Response to non-hate crime scenarios. A crime scenario was included in each

protected category that did not include indicators of a bias crime. There were a total of

five crime scenarios that were not hate crimes. A separate sub-category consisting of

these items was formed for the purpose of analysis. The non-bias crime scenarios were

rated markedly lower on the scale compared to the bias crime items (M= 10.17, SD =

4.67, Range = 5 to 31) (See Figure 4.1). The ANOVA indicated a significant interaction

across all categories, including the non-hate items F(5,1096) = 589.26, p > .001.

Levene's test was significant F(5,1096) = 17.51, p > .001 indicating the assumption of

equal variances had been violated. Tamhane's test was used to correct for the violation

and the mean score difference between non-hate items and the five sub-scales were

significant <.001. These results are consistent with the pilot study results for the same

items and provided evidence of the discriminant-validity of participant responses to the

instrument.

 The non-hate crime items were removed from the analysis when the ANOVA's for

subsequent analyses were performed. The initial purpose of the non-hate crime items was

to establish LEO's ability to discriminate between hate crime and non-hate crime items.

Once established, the items were eliminated from the sub-scale scores for each protected

category. The aim of the study was to look for variability across similar items. It was

believed the non-hate crime items would introduce unnecessary variability into the

subscales. Therefore, no other analysis in this study included the non-hate crime items.

 Two-way ANOVA. A 5 x 2 two-way (repeated measures) ANOVA was conducted

to evaluate the effect of five protected category and gender on law enforcement officer