Grouping Variable by combining two metric variables

[Philipp Schleicher]

Hi there,

I'd like to create a grouping variable "category" from a combination of two other metric variables (let's say "var1" and "var2").
The categories shall be:

category 1 = var1<50 AND var2<25
category 2 = var1<50 AND var2>25
category 3 = var1>50 AND var2<25
category 4 = var1>50 AND var2>25

How do I do this in STATA?
I tried to use "egen" and "cut", but this works only with one source variable, doesn't?

Thanks in advance,

Philipp

[Carlo Lazzaro]

Philip:
welcome to this forum.
Perhaps it is s -group- from -egen- what you're looking for.

[Guest]

try using the "gen" command. example:

clear all
sysuse auto

gen category = .
replace category = 1 if price < 5000 & weight < 2500
replace category = 2 if price < 5000 & weight > 2500
replace category = 3 if price > 5000 & weight < 2500
replace category = 4 if price > 5000 & weight > 2500

[Philipp Schleicher]

Thank you a lot, Carlo!
It worked... :-)))

[Philipp Schleicher]

Guest - Thank you, too! This looks nice!

[Nick Cox]

As a twist on #3 from Guest note also

sysuse auto, clear

gen category = .
replace category = 1 if price < 5000 & weight < 2500
replace category = 2 if price < 5000 & weight > 2500
replace category = 3 if price > 5000 & weight < 2500
replace category = 4 if price > 5000 & weight > 2500

gen category2 = 1 + (weight >= 2500) + 2 * (price >= 5000)

assert category == category2

The new code takes care of one possible nuance, values exactly equal to the thresholds.

In addition, you need an extra qualifier

... if !missing(weight, price)

if there are missing values.