meinequality

Inequality summary effects of nominal and ordinal independent variables

meinequality (Marginal Effects Inequality) is a companion package to Mize and Han’s 2025 Sociological Science article, “Inequality and Total Effect Summary Measures for Nominal and Ordinal Variables.” A nominal independent variable with L categories has L(L-1)/2 pairwise marginal effects – one for every pair of categories. meinequality summarizes them in a single ME inequality statistic: the average absolute pairwise difference in predictions across the categories, with a standard error. It answers “how much does the outcome differ across the levels of this variable, overall?” and it can be compared across predictors, across models, and across groups.

The command calculates weighted ME inequality (the default), which accounts for the relative size of each category in the sample, and/or unweighted ME inequality, which gives every pairwise comparison the same weight. It works after most regression models, for one model or for two; with two models the difference in ME inequality across the models is tested.

Installation

net install meinequality, from("https://tdmize.github.io/data") replace

meinequality requires the suest2 package and Stata 16 or later.

Where to start

  • Options goes through every option with an example of each.
  • The Examples in the sidebar reproduce the examples from Mize and Han (2025): inequality as a summary measure, inequality in categorical models, and comparisons of ME inequality across models and across groups.
  • Examples from the help file runs the help file’s examples, and the help file is on the site.
  • The same summary is available as the meinequality option of mecompare, which places it above the pairwise contrasts it summarizes.
  • To calculate it in R, see ME inequality in R.

Basic syntax

For a single model, fit it with the nominal variable entered as a factor variable (i.), then call meinequality with that variable:

regress dv i.nominalvar iv2, vce(robust)
meinequality i.nominalvar

For two models, store them and name both in models():

logit dv i.nominalvar iv2, vce(robust)
est store basemod

logit dv i.nominalvar iv2 med1 med2, vce(robust)
est store medmod

meinequality i.nominalvar, models(basemod medmod)

For two models fit on separate samples, add groups:

meinequality i.nominalvar, models(grp1mod grp2mod) groups

A first example

sysuse nlsw88, clear
(NLSW, 1988 extract)
drop if missing(union)
(368 observations deleted)

. 
quietly logit union i.race i.collgrad c.age i.south, vce(robust)
. 
meinequality race
ME Inequality Estimates (N = 1878)

                         |   Estimate   Std. err.           z       P>|z| 
-------------------------+-----------------------------------------------
ME Inequality            |                                               
                    race |      0.106       0.025       4.225       0.000

With no models() option the summary is computed for the model in memory; with two stored models in models() it is computed for each and the difference is tested.

Citation

Please cite the use of meinequality by citing the corresponding article:

Mize, Trenton D. and Bing Han. 2025. “Inequality and Total Effect Summary Measures for Nominal and Ordinal Variables.” Sociological Science 12: 115–157.

Authors

meinequality and its sister command totalme are written by Bing Han (Population Research Institute, Penn State University) and Trenton D. Mize (Purdue University).

Back to top