I found b= consistent under Ho and B= inconsistent under Ha, efficient under Ho, its obtained form xtreg.
In this situation, i need to choose re . Am i right ?
Text Ho: difference in coefficients not systematic
chi2 (0) = 0.00
prob>chi2= .
( v_b-V_B is not positive definite)
what does mean ? how can i interpret its significance level in testing my hypothesis ?
Not sure I'm understanding your question right, but the Hausman test is used to distinguish situations in which it is OK to assume that some predictor in a mixed effects mode can be treated as a random effect, meaning it is orthogonal to the fixed effects and is normally distributed around zero (sometimes the ranef adds something, sometimes takes away, not systematic).
If the Hausman test is not significant ("difference in coefficients not systematic"), then a model assuming the random effects as fixed effects is not significantly different, meaning you can retain the null hypothesis, which is that the random effects are random. This doesn't tell you anything about the significance of the model itself (i.e. whether it's a good model to predict the response variable), but it does tell you whether it's OK to assume that your random effect is random.
Hope that answers your question!
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation