Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down at the end of May, and its replacement, statalist.org is already up and running.


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

Re: st: Determining mutual exclusivity for a series of dummy variables.


From   Jeremy Page <miahpage@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: Determining mutual exclusivity for a series of dummy variables.
Date   Fri, 8 Mar 2013 15:30:49 -0500

Dear Stephen,

I find the command -tablist- (findit tablist) helpful in these
situations. Below is an excerpt from the help file.

The tablist command is useful when you want to make crosstabs of
variables but you want the results to be in a list type format.  An "n" way
crosstab implies the need for some kind of an "n" way table.  Instead,
tablist treats each variable as a column makes an "n" way table by using
"n" columns for the variables.  This is especially useful when each variable
has a small number of values, since this can yield a very compressed table
summarizing the data.


Best,
Jeremy



On Fri, Mar 8, 2013 at 1:49 PM, Nick Cox <njcoxstata@gmail.com> wrote:
>
> That's the first thing I thought of too.
>
> Another thing is to look at the correlation matrix using -correlate-.
>
> The diagonal is all 1s as usual. A variable out of line will have
> correlations distinct from others in the same row and column, or so
> experiment indicates.
>
> Nick
>
> On Fri, Mar 8, 2013 at 6:40 PM, Richard Goldstein
> <richgold@ix.netcom.com> wrote:
> > assuming you have 20 variables each coded 0/1, use -egen- with the
> > rowtotal option to get sum per unit across the 20; you could also check
> > the min/max values of the variables (to see if any have codes other than
> > 0/1) using rowmin/rowmax (or using -assert- as a different strategy
> >
> > Rich
> >
> > On 3/8/13 1:35 PM, Stephen Cranney wrote:
> >> Hi all, quick question. I have about twenty separate dummy variables
> >> that, in theory, are supposed to be mutually exclusive. I did a
> >> regression with all the variables and it didn't kick out one of the
> >> variables automatically, so I assume that there is either some overlap
> >> or there are cases with zero values for all the variables. I checked
> >> the cases used in the regression using the "estimates store sample"
> >> command, and none of them have zero values for all the dummy
> >> variables, which leads me to believe that they aren't mutually
> >> exclusive. I've checked my coding numerous times and the coding for
> >> the variables seems accurate.
> >> So my question is: is there a parsimonious way to quickly check
> >> whether the series of dummy variables are mutually exclusive without
> >> having to check every combination by hand?
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index