The questionnaire data was analysed using the statistics package, SPSS version 11. The data was coded and transformed into a number of different types of variable namely nominal (binary), nominal (categorical), ordinal and scale. A selection of four different non-parametric statistical tests was used to test for the existence of statistically significant associations or correlations between different combinations (pairs) of variables. The rationale behind the choice of the four tests was that the most powerful and appropriate test available should be used for any given combination of variables. Table 1 indicates which test was used for a particular combination of variables.
Table 1. Statistical tests used on variable combinations
There were essentially three types of variables used in the study namely the independent variables (see following paragraph), the demographic variables (gender, age, occupation and educational attainment) and the dependent, perceptual variables representing the four research themes (more information is given about these in the thematic sections below).
Three independent variables are referred to in this paper and these are “vegetation density”, “housing density”, and “location in relation to Birchwood”. The scores for the vegetation and housing density of the 12 HCA’s were simply transformed into the scale variables “vegetation density”, and “housing density”, where the values consisted of the vegetation scores and dwellings per hectare respectively. The variable “location in relation to Birchwood”, was a nominal (binary) variable, where the values 1 and 2 denoted whether the respondent lived in Birchwood, or in one of the three control HCA’s outside.