Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <njcoxstata@gmail.com> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: Using value labels to create varlist |
Date | Wed, 13 Mar 2013 10:55:22 +0000 |
If you like -ds- for this purpose, -findname- (SJ) provides the same functionality -- and more besides, which could be helpful for other problems. It is a personal second take on -ds-. The help for -findname- comments on the complicated history of -ds-. The syntax here could be findname, vallabel(LABA) In addition to the examples in the help, here are two more . findname, all(@ == @[1]) finds all variables that are actually constants.(If values are all equal to each other, they are certainly equal to the first value, and conversely.) . findname, all(missing(@)) should then be guessable. Nick On Wed, Mar 13, 2013 at 9:31 AM, Maarten Buis <maartenlbuis@gmail.com> wrote: > On Wed, Mar 13, 2013 at 12:18 AM, <J.Cunliffe1@lse.ac.uk> wrote: >> So I got a dataset with about 1,500 (categorical) variables, but about >> 70 value labels mixed in pretty randomly though out the variables. All >> the value labels are slightly different and the categories change from >> one value label to the next (some binary, some likert 5 point scales, >> some 4 etc and so forth). For each variables with the same value label I >> need to send some to missing (for instance with variables with value >> label LABBG - need 5 and 6 to go to missing, for LABA, need 3 to go to >> missing...). The changes are consistent for each variable with the same >> value label... >> >> Is there a way to pick variables by their value label, alter the >> variable by sending a few values to missing, then picking the next value >> label... appreciate that I'll have to code up _how_ I want each value >> label to behave, but it's the picking of the variables that has got me >> stumped, lookfor doesn't seem to pick them... > > ds, has(vallabel LABA) > di "`r(varlist)'" > > > ds, has(vallabel LABBG) > di "`r(varlist)'" > > Hope this helps, > Maarten > > Ps. It is a Likert scale not a likert scale as it is named after a > person: Rensis Likert. > * * 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/