Australian Institute of Criminology

Skip to content

Appendix A

Factor and principle components analysis

First, the factorability of the 17 items was examined. The Kaiser-Meyer-Olkin measure of sampling adequacy was 0.82 (over the recommended value of 0.6); the scale reported an alpha of 0.76 (above the recommended 0.7 that suggests reliability); and all but one item had communalities above 0.3. This suggests that the measures were reliable and valid and that a significant proportion of the variance was explained by all possible factors.

The factor analysis was conducted with all 17 items in several stages. A principal components analysis was performed including all possible factors. The initial eigen values indicated that the first factor explained 24 percent of the variance, the second factor 12 percent, the third factor eight percent and the fourth factor six percent. Therefore, a four-factor solution, which explained 40 percent of the variance, was preferred because of the drop in eigen values on the screen plot after four factors and the insufficient number of primary loadings on subsequent factors. During several steps, four items were eliminated as they failed to contribute to a simple factor structure and did not meet the minimum criterion of having a primary factor loading of 0.5. Factor loading refers to the degree that the variance of an item’s responses can be attributable to an underlying attitude being measured. The items: ‘The government should focus on prosecuting traffickers’, ‘It is reasonable for victims of trafficking to not contact the authorities because they fear what traffickers may do to them or their families’, ‘A victim of trafficking would attempt to contact authorities at the first opportunity’ and ‘People who enter Australia and work illegally have done so because they cannot do so legitimately’ did not load above 0.5 on any factor. This suggests that the responses to these items did not correlate to a single construct, or in this context, a single attitude.

A principal components analysis of the remaining 13 items, using varimax rotation, was conducted on the four-factor solution, with the factors explaining 60 percent of the variance. All items had primary loadings above 0.5 and none had cross-loadings above 0.3.