Hereabout snp defaults to a lower case version of the Scottish National Party, although perhaps Qiang meant single-nucleotide polymorphism. It's a fair guess that at least some of the econometricians and applied economists and many others on this list are not familiar with either, although the former might want to enlist them as instrumental variables.
Fortunately Qiang's question was clear without this understanding, but here's a reminder to those who need it that this is an _interdisciplinary_ list based on a Venn diagram of shared ignorance. Please try to avoid subject-specific jargon.
Nick
[email protected]
Martin Weiss
Nobody prevents you from creating the dummies beforehand via -tab, gen()-. The interaction would have to be created by hand, though. See also
ssc d dummieslab
-xi- automatically prevents you from falling into the dummy trap and overrides previously created "_I" vars, which is a neat feature...
Qiang (Danny) Chen, Ph.D.
When I use xi (Interaction expansion), I got the following report. I am wondering if there is a way to keep the original snp names. For example:
_Irs1Xrs~1_1 ==> _Irs1234567_1Xrs7654321_1
Thanks!
xi: logit case sex i.rs1234567*i.rs7654321, or
------------------------------------------------------------------------------
case | Odds Ratio Std. Err. z P>|z| [95% Conf. Interval]
-------------+----------------------------------------------------------------
sex | .2390816 .0457164 -7.48 0.000 .1643547 .3477845
_Irs123456~1 | .2017002 .2995979 -1.08 0.281 .010974 3.707226
<snip>
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
