Margins plot with continuous variable

[Joan Marti]

Hi,

I am running a regression of the form:

reg y_var c.x_var##c.inter_var

where I interact two continous variables. I would like to create a margins plot with the following format:



Normally I would use:

margins, dydx(x_var) at(inter_var=(?))

where in "?" I would specify the levels, but I am not sure how to do it with a continous variable.

An example of my data is:

* Example generated by -dataex-. For more info, type help dataex
clear
input float(y_var x_var inter_var)
 .5714286          0  .4047619
 .4047619          0  .4558823
.59366393          0  .2712224
 .3471569  .05300756 .10421521
 .6285715          0       .11
       .5 .028987536  .2352941
 .5914634          0 .22988385
 .3675947   .0891891  .1805922
 .7055879          0 .24351773
  .456893  .07331297 .18184166
 .6387016          0 .25089714
.39703155   .1042697  .1649172
 .7121523          0 .21293853
 .4856865  .08149637  .2257634
.55119824          0  .2364341
 .3637593   .0717025 .18970737
 .6414383          0 .23691233
 .3628412  .09411559 .12715682
  .640048          0  .1985592
 .3878999  .08719668 .13754296
 .5907591          0  .2010135
 .3855799   .0924699  .2052545
.54418606          0 .27065304
 .3584131  .05004835 .13919128
 .6579425          0 .23953563
 .4383174  .09527703 .16684346
  .653834          0 .26696637
 .4039878   .0818624 .12373408
 .6480492          0  .2597146
 .3564334 .063836135 .05905572
 .6461539          0  .1011236
  .578125          0 .23333333
 .6404495          0   .206057
  .399264  .04742869 .15085714
 .6046967          0  .2084354
 .3789108  .09445875 .15795745
 .6496881          0 .27283654
 .4826884  .07605837 .16696376
.58762884          0  .1707317
 .5603448   .2177235 .26576576
 .6180905          0  .2026316
 .4987893   .0874403 .23938224
 .5366717          0 .18856515
.26013118   .0686732 .11280713
 .6523408          0 .19432606
  .425641  .08537883  .1207099
 .5263158          0 .27317074
       .3  .10584583   .184466
 .3624595          0 .21183206
 .2962963  .14848815  .2826087
 .5794066          0  .2404061
 .3498973  .07542813  .1222119
 .6213018          0 .22983015
 .4231651   .0911715 .18172537
 .6965812          0 .20095885
.56186867  .04572668  .1550357
.58810574          0 .18121693
  .449807  .09124868  .1643312
 .6179941          0  .2407199
 .4827109  .10437319 .16703217
  .535093          0 .24053966
 .3935103   .0888148 .21496554
 .6848225          0  .2345965
 .3805128  .06090141  .1110663
 .5519341          0 .26956522
  .306804   .0999722 .12772585
 .5709459          0  .1588367
 .4089347  .07745805  .2751678
 .6393443          0 .24676616
 .4420732  .08030143 .17635752
 .5981839          0  .2456326
 .4130536  .08042275  .1042913
 .6356142          0  .2326611
 .4576547   .0828319 .26104024
 .6965265          0 .23413037
 .4830468  .09650812 .17099236
 .6481481          0 .13333334
 .6057692  .09418721  .2962963
 .6789274          0  .2558149
 .4481335  .08645687  .1385835
 .6109527          0 .23182553
.34515965  .08436964 .08669516
 .5108932          0  .2336232
 .3667992   .0760269   .194958
.57123286          0 .22833665
 .3592096  .08799201 .18285504
 .6652968          0  .2174151
 .4391356   .0711862  .1184182
 .6388318          0  .2650733
.42177585  .07909735 .13423698
 .6167904          0 .20551883
 .4166256 .070240036 .14834623
 .6844171          0 .25854677
 .4452518   .1086528 .14206716
 .7015707          0 .27769986
.51666665  .04727416  .1604426
 .7541912          0 .25053576
.52555406  .09031504 .19982554
 .5801282          0  .2161235
 .3344848  .09601896   .111332
end

[Justin Niakamal]

Sounds like you want

margins, dydx(x_var) at(inter_var=(.05(.1).45))
marginsplot, recastci(rarea) recast(line)

where you just have a numlist of interesting values at(inter_var=(numlist))

[Joan Marti]

thank you very much, works perfectly!