Dear all,
I conducted a CBC study and I have a few questions:
1) Before data collection I ran the advanced test and determined my minimum sample size by looking at the standard errors from the output. As recommended, the standard error for my interaction terms were below 0.1. Now I look at my HB estimation results and I wonder how to test how good my data actually are. I can only see the Standard Deviations and nothing else. What can I do?
2) Should I look at my Logit estimation results if I plan to use HB data? The manual says that using Logit results is not recommended.
3) Finally, what is the main difference between the effects from Logit estimation and the average (!) utilities from zero centered diffs in HB estimation? Can you say that the utilities from HB estimation account for heterogeneity while Logit does not?
Many thanks!