2016}) is that in some ways these employees also vindicate the use of symbols, ceremonies, and formalisms to legitimize the (often authentic) CSR practices of their organizations. In other words, there is a kind of validation of the positive, aspirational discourse (^{Christensen et al., 2013}) where it would seem that there would not be much controversy. It must be accepted that organizations have the right to make use of various mechanisms to disseminate the fulfillment of a purpose-oriented to CSR when it exercises a genuine presence in organizational aspects (policies, processes, operations, and goals, among others), in addition to making use of them to reinforce their positions and define objectives (^{Berrone, 2016}).

Nevertheless, the results could also be a clear sign of concern where the concept of CSR could have alterations according to past academic or work experience. Therefore, it is cause for concern that some empty practices such as the validation of discourse without action, environmental certifications seeking only the fulfillment of a requirement, donations see-king tax benefits or even claiming having contributed to the development of society through compliance with the law, are recognized as CSR. They all represent forms of greenwashing through the use of subtle mechanisms that attempt to create an altered image of reality (^{Mahoney et al., 2013}; ^{Milne & Patten, 2002}). The fact is that this eventual situation would imply an enormous challenge at the educational level to redirect efforts to make clear the real definition of CSR (^{Berrone, 2016}). Moreover, while the latter is a comprehensive topic at all levels and for all areas of knowledge, it is particularly sensitive at the management training level, as the participants are the ones who, subsequently make decisions in the organizations.