Perhaps I am confused, but if (in equation 1) b2 = .29 than would not a
one-unit change in group be a 29% increase in y?
Regarding Eq. 2: The marginal effect of group on y is not simply b2 but:
b2 + b3*time
and the variance would be:
var[b2] + time^2*var[b3] + 2*time*cov[b2,b3],
where time is at some given value; so whether or not the variable group is
statistically significant cannot be determined solely by the looking at the
coefficient and standard error on group (b2).
Scott
> -----Original Message-----
> From: [email protected] [mailto:owner-
> [email protected]] On Behalf Of [email protected]
> Sent: Thursday, October 06, 2005 11:28 AM
> To: [email protected]
> Subject: st: Interaction
>
> Dear Statalisters,
>
> My question is in relation to my previous query about interaction terms. I
> have two regression models:
> 1) ln(y)= constant + b1*(time) + b2*(group)
>
> In this model group is an indicator variable denoting membership in
> group1 or group2. On fiting this model, the value of b2 was .29 indicating
> a 34 percent additional increase in y for the people in group2.
>
> The second model that I fitted included the interaction between group and
> time.
>
> 2) ln(y)= constant + b1*(time) + b2*(group) + b3*(group*time)
>
> In this model , the group is not significant but the interaction effect
> is. But the effect of the interaction reduces to 2.6 percent (b1=.0822,
> b3=.026).
>
> How does one explain (In plain English) the drastic reduction in the
> effect of time on y for the group2 people between the two models?
>
>
> Thank you in advance for any suggestions
>
> Leny Mathew
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
