Dissertation for my MSc degree:
- I have a sample of 9 pax. The questionnaire is adopted without any changes. Previously proven to be reliable. Nevertheless, I decided to run a pilot study. It was mostly for face validity, but I applied Cronbach's alpha just for the fun of it. Two of the 6 variables were giving negative scores. (one variable has 5 questions, the other has 6 questions) I have already double-checked the scoring key, and there seems to be no mistake there. Does anyone have some advice for me?
Many many thanks in advance,
Ben
Cronbach's alpha measures the average correlation between items in a research instrument. Negative values may indicate the opposite direction of the item's content in relation to the other items in the questionnaire.
Delete the items that decrease your alpha, they do not fit your population. I give you an example and you can check this paper here on RG;
Student nurses and the general population in Sweden: Trends in attitudes towards mental illness
Full-text available
- July 2012
- Nordic Journal of Psychiatry
- 📷Béatrice Marianne Ewalds-Kvist
- 📷Torbjörn Högberg
- 📷Kim Lützén
Forzosamente tienes ítems que están negativamente correlacionados. Puedes obtener una matriz de correlaciones y ver dónde está el problema (qué ítems mantienen correlaciones negativas). No obstante, con una muestra tan pequeña prácticamente cualquier cosa puede ocurrir.
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Statistics
- Probability
- Probability Theory
- Reliability