Stat Computing > Stata > Notes

Stata Class Notes 3.0
Reshaping Datasets

1.0 Stata commands in this unit

. reshape

2.0 Demonstration and Explanation

There are two major varities of data files; those that are long and those that are wide. An example of a wide file is shown in the table below.

Example Wide File

 id       grp     lrn95     lrn96     lrn97
  1         1         7        12        16 
  2         1        13        14        15 
  3         2        15        14        20 
  4         2        21        27        24 
  5         3         9         9        12 
  6         3        16        17        18 

We would like to change this file to the long format that looks like this:

Example Long File

 id       year       grp       lrn 
  1         95         1         7  
  1         96         1        12  
  1         97         1        16  
  2         95         1        13  
  2         96         1        14  
  2         97         1        15  
  3         95         2        15  
  3         96         2        14  
  3         97         2        20  
  4         95         2        21  
  4         96         2        27  
  4         97         2        24  
  5         95         3         9  
  5         96         3         9  
  5         97         3        12  
  6         95         3        16  
  6         96         3        17  
  6         97         3        18  

2.1 Wide to long example

. use http://www.ats.ucla.edu/stat/stata/notes3/wide, clear
describe
list
reshape long lrn, i(id) j(year)
describe
list
table year, contents(n lrn mean lrn sd lrn)

Here, we use the reshape long command with lrn as the new dependent variable. The major index (i) variable is id and the sub-index (j) variable is to be called year.

2.2 Back to wide again

. reshape wide
describe
list
reshape long
describe
list

Once you reshaped a dataset from wide to long, Stata remembers how it was done and you can change back and forth without using the index variables. In fact, once you save the file Stata will be able to change from wide to long and back at any later time.

2.3 Split-plot factorial example

. use spf2, clear
describe
list
reshape long y, i(s) j(b)
describe
list
anova y a / s|a b a*b, repeated(b)

This second example occurs frequently in experimental design situations. Data files for repeated measures designs often have all of the responses for one subject on the same line. In fact, this is the same example that we saw in the Statistics Revisited unit in the long format. The i index is the main index that defines the separate observations. In the spf design, s, the subjects id is the main index. The j index defines the secondary level of the observations. In this case, b, is the repeated factor and is represents the secondary level of the observations.

3.0 Try the commands on your own

. use wide,clear
describe
list
reshape long lrn, i(id) j(year)
describe
list

4.0 For More Information

• Statistics with Stata 5 Chapter 2 (pages 45-52)
• Stata Learning Modules
  Reshaping Data- Wide to Long
  Reshaping Data- Long to Wide

UCLA Researchers are invited to our Statistical Consulting Services
We recommend others to our list of Other Resources for Statistical Computing Help
These pages are Copyrighted (c) by UCLA Academic Technology Services