tabgen

Generate indicator variables for each value of a categorical variable.

Installing via net install

The current version is still a work in progress. To install, user can use the net install command to download from the project’s Github page:

net install tabgen, from("https://aarondwolf.github.io/tabgen")

Syntax

    tabgen varname [if] [, generate(stubname)]

Description

tabgen creates a new indicator for each unique value/label pair for a labelled categorical variable.

tabgen is very similar to tabulate varname, gen(stubname). The main difference is that tabgen ensures that the value at the end of stubname represents the numerical value for that label, rather than assigning the subnames the values 1, 2, 3, etc.

Options

• generate(stubname) a stub that Stata will use when generating new indicator variables. If no stubname is specified, tabgen will use varname_ by default.

Notes

tabgen will replace all negative (‘-‘) signs with an underscore (‘_’) for variable names when generating new variables. For example, if your variable has a label for a values of -99 (for example, “Don’t Know”), the indicator for that values would be stubname_99, which is 1 everywhere varname == -99.

tabgen requires Daniel Klein’s labellist in order to run (ssc install labellist). If labellist is not installed, tabgen will install it first before proceeding.

Examples

Load auto data:

    .  sysuse auto, clear

Generate indicator variables for the foreign variable:

  .  tabgen foreign

You can optionally specify a different stub:

   .  tabgen foreign, generate(new_)

Stored results

tabgen stores the following in r():

Scalars

• r(varlist) - The variable name specified.
• r(varlabel) - The variable label for the variable specified.
• r(lblname) - The value label for the variable specified.
• r(values) - The list of labelled values for the variable specified.
• r(labels) - The list of labels for the variable specified.

Authors

Aaron Wolf, Northwestern University aaron.wolf@u.northwestern.edu