-------------------------------------------------------------------------------
Help for ttestplus
-------------------------------------------------------------------------------

Description

ttestplus automates and tabulates t-tests on one or more dimensions across two
populations indicated by a categorical variable. It can also cut a continuous
variable at a specified value to serve as a categorical.

Alternatively, ttestplus automates and tabulates t-tests on one analysis
dimension across a population with characteristics defined by two or more
binary categories.

Syntax

ttestplus varlist [if] [in], 
        by(groupvar | varlist) 
        [cut(value | mean | median)]
        [cluster(groupvar)]
        [t] [se]

    Options               Description
    -------------------------------------------------------------------------
    * by(groupvar)        Variable defining groups for comparison by varlist.
    * by(varlist)         List of binary variables for comparison. Cannot be
                            combined with multiple analysis dimensions or
                            with cut.
    cut()                 Cuts a continuous groupvar at the specified value,
                            at its mean, or at its median.
    cluster               Cluster results using the specified categorical
                            variable. Requires the most recent version of 
                            clttest.
    t                     Returns t-statistics instead of p-statistics.
    se                    Returns standard errors below group means.
    -------------------------------------------------------------------------
    * Exactly one version of by() is required.

Examples

. sysuse nlsw88
(NLSW, 1988 extract)

. ttestplus married grade wage, by(collgrad)

             |   Group 1    Group 2     p-Stat 
-------------+---------------------------------
     married |  .6406068   .6466165   .4003468 
       grade |  11.97371   16.71992          0 
        wage |  6.910561   10.52606   2.63e-38 

. ttestplus married grade wage, by(hours) cut(median) se

hours cut at 40 (median)

             |   Group 1    Group 2     p-Stat 
-------------+---------------------------------
     married |  .7220026   .6021578   9.53e-09 
          SE |  .0162725   .0127141          . 
       grade |  12.88801   13.20999     .00209 
          SE |  .0869087   .0669262          . 
        wage |  6.492085   8.430878   1.58e-14 
          SE |  .1552967   .1631842          . 

. ttestplus married grade wage, by(hours) cut(median) cl(race) se
 
hours cut at 40 (median)
Standard errors clustered by race

             |   Group 1    Group 2     p-Stat 
-------------+---------------------------------
     married |  .7220026   .6021578   .2720061 
          SE |   .131747   .1240575          . 
       grade |  12.88801   13.20999   .3269942 
          SE |  .4856302   .4555908          . 
        wage |  6.492085   8.430878   .0790578 
          SE |  .8183053   .7628162          . 

. ttestplus wage, by(collgrad married union)

             |   Group 1    Group 2     p-Stat 
-------------+---------------------------------
    collgrad |  6.910561   10.52606   2.63e-38 
     married |  8.080765   7.591978   .0268258 
       union |  7.204669   8.674294   1.91e-11

. ttestplus wage, by(collgrad married union) cl(race)
 
Standard errors clustered by race

             |   Group 1    Group 2     p-Stat 
-------------+---------------------------------
    collgrad |  6.910561   10.52606   .0120197 
     married |  8.080766   7.591978   .3471417 
       union |  7.204669   8.674294   .1185467 


Saved Results

ttestplus saves its output in the matrix results.

Author

Benjamin Daniels
bbdaniels@gmail.com

Acknowledgements

Thanks to Jeph Herrin, author of cltest, for updating his program for
compatability and suggesting a correction for stratified data.