Hi! In my bachelor thesis am using a new scale (á 12 items) for sexual orientation and the reliability analysis resulted in Cronbachs Alpha .75. After consulting the inter-item-correlation, I decided that due to high correlations, high Alphas and due to the content to aggregate the items, which led me to only having 6 items. But now I am left with an Alpha of only .52. I could exclude one of the aggregated items, which would lead to a Alpha of .63. But that would would exclude "Attraction towards men" alltogether, which doesn't appear to be the most reasonable course.
How do I proceed? Is it valid to not do the aggregation and state why in my thesis?
Actually you should have done a face validity and substantive validity to avoid this situation. Best way to re do if the study.
Otherwise report the findings as it is and dont aggregate, beacuse it dilutes the theortical base.
Your initial Cronbach's value of 0.75 is adequate, and there is no need to exclude any items (only consider to exclude item when alpha value is above 0.90).
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning