[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: re: rollreg
Christopher F Baum wrote:
Ah! Thank you. I'd offer to help investigate, but I'm a complete novice
as far as programming is concerned. Still I've give it a shot. Anyway,
thanks for both your answer and for writing this useful routine.
The coefficients which you are getting from rollreg correspond to those
calculated from observations 2/6, not 1/5 (that is, for 1981-1985 rather
than 1980-1984): that is why they are labeled as aligned with 1985. Now
why rollreg does not produce the estimate for obs. 1-5 I do not know,
and will have to investigate. But the rollreg ests. match those
generated by "reg y x1 x2 in 2/6" exactly.
* For searches and help try: