In this chapter, the results simulated from SPSS will be presented, explained and discussed. Factor analysis, by its nature, is an iterative process. As such, the discussion will be conducted in a step by step approach. During the procedures, the various assumptions and rationales for the underlying steps to be conducted will be articulated. At the end of this chapter, a total of 10 factors will be extracted from the 46 variables or research questions designed to investigate the students’ perceptions towards Corporate Social Responsibility.
In this chapter, the process and the output from SPSS in the conduct of factor analysis will be presented. Before, the discussion on the process, procedures and implementation of factor analysis is outlined; the variables/ symbols used in this chapter will be firstly defined. There are a total of 46 symbols to be used in this chapter, ranging from VAR00001 to VAR00046, whereby each of these variables will correspond to the respective research questions directed to the students/ research participants. The details of the symbols/ variables to the specific research questions are shown in Table 4.1 as follow.
After defining the symbols to be used in the process of conducting factor analysis in the subsequent section, the discussion on the procedures and steps relevant to the conduct of factor analysis is presented in the next section.
As discussed before, Principle component analysis will be used to reduce the variables into a smaller set of representative variables, whereby these representative variables can still be reasonably good in characterizing the original set of variables. Besides, it is also discussed before that factor analysis start with identifying the total number of factors to be extracted from the many variables (i.e., in this study, from a 46 variables, represented by each research questions). There are two methods widely used to identify the amount of factors to be elicited, namely: (a) latent root criterion; and (b) scree test criterion. Both of the methods will be discussed and used in the following paragraphs.
Latent root criterion. Under the latent root criterion, it is assumed that for any single variable to be retained for further interpretation, these variables should be at least able to contribute minimally a value of 1 to the total eigen-value. As such, technically speaking, only those variables that have an eigen-value of greater than 1 will be considered as significant. The rest of the variables should be dropped, as these variables are already being largely explained by the variation in other variables. As discussed by Hair et. al. (2006), employing eigen-value as a reference point to determine the cutoff point is most reliable when the numbers of variables are between 20 and 50. As such, to employ latent root criterion (i.e., eigen-value as cutoff point) will be relevant to our study here. As shown in Table 4.1 below, the results simulated from SPSS, for identifying the numbers of factors to be elicited from the entire data set, based on latent root criterion is shown as follow. According to the results, it is suggested that a total of 14 factors should be extracted from the list of 46 variables.
Scree test criterion. As discussed by Hair et. al. (2006), scree test is performed by plotting the latent root against the amount of factors, whereby the shape of the curve can be used to assist in determination of the cutoff point. As shown in Figure 4.1, the scree test for the study performed in this dissertation is presented. In the scree test criterion, the point at which the curve shown below started to straighten can be used as the ways to indicate the maximum number to be extracted. From Figure 4.1, it is apparent that the number of factors to be extracted from scree test criterion is about 10 factors.
As such, it is found that from latent root criterion, a total of 14 factors are selected, while from the scree test criterion, a total of 10 factors are selected. For the purpose of the analysis here, a total of 10 factors are selected, because selecting more than 10 factors will complicate the study significantly. Furthermore, if more than 10 factors are selected, then it will defeat the purpose of data reduction in this study. From the latent root criterion, it is observed that if 10 factors are selected, the total cumulative extracted sum of square is 65.4%. As asserted by Hair et. al. (2006), as long as sufficient factors are selected to be able to meet a 60% of variance explained in the set of variables, it can be considered that those selected factors can reasonably used to represent the set of variables under investigation. As such, to select a total of 10 factor in this analysis is indeed justifiable.
After deciding the relevant number of factors to be extracted, factor rotation process will be conducted. In Table 4.3 below, the descriptive statistics for all variables (i.e., answers from all of the research questions) are presented. In the table, it is found that the variables or research question yielding the highest score/ mean value is VAR00020. VAR00020 is about companies should ensure environmental protection in acting in a socially responsible manner. The mean value for VAR00020 is 6.095 with a standard deviation of 0.6916. Considering that the research question is a 7 scale Likert style question, such a score should be considered as a high one. This strongly suggests that the students believe that companies that exhibit Corporate Social responsibilities should and will ensure environmental protection. The second highest scoring variable is VAR00015, which is about if the company should ensure judicious use of natural resources. VAR00015 has a mean value of 5.955 and the standard deviation of 0.6747. In a similar vein, the third highest scoring variable is VAR00005, which is about if the company should maintain ecological balance. Similarly, VAR00005 has a mean value of 5.955 and a corresponding standard deviation of 0.6747. This is a somewhat surprising finding, as the top three questions indicate that from the student perspectives, to act in a Corporate Social Responsible manner is highly associated with issues related to environmental protection and preservation. Then, the fourth highest scoring variable is VAR00025, which is about if the company should fulfill their responsibilities towards customers. The variable has a mean value of 5.86 and a standard deviation of 0.8913. Then, the fifth highest scoring variable sis VAR00038, which is about if a company, should adhere to law. This should be nothing surprising because to act in a Corporate Socially Responsible manner surely demand any company to firstly adhere to the respective rules and regulation governing the respective business and industry. VAR00038 has a mean value of 5.77 and a standard deviation of 0.9456.
From another perspective, the lowest scoring variable is VAR00034, which is about if the company should fulfill responsibilities towards suppliers. VAR00034 has a mean value of 3.25 and a standard deviation of 1.4861. Considering that the research question is a 7 scale Likert style question, a score lower than 3.5 should indicate that the student may not be agree that a corporate socially responsible company should focus on fulfilling its responsibilities to the suppliers. This is perhaps from the customer centric perspectives; student may think that it is the suppliers that should act in a responsible manner to the company. However, this is purely speculation and more investigation should be conducted to understand the rationales of students to perceive in such a way. Then, the second lowest variable is VAR00040, which is about if the company should act responsibly towards intermediaries. This variable has a mean value of 3.30 and a standard deviation of 1.4870. Last but not least, the third lowest scoring variable is VAR00036, with a mean value of 3.445 and a standard deviation of 1.4552. This variable is about if the company should fulfill responsibilities towards competitors. Apparently, the students under this study do not quite agree that a company should also act in a responsible manner towards competitors. Anyway, such a discussion should provide some insights to the readers, but as the objective of this dissertation is to ferret out the minimum number of factors that able to describe the attitudes and perceptions of management student towards the concepts of Corporate Socially Responsibility in a parsimonious manner, the discussion will revert back to factor analysis in the subsequent sections.
In Table 4.4, the ranking of all of the variables or research questions is presented. The variables with the highest score is arranged to be on top of the list, and so on, until the variable with the lowest score. Overall, it can be seen that variables related to environmental protection, ethical behaviors, being responsible to customers and employees, or to act responsibly to the community rank higher than other variables. Those variables with lower scores are found to be those issues either less discussed in the context of CSR, or most likely perceived by the students not as important as other variables. A close investigation on the results findings indicates that the student apparently rank those variables that is more popular and hotly discussed in the context of CSR, as compared to the other less discussed or pressing issues.
Before entering to stage two, whereby factor rotation is to be conducted, the communalities of each of the variables should be investigated. Communalities can be technically defined as the amount of variance in a particular variable that is accounted for by two of the factors taken together. As asserted by Hair et. al. (2006), the degree of communalities is a practical reference point to judge how much variance in the respective variables that can be accounted for by the factor solution. The authors suggested that it is probably useful for the researchers to delete the variables with low communalities from the analysis. However, the degree either the communalities figure can be considered as large or small will be subjective to researcher interpretation and judgment call. In this study, variables with communalities lower than 0.40 will be deleted. Those communalities with a score lower than 0.40 are highlighted in Table 4.5 below. All of these variables will be deleted and excluded from the factor analysis, as a huge portion of the variances of these variables are not accounted by the factors.
Besides, a review of these variables with low communalities indicates that these variables may be excluded from the context or discussion of Corporate Social Responsible for sensible reasons. For example, VAR00017 (with a communalities of 0.346) can be reasonably excluded because it is about if a socially responsible company should ensure effective grievance handling. This is indeed a rare issues discussed in the context of CSR, and such issues is hardly related to the other forty five variables. Indeed, to delete the variables from the discussion is reasonable as the objective of this study is to reduce the many variables into a selected few representative dimensions. Apart from that, VAR00033 can be deleted as well, because it is about argument that socially responsible companies should invest in innovative technologies. This research question is apparently different from other research questions. Similarly, VAR00035 (i.e., socially responsible companies should follow proper business model) and VAR00045 (i.e., socially responsible companies should follow core values) are deleted as well because the variables are actually harder to be associated with other variables. Overall, it can be observed that variables with low score of communalities are issues less discussed in the context of CSR, and that the low communalities values probably are contributed by the more diverse opinions from the students. Indeed, without performing factor analysis, purely based on subjective judgment process, the many research questions can be categorized into several groups as follow: (a) environmental protection, (b) ethical behaviors, (c) being responsible to customers and employees, or (d) to act responsibly to the community. These variables, namely VAR00017, VAR00033, VAR00035 and VAR00045 can hardly be associated with any of the groups identified above.
Upon deleting the variables with low communalities value, factor rotation will be performed. As previously discussed, a total of 10 factors will be used to describe the set of variables in this study. The factor rotation method employed is Varimax rotation method. The results simulated from SPSS are presented in Table 4.6 below. In the table, each of the factor loadings of the variables, correspond to each of the factors are shown. It is observed that most of the variables can be easily associated with one of the ten factors shown below. In the table below, the respective factor loading that is above 0.50 will be highlighted. However, it is found that two variables may be deleted or excluded from this study due to issue of cross factor loading, and secondly, having a factor loading less than 0.50 to any one of the factors. For this, VAR00029 will be deleted from the analysis because it has two factor loadings with score exceeding 0.50. That variable actually is above if socially responsible company should prevent insider trading. Theoretically, this issue is about a company should take proper actions to safeguard shareholders’ interests, as insider trading will cause harm to the mass shareholders. As it is discovered that there are other useful variables describing the issue that a company should concern about shareholders’ interest, the deletion of such variable will not cause huge impacts to this study. Then, VAR00044 will also be deleted, as the variable has essentially no factor loadings score of above 0.50 to any of the ten factors outlined in Table 4.6. VAR00044 is about if the company should promote social values. As this is a topic quite distinctive to the theme of other research question, it is probably appropriate to exclude such variables from the analysis, as the objective of this study is to reduce the many variables into a few critical representative variables. Upon deletion of the two variables from the analysis, the remaining variables will be re-simulated and rotated to yield a better model, as will be presented in Table 4.7.
Again, the re-simulated results are shown in Table 4.7. The variables are found to have pretty satisfactory factor loadings (i.e., above 0.500) to their respective factors in the results. For each of the variable, the column with the highest factor loading is highlighted. Besides, referring to Table 4.8 below, the model is found to be able to explain more variances among the variables as compared to the situation before the several variables (namely, VAR00017, VAR00029, VAR00033, VAR00035, VAR00044, and VAR00045) are deleted. To be specific, the new model is found to be able to explain a total of 71% of the variances among the variables (i.e., from table 4.8), as compared to the earlier model that can only explain a total of 65% among the variables previously (i.e., from table 4.2).
Upon several iterative simulation processes, the set of variables had already been successfully summarized in into a total of ten factors, as shown in Table 4.7. In this section, the results findings from the factor analysis performed above will be articulated and summarized.
It is found that there are a total of 8 variables correspond to factor 1, namely VAR00001 (i.e., company should proactively promote public interest); VAR00004 (i.e., company should promote community volunteering); VAR00006 (i.e., company should undertake charitable activities); VAR00007 (i.e., company should build healthy public relations); VAR00012 (i.e., company should fulfill responsibility towards society); VAR00021 (i.e., company should build efforts towards social upliftment); VAR00023 (i.e., company should fulfill responsibility towards society); and VAR00043 (i.e., company should preserve the heritage). A review of these variables found that they have a common theme, namely, responsibility and delivering benefits to society. As such, factor 1 can be regarded as the dimensions where companies are expected to act responsibly and delivering benefits to society.
Next, it is found that there are a total of five variables, namely VAR00002 (i.e., company should prevent corruption); VAR00003 (i.e., company should follow ethical standards); VAR00013 (i.e., company should promote ethical advertising); VAR00018 (i.e., company should avoid unethical business practices); and VAR00024 (i.e., company should follow corporate behavior) correspond to factor 2. A review of these variables found that they have a similar theme, whereby it is expected that companies should adopt business ethics in the conduct of the business. Thus, it is reasonable to name this factor as the practice of business ethics.
Then, the variables correspond to factor 3 include: VAR00016 (i.e., company should ensure corporate governance); VAR00026 (i.e., company should ensure transparency); VAR00028 (i.e., company should undertake corporate governance disclosure); VAR00030 (i.e., company should fulfill responsibility towards shareholders); VAR00031 (i.e., company should undertake social accounting); and VAR00032 (i.e., company should incorporate self-regulations). These variables can be perceived as something related to shareholders and the financial market. As such, it can be perceived that factor 3 should be safeguard of shareholders’ interests.
Then, the variables correspond to factor 4 include: VAR00005 (i.e., company should maintain ecological balance); VAR00015 (i.e., company should ensure judicious use of natural resources); VAR00019 (i.e., company should ensure sustainable development); VAR00020 (i.e., company should ensure environmental protection); and VAR00022 (i.e., company should ensure corporate sustainability). There are two themes on these variables. Firstly, it is bout environmental protection, and secondly, it is about sustainable development. As such, this dimension can be considered as responsibilities of companies towards natural environment and sustainability.
There are four variables correspond to factor 5, namely: VAR00034 (i.e., company should fulfill responsibilities towards suppliers); VAR00036 (i.e., company should fulfill responsibilities towards competitors); VAR00037 (i.e., company should fulfill responsibilities towards creditors); and VAR00040 (i.e., company should fulfill responsibilities towards intermediaries). In this factor, it is apparent that these are issues related to responsibilities of companies towards other stakeholders (except employees and customers). For this, it is found that the issues on companies’ responsibilities towards employees and customers are characterized in factor 6 and factor 9.
Then, it is also found that there are three variables correspond to factor 6, namely: VAR00025 (i.e., company should fulfill responsibility towards customers); VAR00039 (i.e., company should produce fair products); and VAR00041 (i.e., company should offer fair price to the customer). All of the three variables have common theme, whereby it is about company responsibilities and interaction to customers or consumers. As such, factor 6 can be reasonably named as company responsibilities towards customers.
Then, there are three variables correspond to factor 7, as follow: VAR00010 (i.e., company should promote health); VAR00014 (i.e., company should improve quality of life); and VAR00046 (i.e., company should promote safety). There of the variables are about safety, health and life. As such, factor 7 can be considered as a dimension regarding health, safety and lifestyle.
The two variables correspond to factor 8 are VAR00009 (i.e., company should adopt global standards) and VAR00011 (i.e., company should follow international norms). The two variables are similar in the sense that they are about adoption of certain practices or standards. As such, factor 8 will be termed as global norms and standards.
The two variables correspond to factor 9 are VAR00008 (i.e., company should maintain healthy labor standards) and VAR00042 (i.e., company should fulfill responsibilities towards employees). For the two variables, it can be seen that they are about employees’ related issues. Thus, factor 9 will be interpreted as responsibilities of company towards employees.
The two variables correspond to factor 10 are VAR00027 (i.e., company should provide stable business environment) and VAR00038 (i.e., company should adhere to law). These two variables have lesser common theme in place. The possible explanation is that this factor is about business environment and regulation. As such, factor 10 will be specified as company’s responsibilities towards business environment and regulation. Overall, the 10 factors identified from the factor analysis process are sensible and easily understood. The 10 factors are listed in Table 4.9 as follow.