*3.5. Analyses*

Each household's recycling behavior for paper and packaging waste is calculated by averaging the frequency of recycling (one variable) and how much of each of four recyclable materials, namely paper, plastic, glass, and metal (four variables), is recycled. The households are then grouped into clusters based on their recycling behavior [26,49]. One group comprises the households that reported no recycling behavior (B = 1). The recycling households (B > 1) are further grouped into three groups, according to their various levels of reported recycling behavior. Borrowing from both Thomas and co-workers' description of high, medium, and low recyclers [49] and Martin and co-workers' description of full (recycle five materials including plastic, paper, glass, cans, and cardboard), casual (recycle 1 to 4 materials), and non-recyclers (recycle no items) [26], for this study, the terms with definitions are adopted as follows:


For the purpose of this paper, the segmentation into groups diverts from the Thomas et al. example [49] by also differentiating between the non-recyclers (B = 1) and the low (sporadic) recyclers. These two groups were analyzed separately because the large number of non-recycling households would overshadow the responses from the sporadic-recyclers. The author's classification also differs from the Martin et al. example [26] by splitting off a group called the sporadic recyclers (the households that almost never/seldom recycle very little/some of one material) from the casual recyclers (those households that recycle more materials and more frequently than the casual recyclers but less than the dedicated recyclers).

Using MS Excel, the total number of times each reason for non-recycling was selected was calculated and expressed as a percentage of the total number of reasons mentioned. First, the total number of times an item was selected as the one most important reason was expressed as a percentage of the total number of responses. Second, each of the three most important reasons selected was totaled and each statement expressed as a percentage of the total number of reasons mentioned. These calculations were repeated for the total sample group (*n* = 2004) as well as for each of the sub-groups as described above. According to the popularity of each selected reason, i.e., times selected, the factors were ranked [24] (p. 64).

#### *3.6. Research Ethics*

The impersonal phrasing of the question "what would you say are the reasons why people do not recycle?" has value in an ethical sense by not putting the respondents in a position where they feel guilty about their personal recycling behavior. It can also be argued that respondents would be more willing to share information and share it accurately with less desirability bias when a question is posed in a more impersonal manner. However, a more direct question: "why don't you recycle?" would have allowed a more confident comparison. Due to the impersonal phrasing of the question, interpretation of the results should thus be done with caution by keeping in mind that the results portrayed are the perceptions of the respective groups about the "people" and not their perceptions or knowledge about the specific group within which they are placed.
