Factorial Validation of a Corporate Social Responsibility Perception Scale for Micro, Small, and Medium Enterprises

: The aim of this study is obtaining a valid and reliable scale to evaluate the Corporate Social Responsibility perception level (pCSR) of micro, small, and medium-sized Mexican enterprises. A literature review revealed 58 items among 11 dimensions: Ethics, performance, business model changes, quality of working life, the environment, community outreach, marketing, philanthropy, competitiveness, and the interests of stakeholders. A total of 296 companies located in Mexico were considered for the study. Principal axis factoring with varimax rotation was performed with the Exploratory Factorial Analysis Technique (EFA) to give an interpretation to the pCSR in the Mexican context. The factors found are the environment and social commitment, performance, ethics, change in business model, restrictions, and competitiveness. The outcomes show that the scale is valid and reliable for use in evaluating pCSR in micro, small, and medium-sized enterprises. This scale is a tool that allows us to evaluate the social responsibility of micro, small, and medium-sized enterprises, we conclude that considering the characteristics of Latin America, we have provided an original scale to measure pCSR.


Introduction
Corporate Social Responsibility (CSR) is a methodology that organizations are incorporating into their businesses for philanthropic reasons due to pressure from stakeholders or for convenience. They adopt CSR as a strategy to improve competitiveness, to become more sustainable, and to improve their quality of life (FUNDES, 2005:26). Both at national and international levels, the ethical behavior of companies, their commitment to the community, the preservation of the environment, and the quality of life of their human capital, are aspects observed by governments, organizations, associations, and researchers. Companies making a product, retailing, or offering a quality service must also commit to aspects such as social marketing, economic retribution to their human capital, community outreach, paying taxes, revenue generation, job creation, environmental training, health, and job security. Among companies' CSR actions are their commitment to develop programs that have a sustainable impact, as much within their internal administration as on the communities in which they operate (CEMEFI, 2008), considering the demands of their different stakeholders (customers, suppliers, workers, community, among others). MSMEs boost innovation, entrepreneurial initiatives, and competitiveness (Spence et al., 2000;Enderle, 2004;Sweeney, 2007;Fisher et al., 2009). However, many believe that small and medium-sized businesses are not interested in being socially responsible, or they do not have the capacity to do so (Castka et al., 2004;Shoenberger-Orgard, 2005;Jenkins, 2004). This could not be further from the truth, as these kinds of companies have flexible operations, rapid communications channels (Margolis, 2001), their business value chain (Argandoña, 2008), and their local involvement. All of this makes these types of companies aware of all their overriding problems.
Enterprises, organizations, and researchers around the world are concerned about the RSC study; it is also true that most Mexican research is based mainly on: ISO26000 (2007), Social Accountability Standard 8000 (Accountability, 2004), the Global Reporting Initiative (GRI, 2009), and the Mexican Norm on Responsibility (IMNC, 2004). However, most of them also take into account certain dimensions from other models. García de la Torre, Portales, and Camacho & Arandia (2010:98) comment that with the GRI, which includes three elements from TBL, all the others are focused on one or two elements from the Elkington Model. This model called the Triple Bottom Line (TBL) includes economic, social, and environmental dimensions. There are few works about developing constructs about local business in the national and cultural environment, especially in SMEs (Pérez y Veloz, 2007;IMNC, 2004). It is important to know the CSR perception level of the MSMEs. Therefore, this work presents the validation of a scale for Mexican enterprises, using Exploratory Factor Analysis (EFA) with extraction factors of the principal components using the Varimax method of orthogonal rotation. This research is considered relevant due to the lack of similar studies on this subject in Mexico, and the need to design specific tools following the ideologies of developing countries. This paper offers two contributions. The first one is to provide a customized scale to evaluate the level of social responsibility on Mexican SMEs. The second one provides SMEs with feedback on their own business as a guide on how to become a socially responsible enterprise. This research has been structured into five sections. The first one, the introduction, mentions the importance of CSR in SMEs and the research objectives. The following section is a review of the literature, substantiating the importance of the study of CSR and validation through Exploratory Factor Analysis. The third section is concerned with the methodology, describing the original scale. In the fourth section the analysis and results obtained are shown, using the Exploratory Factor Analysis (EFA), with the extraction factors of the principal components, and using the Varimax method of orthogonal rotation.
Tools have been developed in Mexico to diagnose CSR in organizations such as the Centro Mexicano para la Filantropía A.C. [Mexican Center for Philanthropy, registered charity in Mexico] (CEMEFI, 2014), which produced a survey that is applied to companies that seek to obtain the ESR® certification. The fundamental areas of this survey are: Quality of life in the company, corporate ethics, community outreach, and care and preservation of the environment. Another organization is the Fundación delEmpresarioChihuahuense, A.C.
[The Entrepreneur's Association of Chihuahua, registered charity in Mexico] (FECHAC, 2014), which evaluates four indicators: Quality of life in the company, community outreach, care and preservation of the environment, competitiveness of the company, and the relationship with its stakeholders; this instrument is entirely web based. The agency responsible (Aguilar, 2013), provides statistical data on the CSR of large and MSMES, which is centered on five axes: Reach and structure; implementation; communication; benefits and profitability; and consumer perception. Also, the Mexican Norm in Social Responsibility (IMNC, 2004), which is based on seven dimensions: 1) Ethical values, 2) mutual benefit between the interested parties, 3) leadership,4) personal involvement, 5) process approach, 6)system approach to management, and 7) continuous improvement. Several researchers, such as Husted & Salazar (2005), contributed an instrument to analyze factors that contribute to companies' motivation to participate in social projects: Profitability and relationships; human capital; social responsibility; competition; and social inclusion. Mercado & García (2007) evaluated the following variables: Business ethics, environmental preservation, quality of working life, and community outreach.

Initial pCSR Scale:
The majority of the aforementioned tools evaluate the environment, quality of life, ethics, competitiveness, profitability, and community. However, despite the range of dimensions that each survey involves, we consider it necessary to look into the indicators in order to determine the companies' CSR perception levels. Therefore, the objectives of this research are: 1) To develop a scale to measure pCSR, 2) to determine the dimensions of the scale, 3) to evaluate the reliability and validity of the scale. Taking into consideration aspects such as ethics, philanthropy, labor conditions, the environment, commitment to the community, social marketing, restrictions, performance, changes in the management model, obtaining benefits, and the interests of stakeholders, we have the goal of obtaining a tool that measures the companies' socially responsible behavior. In order to determine the dimensions of the new scale proposed in this study, a literature study was first carried out. Eleven dimensions were found: Ethics, performance, changes in the business model, restrictions, quality of working life, the environment, community outreach, marketing, philanthropy, competitiveness, and the interests of the stakeholders (see table 1).  (2001); Creighton (1999); Riera (1996); Van Ginkel (1996); Meyerson & Massy (1995).  Richardson et al. (1999); Baram (1984); Zenisek (1979

Methodology
With the aim of obtaining a valid and reliable scale to evaluate the pCSR of micro, small, and medium-sized Mexican enterprises, the Exploratory Factorial Analysis Technique (EFA) was applied, with principal components factor extraction, using the Varimax method of orthogonal rotation. Before this analysis we made a content validation by an expert's panel to ensure the content of each item in the scale. After the EFA, we also run a Reliability analysis: Cronbach's Alpha, Compound Reliability, and Analysis of Variance by Extraction, for the extraction scales were the measures used. Results from these measures are presented later. The quantitative tool was developed based on the Likert five points scale, with answers ranging from completely disagree(1), to completely agree (5). The survey asked 58 items based on the 11 dimensions (see appendix 1).

Technical information about the research:
The sample comprised 296 micro, small, and medium-sized enterprises (MSMEs) located in Celaya, Guanajuato, Mexico. There were considered to be companies with at least five employees and an sufficient number of them to determine its social commitment (see table 2).In order to classify the MSMEs, we used Mexican criteria for this (DOF 2009).Regarding the number of employees (micro from 0 -10, in the commercial, industrial, and service sectors; small from 11 to 50 in the industrial and service sectors, and from 11 to 30 in the commercial sector; medium from 51 to 250 employees in the industrial sector, 51-100 in the service sector, and 31 to 100 in the commercial sector).

Analysis of the pCSR scale and Results
Content Validity: Once the 58 items of the initial pCSR tool had been determined, it was reviewed by a panel of experts (businessmen and researchers) with the aim of validating its content. Once corrections had been made to the pCSR, we proceeded to apply it in person to the executives and owners of MSMEs. A global Cronbach's Alfaof = 0.895 was obtained, which is adequate according to Werts et al. (1974) and Nunally (1978) (considering an acceptable level equivalent to 0.700). However, on analyzing the results by dimension, we observe (see table 3 Exploratory Factor Analysis: With the purpose of obtaining a robust tool, we proceeded to analyze the data using Exploratory Factorial Analysis (EFA; Reise et al., 2000). In order to ensure the practical meaning, an absolute value of factorial load was considered to be at least 0.600, considering that this explains an adequate proportion of the variance, and is not elevated (0.800) or abnormal (Hair et al., 2007). At the first attempt, 11 dimensions were considered in the initial tool, finding that the rotation coincides with 12 interactions of 11 extracted components. Of all the variables suggested, 26 were ruled out (see Attending to the demands of their clients 0.526 20 FIL1 Self-motivation in CSR behavior 0.646 21 FIL3 Supporting social aspects that the government cannot resolve 0.544 22 COMP3 Improving job satisfaction 0.647 23 COMP 4 CSR bringing more benefits than costs 0.551 24 PRESGI2 Regulation of CSR behavior 0.585 25 PRESGI3 CSR behavior having reduced fiscal responsibilities 0.563 26 MODEG1 Impact on production processes 0.534 Source: Own elaboration based on an extraction method: Principal Component Analysis. Regarding the explained variance, the process extracts 22 factors in 11 components, explaining 58.427% of the original variation (100%). With the PEARSON correlation matrix, the items were contrasted and confirmed with the Kaiser-Meyer-Olkin coefficient (Kaiser 1970;Cerny and Kiser, 1977), considered a high value (Kaiser 1970(Kaiser , 1974. Bartlett's sphericity test (1950) contrasted the existence of a significant correlation between variables 2 = 7056.929, < 0.01 at a critical level (significance) of 0.000. However, the data obtained indicate that 4 of the proposed factors to determine pCSR are removed by EFA: Marketing, philanthropy, pressure from stakeholders and community outreach; this last one had 3 items (of which 5 comprised this dimension) with values below 0.580. The reliability analysis of the tool, observed in table 5, is = 0.833, and although it is above 0.700 (recommended by Nunnally, 1978;Churchill, 1979), the dimension of quality of work life obtains = 0.665 representing a weak reliability according to George & Mallery (1995). Therefore, we proceeded to carry out a second try based on 6 dimensions, with an absolute value of 0.600, obtaining the following results: The rotation converged to 7 interactions in 6 extracted components. The results indicate that 6 components are capable of reproducing 58.466% of the original variability (100% of the variance). In relation to communalities, 6 variables were deleted (DES8, REST5, ETIC4, LAB4, LAB5 and FIL2), as shown in table 6. Having deleted the items with values less than 0.600, we applied exploratory factorial analysis to the remaining 26 items, with the same data, obtaining a rotation convergence in 6 interactions and 6 extracted components. The results indicate that 6 components are capable of explaining 63.258% of the original variance (100%). Contrasting the factors, the Kaiser-Meyer-Olkin (KMO, Kaiser 1970, Cerny & Kiser, 1977, determined the sample adequacy measurement, = 0.841 > 0.600 which is considered to be a high value (Kaiser 1970(Kaiser , 1974. Regarding the null hypothesis (Bartlett, 1950), a significant correlation was found between variables 2 = 3049.930, < 0.01 at a critical (significant) level of 0.000. Therefore, the factorial model is adequate to explain the data.

Reliability Analysis: Cronbach's Alpha, Compound Reliability, and Analysis of Variance by Extraction-
In order to determine the internal composition of the tool, a Cronbach's Alpha test was applied to determine the compound reliability and analysis of variance via extraction (AVE). As can be seen in table 7, the internal composition of all dimensions above > 0.700are considered an adequate alpha value, followed by compound reliability above > 0.841, an acceptable value considered to be 0.700 (Werts et al., 1974;Nunally, 1978). Finally, the convergent validity was tested using variance via extraction. All dimensions obtained values above 0.594 (Fornell & Larcker, 1981) recommend an > 0.594. Therefore, the factorial model is adequate for explaining the data.

Conclusion and Recommendations
It is evident that CSR is an important topic worldwide for individuals, businesses, and countries, all concerned about the repercussions of their actions. This research validates a scale that measures the social responsibility perception of MSMEs in a Latin American country, such as Mexico. A robust scale using 26 indicators was developed to measure the pCSR. According to the hypotheses stated in this work, the first hypothesis "H1 was rejected. The level of the perception of corporate social responsibility (pCSR) in micro, small and medium-sized Mexican businesses (MSMEs) is determined by ethics, performance, changes in the business model, restrictions, quality of work life, the environment, community outreach, marketing, philanthropy, competitiveness, and pressure from stakeholders", as the pCSR obtained shows only six dimensions: The environment and social commitment, performance, ethics, changes in the business model, restrictions, and competitiveness, accepting the Ha hypothesis and rejecting H0. A significant aspect is that the first scale we proposed included the environment as a separate factor from social commitment, and on the new scale these two factors are merged. Therefore, the AFE allowed a pCSR scale to be established that is adequate to explain the data.
Most of the scales are translated to another country in a simple base and they are applied as if they were equal in both countries contexts. In this case we search items according to the literature and integrated them on a scale that was tested in a local context. One limitation of the study was the sample size, which although statistically adequate, the results may have been different if the study were conducted in other states or at a national level, for which one suggested future line of research is to apply the tool to a sample of the entire country. We also suggest that in addition confirmation analysis be performed in order to obtain a structural model of CSR. It is recommended that in the second part of this study a multiple regression analysis be carried out in order to analyze the influence of the factors obtained in the AFE on the level of CSR perception, which may allow the validation from the econometric point of view as a model of the determinants of the level of perception of CSR in a sample of MSMEs for an emerging economy, such as Mexico. Considering that the proposed scale is a tool that allows us to evaluate the social responsibility of micro, small, and medium-sized enterprises, we conclude that considering the idiosyncrasies of Latin America, we have provided an original scale to measure pCSR.