Changing the Y axis scale on a twoway graph

[Rajdeep Chaudhuri]

Hi all,
I am trying to graph raw trends to show parallel trends for a DiD estimation. Now for the unadjusted trends, I do not need to run the regression, so I was using separate and graph to generate the parallel trend graphs. However, since the point estimates of the excercise is very small, it does not visibly show up on the graph. My understanding is that if I rescale the Y axis, I might be able to show the changes. Is there a way to rescale the axis. I used

preserve
separate drinkalcohol, by(Bihar)
collapse (mean) drinkalcohol? , by(Bihar Year)
graph twoway line drinkalcohol? Year, sort
restore

to generate the graphs.
Here Bihar is basically the group assignment and Year is the 3 years for which data was collected. In fact 1 stands for 2003, 2 stands for 2014 and 3 stands for 2019. so my typical regression for the DiD would be

reg drinkalcohol i.Bihar##i.Post i.Year

Post is 1 for Year = 3 and 0 otherwise
Here's a data example

* Example generated by -dataex-. For more info, type help dataex
clear
input float(Bihar Year) byte drinkalcohol
0 3 0
0 3 0
1 2 0
0 1 0
0 3 0
1 2 0
0 2 0
0 1 0
0 3 0
1 3 0
0 1 0
0 1 0
0 3 0
0 2 0
1 3 0
0 3 0
0 2 0
1 2 0
1 3 0
0 1 0
1 3 0
0 3 0
1 3 0
0 1 0
0 3 0
0 2 0
1 3 0
1 3 0
0 3 0
0 1 0
0 2 0
0 2 0
1 3 0
0 3 0
0 2 0
0 2 0
0 3 0
0 3 0
1 3 0
0 3 0
0 3 0
1 3 0
0 2 0
0 2 0
0 2 0
0 1 0
0 2 0
0 2 0
0 2 0
1 2 0
0 2 0
0 1 0
0 2 0
0 2 0
0 2 0
1 2 0
0 2 0
0 1 0
1 1 0
0 2 0
0 3 0
0 2 0
0 2 0
0 2 0
0 2 0
0 2 0
1 2 0
0 3 0
0 1 0
0 3 0
0 2 0
0 2 0
0 3 0
0 3 0
1 3 0
0 1 1
0 3 0
1 2 0
1 1 0
1 2 0
0 1 0
0 2 0
1 3 0
0 2 0
1 2 0
1 2 0
1 2 0
1 2 0
0 3 0
1 3 0
0 3 0
0 3 0
0 1 0
0 2 0
1 2 0
0 2 1
0 1 0
0 2 0
1 2 0
0 1 1
end
label values drinkalcohol S720
label def S720 0 "no", modify
label def S720 1 "yes", modify

[Nick Cox]

In your data example, one variable's means are always 0, so the scope for changing the scale usefully seems limited.

In your full dataset, if all means are positive and some are very small, changing to logit scale could be a good idea. The summaries concerned are proportions after all. (What that implies for your methodology using plain regression I take to be a larger question.)

More discussion at

SJ-8-1 gr0032 . . . . . . . Stata tip 59: Plotting on any transformed scale
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N. J. Cox
Q1/08 SJ 8(1):142--145 (no commands)
tip on how to graph data on a transformed scale

Stata Tip 59: Plotting on Any Transformed Scale

[Rajdeep Chaudhuri]

Hi Nick, thanks for the response
I tried using

gen logit_drinkalcohol =  logit(drinkalcohol)

but stata gives me

. gen logit_drinkalcohol =  logit(drinkalcohol)
drinkalcohol ambiguous abbreviation
r(111);

[Nick Cox]

The function logit() may be useful if and only if applied to values between 0 and 1, but not 0 and 1. Otherwise put, I am suggesting it only for means greater than 0 and less than 1.

Otherwise your code is failing for the reason stated: two or more variable names abbreviate ambiguously to the name you give,