st: Re: Uncollapsing Likert Scales

["Michael Blasnik" <michael.blasnik@verizon.net>]

Are you sure you can't work with Stata's frequency weights instead? 
Assuming you have figured out why you want this format, and assuming I 
understand you correctly, I think a reshape and expand may do what you want.

First, the example data you show is obviously not in Stata format since your 
column (variable) names are not legal in Stata.  Let's assume you can rename 
the 5 Likert categories as cat1...cat5 and also assume that the question 
variable is called question.  Then:

reshape long cat, i(question) j(likert)
expand cat
drop cat
label define Likert 1 "S. Disagree" 2"...........

should get you something like what you want.  Beware that there will be QxN 
observations overall, where Q is the number of questions and N is the number 
of respondents.

Michael Blasnik


----- Original Message ----- 
From: "Donald Spady" <dspady@ualberta.ca>
To: <statalist@hsphsun2.harvard.edu>
Sent: Wednesday, September 20, 2006 1:55 PM
Subject: st: Uncollapsing Likert Scales



> I have data of the sort below.  They are summary reports of Likert scale
> type questions.
>
>                  S.Disgree  Disgree Neutral Agree S.Agree
> I like school        2           3    5        10     9
>
> I want to create a variable that will take the individual values and give 
> me
> something like
> Var Label: I like school  Value label 1:S.Dis 2:Dis 3:Neut 4:Agree 
> 5:S.Agree
> which then gives me, in this case, 29 records for the variable labelled "I
> like school"
>
> I have searched the FAQs and the Statalist archives and have read the
> section on Expand in the manuals but have not come up with anything to 
> help
> me here.
>
> I initially thought of creating 5 variables, one for each Value Label
> e.g.
> Variable         Value
> I like school A    2
> I like school B    3
> I like school C    5
> I like school D   10
> I like school E    9
>
> But then I do not know how to collapse these 5 variables into 1 variable
> with the appropriate associated values
>
> Is there a simple way to do this?
*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/