Validation of a scale to measure consumers’ perceptions of clothing retailers' corporate social responsibility (CSR) in South Africa

Abstract

The clothing and textile industry plays an enormous role in contributing toward environmental and social issues (Sweeny, 2015). In order to address these issues and conserving our planet, future generations, stakeholders and consumers, should alter their behaviours to incorporate more sustainable practices. Stakeholders in the clothing and textile industry should implement more sustainable manufacturing and production methods and through that enable and encourage consumers to make better lifestyle choices when it comes to clothing. One way of implementing more sustainable practices is through Corporate Social Responsibility (CSR), defined as a voluntarily contribution by firms to create a better society and a cleaner environment (Fatma et al., 2016). It does not only improve a company’s image, but is said to have a positive impact on consumers’ perceptions towards a company, thus influencing loyalty, purchase intentions and potentially even their behaviours. To date, little research has been done on consumers’ perceptions of clothing retailers’ corporate social responsibility (CPCRCSR) in emerging market contexts. Thus, more comprehensive studies relating to the aforementioned topic are crucial to enhance and contribute to the research domain. Before such research can be conducted, it would require appropriate measurement scales to measure perceptions that consumers have regarding clothing retailers’ CSR to ultimately provide reliable and valid results. Therefore, this research focused on the validation of a scale to measure consumers’ perceptions of clothing retailers’ CSR in South Africa.

Existing literature states that CSR’s dimensions are in line with the triple bottom line’s (TBL) dimensions, namely social issues (people), economic issues (profit) and environmental issues (planet) (Wilson, 2015). Furthermore, Öberseder et al. (2014) sub-divided CSR into seven domains stating it makes it easier to assess, namely: employees, customers, environment, suppliers, the local community, shareholders and society at large. Each of these seven domains falls within one of the three CSR dimensions. In order to test and produce the outcomes as set out in this study, the scale development process as implemented by Öberseder et al. (2014) was used as a guide throughout this study. The scale includes various stages, namely the initial scale item generation and purification (phase one), the measurement model development and refinement stages (phase two) as well as the measurement model validation (phase three) (Öberseder et al., 2014). The scale item generation and purification took place in 2018. During phase one, a structured, self-administered online questionnaire was developed as part of the Consumer Science Clothing Retail Management final year students’ research projects. The development of the questionnaire entailed adapting existing scales from Öberseder et al. (2014) who in turn based their scale development process on a process developed by Churchill (1979) and later used by DeVellis (1991) and Netemeyer et al. (2003), to relate to the emerging market and its population, after which the items were pre-tested to ensure valid and meaningful content (as part of the purification stage). As part of phase two, data was collected online throughout South Africa, with the majority of the responses originating in urban areas. A quantitative research approach was followed together with the use of a cross-sectional survey design to obtain a total of 1632 responses for exploratory purposes. SPSS software was utilised during data analysis and the entire dataset (N = 1632) was randomly split into two datasets to ultimately be able to perform measurement model development and refinement on dataset one (n = 816) as well as measurement model validation on dataset two (n = 816).

The measurement model development and refinement (part of phase two) entailed various exploratory factor analyses (EFA) which ultimately resulted in a six-factor solution labelled as follows: Ethical stakeholders, Social causes, Quality service, Environmental causes, Local sourcing and Pricing policies. These six factors resemble the various domains of consumers’ perceptions of CSR, as proposed by Öberseder et al. (2014), but are more contextual to the emerging market state of South Africa. “Ethical stakeholders” was deemed the most prominent factor, while “Local sourcing” and “Pricing policies” were not as prominent, but remained distinct enough to be retained, and could be deemed valid components in understanding consumers’ perceptions of clothing retailers’ CSR in the South African context. Thereafter, a range of first- and second-order confirmatory factor analyses (CFA) were conducted and convergent as well as discriminant validity methods were used to determine the overall validity of the measurement model. Eventually some items with lower factor loadings were eliminated during the first-order CFAs to ensure further refinement of the measurement model which achieved an overall good fit. Subsequently two second-order CFAs were conducted to determine whether the first-order constructs (Ethical Stakeholders, Social Causes, Quality Service, Environmental Causes, Local Sourcing and Pricing Policies) were true reflections of the higher order construct, namely CPCRCSR. This was performed together with an additional construct, namely purchase intention. All the fit indices were deemed acceptable with several indices being classified as excellent (Hair et al., 2014:617), and ultimately “Ethical stakeholders”, “Social causes” and “Environmental causes” were deemed the most prominent factors relating to consumers’ perceptions of clothing retailers’ CSR.

Phase three, that consisted of the measurement model validation process, once again included a range of first- and second-order CFAs, but contrary to phase two, this phase made use of dataset two (n = 816), to ultimately validate the results as presented in phase two with an independent dataset. Overall, an excellent model fit was achieved, thereby validating the measurement model as seen in phase two. Lastly, structural equation modelling (SEM) was conducted to assess the nomological validity of the scale by inspecting the patterns of the relationship between the construct (i.e. consumers’ perceptions of clothing retailers’ CSR [CPCRCSR]) and the other suggested measure, namely purchase intention (PI). The results indicated that the relationship between CPCRCSR and PI was positive, but very weak, yet still statistically significant. That said, it does not present convincing results that CPCRCSR does in fact affect consumers’ PI when it comes to clothing. This study, together with its validated CPCRCSR scale, could be of utmost importance in further exploring the link between CPCRCSR and PI, and could be used for future research purposes to specifically investigate the relationship between these two constructs to form a solid base for research surrounding CPCRCSR and their overall purchasing intentions and behaviour. Furthermore, future insights derived through valid, reliable measurement scales such as the one proposed in this study could enable companies to better understand their customers to consequently adapt their CSR strategies and identify shortcomings that can be improved.