Tipultech logo

Internal Reliability

Sarah Ogilvie

Others' Views Task. A reliability analysis of this measure revealed two items with item total correlations close to zero. Removal of these items (sociable and verbal) increased the Cronbach's alpha from 0.58 to a more acceptable level of 0.66.

Your Views Task. In calculating the internal reliability of this scale, two items were found to have particularly low item total correlations. These items were not the same as those extracted in the analysis of the OthsVws (they were empathetic and intelligent). When these two items were removed, Cronbach's alpha increased from 0.51 to 0.61.

Ambivalent Sexism Inventory. A reliability analysis of the ASI showed that overall the scale had high internal consistency (Cronbach ╬▒ = 0.83) which was similar to that reported by Glick and Fiske (ranging from 0.83 to 0.92, 1996). The Benevolent subscale showed only moderate internal consistency (Cronbach ╬▒ = 0.65), a finding marginally lower than those found by Glick and Fiske (0.73 to 0.85). However, the Hostile subscale showed high internal consistency (Cronbach ╬▒ = 0.85), a result in keeping with those of Glick and Fiske (0.80 to 0.92). Inclusively these results indicate that while the scale was not as reliable in its measurement of the Benevolent Sexism construct, it was nevertheless reliable in its measurement of sexism overall. For this reason the ASI was included in the major analyses as an entire scale, rather than in the form of its constituent subscales.

Women in society Questionnaire. Despite the fact that there were only six items of this questionnaire employed in this study, an internal reliability analysis was conducted to determine whether these six items were measuring a similar construct. As the Cronbach's alpha was 0.85 it was concluded that this set of items was, in fact, measuring a similar construct.

<< >>