logitcprplot or alternative illustration?

[Samantha Mosb]

Dear all,
I am struggling with a graphical illustration for the following situation:

• "newvar": binary outcome variable indicating whether someone buys a new car or not.
• "incentive": variable of interest is a summary measure that captures different aspects (finances, needs etc.) influencing the purchase decision.

I want to find out whether this incentive variable is adequate to explain the "newcar" variable. Moreover I want to find out whether this differs by age and gender. In other words - do females respond to the incentives more accurately than males?
For a first impression I would like to have a graphical illustration. I tried the following, but I am not sure whether this is the best way for my intention:

logit newcar incentive
logitcprplot incentive, rcspline(ci) by(female)

Do you have any other ideas for a better illustration?

Thanks and have a great weekend,
Sam


I am using Stata 14, version 1.0.4 28jan2009 of logitcprplot. Here is some example data:

* Example generated by -dataex-. To install: ssc install dataex
clear
input byte(ID newcar) double incentive byte age float female
  3 0   .03972538035415816 28 1
  4 0   .16682683928769018 33 1
  5 0 .0012173821879535836 31 1
  6 0   .20561523090611494 39 1
  7 0   .13439488561921117 28 0
  8 0   .09112885675571442 30 0
  9 0   .04690987263185393 26 1
 10 0   .04690987263185393 26 1
 11 0   .08978829754195784 31 0
 12 0   .12111005208256041 26 0
 14 1   .04059250267250229 37 1
 15 1   .20676567212001254 37 0
 17 0   .32137271478198465 38 1
 18 0   .02069449099176917 29 1
 20 0 .0012173821879535836 31 1
 21 1   .16682683928769018 33 1
 22 1   .02126607470428645 37 1
 23 1   .26704398470400553 33 0
 24 1    .1715611891895402 36 0
 25 1   .11721748713877664 31 0
 26 1   .32483437782415053 33 0
 29 1    .3885486604017397 39 0
 32 1   .32137271478198465 38 1
 33 0   .04059250267250229 37 1
 34 0   .23438664769315412 34 1
 35 1   .14764137688628484 35 0
 36 1    .3433477964308977 38 0
 37 0   .11965222819764557 30 0
 38 0  .011503910217765471 32 0
 39 0    .0457394323233342 27 1
 40 0    .0537767431407912 26 1
 41 0   .11191945729808164 25 0
 42 1   .13861102790755997 27 0
 43 1   .13133017107293066 29 0
 44 0  .043554096752659664 25 1
 45 0  .052776501982149414 27 1
 46 1   .07428374347310765 35 1
 47 0   .08841490098710833 33 1
 48 0   .04690987263185393 26 1
 49 0   .11662819014695244 26 0
 50 0   .13133017107293066 29 0
 51 1   .32137271478198465 38 1
 52 1   .21262513175234668 35 0
 53 1   .25588355968243515 39 0
 54 1   .20129658769110298 36 0
 55 1    .0443966984639766 30 0
 56 0    .1023023256345485 28 0
 57 1   .07428374347310765 35 1
 59 1   .08841490098710833 33 1
 60 0   .03268199621188424 27 1
 61 0   .10743767074877218 26 0
 62 0    .0457394323233342 27 1
 63 0  .055220528743400873 25 1
 64 0   .10040844108317014 29 0
 65 1    .2740306391242877 38 0
 66 1   .19683912030636308 36 0
 67 1   .03686372394791417 32 0
 70 1   .02126607470428645 37 1
 71 1    .4323128507444171 39 1
 72 0  .022917451373031236 28 1
 74 0   .07753927869910902 34 1
 75 0    .0537767431407912 26 1
 76 0   .05017170816564163 29 1
 78 0    .1715611891895402 36 0
 79 0   .26182369323299726 34 0
 80 0   .12957276106988652 28 0
 81 0   .11662819014695244 26 0
 82 1   .09861978330573328 27 0
 83 1   .30891493628814753 39 0
 84 0   .05017170816564163 29 1
 85 0    .0537767431407912 26 1
 86 0     .043881497903992 28 1
 87 1   .09112885675571442 30 0
 88 1   .13439488561921117 28 0
 90 0   .03814587588655128 29 1
 92 1   .22506447668557927 38 0
 93 0   .10471420373182062 29 0
 94 1   .20129658769110298 36 0
 95 1   .35704698117849354 34 0
 96 0   .10356424623493078 27 0
 97 0   .07753927869910902 34 1
 99 1   .30891493628814753 39 0
100 1   .26704398470400553 33 0
end

[Marcos Almeida]

Thank you for sharing data and command correctly. It helps a lot.

I wonder whether you wish something like this:

.logit newcar incentive age i.female
.margins female, dydx(age) at(incentive=(0.0012(0.01)0.44))
.marginsplot

Or maybe this way:

.margins female, dydx(incentive) at(age=(25(1)40))
.marginsplot

Hopefully that helps.