Dynamic panel data estimation for 5 years-interval survey data

[Kasbas Bandy]

Can I use dynamic panel data estimation for household survey data that held not annually (it has interval about 3-7 years)? My friend told me that it can be biased, but i didn't get the satisfying explanation. Thank you

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

You don't really tell us the sample or your model or even what you mean by dynamic. If by dynamic you mean including a lagged dv, then yes, it is likely that you need to use an estimator designed for such things - it can be done in xtivreg, xtivreg2 (user written), the xtabond set of estimators, xtdpdml (user written) or several other estimators.

There is a serious problem if some of your panel time periods are 3 years and some 7. Then, what a lag means changes with the panel. I'm not sure what to do about that.