Quantcast
Channel: Statalist
Viewing all articles
Browse latest Browse all 73370

-xtpoisson, fe- Interpreting 3-way interaction

$
0
0
Hello forum members,

Suppose I have the following model:
Code:
 xtpoisson y x1 x2 c.z1#c.z2#c.z3, fe irr
where x1 and x2 are time-variant controls, z1 and z2 are fixed (time-invariant) and z3 is time-variant. As such, the FE model estimation is not feasible, since z1 and z2 are fixed.

Therefore, I create:
Code:
gen RT = z1*z2*z3
Note, RT becomes time-variant and I can estimate:
Code:
 xtpoisson y x1 x2 RT, fe irr
My question then is, how can I interpret the effect of RT, considering its an "interaction" of 3 measures (say, its IRR coefficient equals 1.10 [+10%])?

Thankfully,
Anton

Viewing all articles
Browse latest Browse all 73370

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>