Stata margins command (atmeans) for one standard deviation change

[Jake Smith]

Hi everyone,

I am currently writing my Bachelor thesis and need some help with the margins command. In particular, I estimate the following model:

 reg y v FI c.v#c.FI X

where my dependent variable y is ranged between 0 and 1, v is the main variable (between 1950 and 2017), FI is also ranged between 0 and 1 and X captured the covariates. In order to analyse the interaction term, I want to compute marginal effects at means. I would like to draw three scenarios:

1. v increases by one standard deviation and FI equals to 0
2. FI increases by one standard deviation and v equals to zero
3. Both variables increase by one standard deviation.

The remaining covariates should be at their mean (atmeans).

How can I solve this using the margins command?

Thanks,
Jake

[Maarten Buis]

. sysuse nlsw88, clear
(NLSW, 1988 extract)

. reg wage c.grade##c.ttl_exp i.south i.union

      Source |       SS           df       MS      Number of obs   =     1,876
-------------+----------------------------------   F(5, 1870)      =    162.30
       Model |  9866.09489         5  1973.21898   Prob > F        =    0.0000
    Residual |  22735.3797     1,870   12.157957   R-squared       =    0.3026
-------------+----------------------------------   Adj R-squared   =    0.3008
       Total |  32601.4745     1,875  17.3874531   Root MSE        =    3.4868

-----------------------------------------------------------------------------------
             wage |      Coef.   Std. Err.      t    P>|t|     [95% Conf. Interval]
------------------+----------------------------------------------------------------
            grade |   .2745394    .095182     2.88   0.004     .0878653    .4612134
          ttl_exp |  -.0517401   .0916611    -0.56   0.573    -.2315089    .1280287
                  |
c.grade#c.ttl_exp |   .0251362   .0071125     3.53   0.000      .011187    .0390855
                  |
          1.south |  -1.014563   .1653533    -6.14   0.000    -1.338859   -.6902662
                  |
            union |
           union  |   .8154901   .1900031     4.29   0.000     .4428497    1.188131
            _cons |   .5527386   1.208343     0.46   0.647    -1.817104    2.922581
-----------------------------------------------------------------------------------

.
. sum grade

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
       grade |      2,244    13.09893    2.521246          0         18

. local gr1 = r(mean)

. local gr2 = r(mean) + r(sd)

. sum ttl_exp

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
     ttl_exp |      2,246    12.53498    4.610208   .1153846   28.88461

. local exp1 = r(mean)

. local exp2 = r(mean) + r(sd)

. margins, at(grade=(`gr1' `gr2') ttl_exp=(`exp1' `exp2') south = 0 union=0)

Adjusted predictions                            Number of obs     =      1,876
Model VCE    : OLS

Expression   : Linear prediction, predict()

1._at        : grade           =    13.09893
               ttl_exp         =    12.53498
               south           =           0
               union           =           0

2._at        : grade           =    13.09893
               ttl_exp         =    17.14518
               south           =           0
               union           =           0

3._at        : grade           =    15.62018
               ttl_exp         =    12.53498
               south           =           0
               union           =           0

4._at        : grade           =    15.62018
               ttl_exp         =    17.14518
               south           =           0
               union           =           0

------------------------------------------------------------------------------
             |            Delta-method
             |     Margin   Std. Err.      t    P>|t|     [95% Conf. Interval]
-------------+----------------------------------------------------------------
         _at |
          1  |   7.627585   .1215531    62.75   0.000     7.389191    7.865979
          2  |   8.906998   .1452138    61.34   0.000       8.6222    9.191796
          3  |   9.114166   .1440408    63.27   0.000     8.831668    9.396663
          4  |   10.68575   .1766729    60.48   0.000     10.33925    11.03224
------------------------------------------------------------------------------

[Jake Smith]

Hi Maarten,

thank you so much for your respond. It works!

Best,
Jake

[Liselotte Nel]

Hi Maarten,

I run the same code as you describe but I get a error "invalid numlist has too few elements". Do you know how to solve it?