Sustainable Purchasing in the Retail Industry: A Model of Environmentally Responsible Purchasing
The business community has played an important role in achieving sustainable development over the last few decades (Kalafatis et al., 1999; Laroche et al., 2001; Han et al., 2009). As corporations are under growing pressure from internal and external stak
- PDF / 433,291 Bytes
- 4 Pages / 595.276 x 790.866 pts Page_size
- 74 Downloads / 226 Views
Trevino, 1986). Personal groups or role sets such as peers and a superior such as a top manager are the organizational source factors that may change beliefs to influence attitudes and eventually environmentally responsible purchasing. Due to these organizational factors, while idealism is stable, relativism is more likely to change. RESEARCH METHODOLOGY This study was conducted in the Swedish retail industry. As a developed country with a high standard of living, Sweden has high rates of consumption, supported by a growing number of retailers. The ethics position questionnaire (EPQ) (Forsyth, 1980) was the basis of measuring idealism and relativism in this research. In this study, attitude toward ethics and social responsibility was measured by the existing perceived role of ethics and social responsibility (PRESOR) scale (Singhapakdi et al., 1996; Etheredge, 1999). Moreover, to measure the actual environmental purchasing of professionals, based on social screening criteria used by socially responsible investing firms, the previously developed measure by Park and Stoel (2005) were used. Ultimately, a five-page questionnaire was sent to 1,532 purchasing professionals in the Swedish food retail industry. After two rounds of data collection, a total number of 573 usable questionnaires were returned providing a response rate of 37.4%. DATA ANALYSIS AND RESULTS To identify and test the underlying structure of the scales and dimensions of the constructs, an exploratory factor analysis (EFA) (principal components) with Varimax rotation was employed. The minimum loading criterion of .4 (Nunnally & Bernstein, 1994) and eigenvalue of 1.0 (Hair et al., 2010) were considered while conducting EFAs. Then, CFA using maximum likelihood estimation was conducted on all latent variables with structural equation modeling (SEM) using AMOS 20 to validate the factors proposed by EFA. Once the scales were developed and unidimensionality established, reliability was assessed using coefficient (Cronbach’s) alpha all exceeded .7 (Hair et al., 2007). Moreover, all AVE values were greater than the squared correlations, indicating discriminant validity for all of the latent constructs (Hair et al., 2010). Finally, the construct reliabilities as another indicator of convergent validity were all above the limit of .7, which suggests very high reliability (Hair et al., 2010). Together, these results demonstrate that the scale had sufficient levels of reliability and unidimensionality, the items adequately reflected their corresponding constructs, and all the scales exhibited convergent validity. To be able to test the hypotheses, the final step of the analysis process was to evaluate the structural model and its validity, having all the items evaluated in the measurement models. The results of the analysis demonstrate that the structural model fits the data well (see Table 1). Since evidence exists of adequate model fit and satisfactory reliability and validity, it is suitable to proceed with testing the hypotheses using SEM. The findings of the
Data Loading...
