Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
Nick Cox <njcoxstata@gmail.com> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
Re: st: Creating interaction and square variables |

Date |
Fri, 31 Aug 2012 02:20:54 +0100 |

It seems that you are not getting the point. In the case of gen f`i'f`j'=firm`i'*firm`j' if `i'<=`j' a variable will always be created. However, the qualifier ... if `i' <= `j' will mean that _all_ its values will be missing if that condition `i' <= `j' is not true. That is pointless. Consider any i and j pair for concreteness, say i = 7, j = 3 or i = 3, j = 7. The result of the -if- qualifier being evaluated is the same regardless of which observation is being considered. That is why, as Dimitriy explained, the appropriate check is to use the -if- command, not the -if- qualifier. That way, you get one test, not several. However, -drop-ping variables you don't want is not needed here at all. Just ensure that they are not created in the first place. You don't need any kind of -if- to do this. For example, forv i=1/22 { forv j=1/`i' { gen f`i'f`j'=firm`i'*firm`j' } } It may help to think about this in matrix terms. You want one half of a matrix, and its principal diagonal. Nick On Fri, Aug 31, 2012 at 1:49 AM, Mike Kim <kalisperos@gmail.com> wrote: > Thanks, Dimitriy. > > I changed the code like the following and it works. I am still curious > because the first "if" works well. I am using Stata 12. > > forv i=1/22 { > forv j=1/22 { > gen f`i'f`j'=firm`i'*firm`j' if `i'<=`j' > } > } > > forv i=1/22 { > forv j=1/22 { > if `i' >`j' { > drop f`i'f`j' > } > } > } > > Mike. > > -----Original Message----- > From: owner-statalist@hsphsun2.harvard.edu > [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Dimitriy V. > Masterov > Sent: Thursday, August 30, 2012 7:31 PM > To: statalist@hsphsun2.harvard.edu > Subject: Re: st: Creating interaction and square variables > > Mike, > > You are not understanding how if works. Take a look at this: > http://www.stata.com/help.cgi?ifcmd > > Depending on which version of Stata you have, you also might want to look > into factor variable notation to do this for you on the fly. > > DVM > > > On Thu, Aug 30, 2012 at 3:39 PM, Mike Kim <kalisperos@gmail.com> wrote: >> Hi all, >> >> I would like to create interaction and square variables. Let say, I >> have three variables x1, x2, x3. >> >> I used to the following code to create interactions and square of x1, >> x2, x3. >> >> forv i=1/3 { >> forv j=1/3 { >> gen x`i'x`j'=x`i'*x`j' // if `i'<=`j' : This still creates > variables >> with missing values. >> } >> } >> >> Since x1*x2 = x2*x1, for example, I tried to drop those same variables >> using the following code, but got error message: "invalid syntax, r(198)". >> >> forv i=1/3 { >> forv j=1/3 { >> drop x`i'x`j' if `i'<`j' >> } >> } >> >> Can you please let me know what is wrong in the code? * * 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/

**References**:**st: Creating interaction and square variables***From:*"Mike Kim" <kalisperos@gmail.com>

**Re: st: Creating interaction and square variables***From:*"Dimitriy V. Masterov" <dvmaster@gmail.com>

**RE: st: Creating interaction and square variables***From:*"Mike Kim" <kalisperos@gmail.com>

- Prev by Date:
**RE: st: Creating interaction and square variables** - Next by Date:
**Re: st: RE: Mean test in a Likert Scale** - Previous by thread:
**RE: st: Creating interaction and square variables** - Next by thread:
**st: negative binomial regression accepts non-integer continuous outcomes?** - Index(es):