zindex was developed by the Global Poverty Research Lab at Northwestern University, with support from Innovations for Poverty Action.

*zindex can be installed directly from Github
net install zindex, from("https://raw.githubusercontent.com/PovertyAction/zindex/master") replace 

* or using haghish's Github command
github install PovertyAction/zindex

Title

    zindex -- Generates an index using z-scores

Syntax

        zindex varlist [if] [in], gen(varname) [options]


    options               Description
    -------------------------------------------------------------------------
    Main

    * gen(varname)        Name of index variable to be created

    Optional

    * mean(varlist)       Variables that will be used to standardize against
                            (e.g. baseline versions of main varlist)

    * condition(string)   A condition to specify a sub-group to standardize
                            against (e.g. control group)


Description

    zindex creates a z-score index. The resulting index is the z-score of the
    mean of each index components' z-scores.  Steps of index creation:

    1.     Generate z-scores for each variable included in the index

    2.     Generate means of these z-scores

    3.     Create index by generating the z-score for the means of these z-scores

        The mean() option allows you to specify a list of variables to
            standardize against. The means and SDs of this varlist will be
            used in the initial z-scores (in step 1 above) instead of the
            means and SDs of the main varlist specified for the index. The
            size of this varlist must be the same size as the main varlist
            specified for index. A common input is a list of the baseline
            versions of the index variables to standardize against the
            baseline values.

        The condition() option allows you to specify a subgroup to
            standardize against.  The condition is applied when taking the
            mean and SD of each variable going into the index (in step 1
            above) and when taking the mean and SD of the means of the
            z-scores (in step 3 above).  A common condition is "treatment ==
            0" to standardize against the control group.

        [if] [in] restricts the sample used for ALL STEPS to the observations
            that meet the if/in condition. If {opt:condition()} is specified,
            only the observations that fulfill both {opt:condition()} and the
            if/in condition is used to calculate the mean and standard
            deviation for standardization.


Examples

         sysuse citytemp.dta 

     Create an index of heat, not standardized against a different round or
    subgroup 

         zindex tempjan tempjuly heatdd, gen(hot_weather)

    Standardize against the Northeast

         zindex tempjan tempjuly heatdd, gen(hot_weather_NE) condition(region == 1)
        
    Calculate the index only for the Northeast. Note that hot_weather_NE and
    hot_weather_NE_only are the same for all cities in the Northeast, but
    hot_weather_NE_only does not have any values for cities outside of the
    Northeast.

        zindex tempjan tempjuly heatdd if region == 1, gen(hot_weather_NE_only)

    Standardize against January temperatures and days of cold

         zindex tempjuly heatdd, gen(hot_weather_NE) mean(tempjan cooldd)

Stored results

    zindex stores the following in r():

    Scalars   
      r(m_VAR)       number of missing observations in VAR (one return value
                       per variable in the index)
   


Authors

    Caton Brewster
    Codecheck: Kelsey Larson

    For questions or suggestions, submit a GitHub issue or e-mail researchsupport@poverty-action.org.