Statalist The Stata Listserver


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: using reshape


From   frone@ria.buffalo.edu
To   statalist@hsphsun2.harvard.edu
Subject   st: using reshape
Date   Mon, 11 Sep 2006 17:22:01 -0400

Hello,

I have a data set in wide form.  Part of the data have to do with work 
injuries.  After asking whether a person experienced a work injury during 
the past 12 months and then asking how many injuries were experienced, the 
person reported on a number of potential characteristics of the 
injury(ies).  The series of characteristics were asked for up to 5 injury 
events.  Obviously, not everyone had up to 5 injuries.  I want to convert 
the file into long form where each record represents an injury event. 
There are 2535 respondents in the wide file.  Had everyone experienced at 
least 5 injuries, I would expect the number of records in the long file to 
equal 12675.  However, the actual number of records should be much lower 
than 12675. 

This is a partial record for a respondent in wide format:


respnum_  = respondent number
finalwt2 = sampling weight
gender = gender
ir1  = did an injury occur--yes/no
ir2 = number of injuries experienced
ir3_1 =   injury characteristic 1 for injury 1  -- yes/no
ir3_2 =   injury characteristic 1 for injury 2  -- yes/no
ir3_3 =   injury characteristic 1 for injury 3  -- yes/no
ir3_4 =   injury characteristic 1 for injury 4  -- yes/no
ir3_5 =   injury characteristic 1 for injury 5  -- yes/no
ir4_1_1 =   injury characteristic 2 for injury 1  -- yes/no
ir4_1_2 =   injury characteristic 3 for injury 2  -- yes/no
ir4_1_3 =   injury characteristic 4 for injury 3  -- yes/no
ir4_1_4 =   injury characteristic 5 for injury 4  -- yes/no
ir4_1_5 =   injury characteristic 3 for injury 5  -- yes/no


respnum_        finalwt2                gender          ir1     ir2 ir3_1 
ir3_2   ir3_3   ir3_4   ir3_5   ir4_1_1 ir4_1_2 ir4_1_3 ir4_1_4 ir4_1_5
354             16151.351       male            yes     2       no yes  . 
.       .       no      no      .       .       .


To get the data in long format, I executed these commands:


sort respnum_;

reshape long 
  ir3_@ ir4_1_@,i(respnum_) j(injnum);


I obtained data in this form:

respnum_          injnum                finalwt2                gender ir1 
ir2     ir3_    ir4_1_
354             1               16151.351       male            yes     2 
no      no
354             2               16151.351       male            yes     2 
yes     no
354             3               16151.351       male            yes     2 
.       .
354             4               16151.351       male            yes     2 
.       .
354             5               16151.351       male            yes     2 
.       .


However, because this person only experienced 2 injuries, I  wanted the 
data in this form:

respnum_          injnum                finalwt2                gender ir1 
ir2     ir3_    ir4_1_
354             1               16151.351       male            yes     2 
no      no
354             2               16151.351       male            yes     2 
yes     no


How can I keep the number of records for a given respondent equal to the 
number of injuries experienced?

Thanks,

 Mike Frone


*
*   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/



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