Page 1 of 1

Comparing the significance of 2 datasets

Posted: Mon May 06, 2013 5:16 am
by EmperorBE
Hi,

I have been searching on the forums but couldn't find an answer for my questions yet.
I made 2 regressions about dividends for 2 different time periods, 2003-2005 and 2007-2009. Now I want to proove that the same coëfficiënt is statistically more significant in one sample compared to the other. Is this possible? I was trying to calculate the t-statistic but I could find the std. dev. from the coefficients to do it....

The problem is the coefficients are both significant at the 5% level, but one has a p-value of 0,0281 and the other one is 0,0139. I hope to say that during 2007-2009,which is the second p-value I gave, the coëfficiënt is more significant than during 2003-2005.

Thx in advance!

Re: Comparing the significance of 2 datasets

Posted: Mon May 06, 2013 8:33 am
by startz
That's exactly what the p-values mean, although it's hard to see (without knowing a lot more) why one would care about a coefficient being "more significant."