I would like to conduct ANOVAs and regression analyses with several DV's having a low reliability (cronbach's alpha < .40). This is probably due to the low number of items (2-3) in each scale. Is reliability of all DV and IV a requirement when doing ANOVAs and regression analyses?

  • 1
    $\begingroup$ ANOVA and regression as such have nothing to do with reliability. You should supply more detail about the variables you are using and the model you want to fit. $\endgroup$
    – Placidia
    Aug 12 '15 at 13:08
  • $\begingroup$ OK. I want to fit a model with 5 repeated measures as DV's (unreliable means of questionnaire subscales) and one categorial variable (profession type) and three covariates (age, gender and company). The data is unbalanced as that the cell sizes in profession type are very different (ranging from ca. 20-200). Which type of analysis would you suggest? $\endgroup$
    – teeglaze
    Aug 13 '15 at 14:25
  • $\begingroup$ You could try a MANOVA with the DV's a multivariate response. If the DV's have low reliability, and hence are not highly correlated, doing 5 independent ANOVA's will be just as informative. If you believe that the DV's manifest one factor, you could do a structural equation model. $\endgroup$
    – Placidia
    Aug 13 '15 at 15:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.