rescaling a variable difficulty

[Alan Jeddi]

Hi all.
I have a continuous variable called GLA ranging from 0.10 to 150.0 (and in some cases it was not measured in participants, so was coded as . for those participants)
I would like to rescale this so it is in quartiles
I have been having difficulty, as for some reason my code ends up replacing data that is actually missing (.) with a value of 4.

Here is my syntax: any help at all would be amazing!

gen gla_quart = .
replace gla_quart = 1 if gla>=0.1 & gla<=1.0
replace gla_quart = 2 if gla>=1.1 & gla<3.0
replace gla_quart = 3 if gla>=3.1 & gla<10.0
replace gla_quart = 4 if gla>10.1

[Marcos Almeida]

Welcome to the Stata Forum / Statalist.

Please do read the FAQ. There you'll find advice about posting data/command/output.

That said, you may wish to try this:

xtile myvar = GLA, nq(4)

[Nick Cox]

Note that 10.1 is to Stata less than missing. See e.g.

FAQ . . . . . . . . . . . . . . . . Logical expressions and missing values
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . W. Gould
2/03 Why is x > 1000 true when x contains missing value?
http://www.stata.com/support/faqs/data-management/
logical-expressions-and-missing-values/

With the original code, what about values between 1.0 and 1.1, and so forth?

Consider this way of thinking about the problem:

gen gla_quart = cond(gla <= 1, 1, cond(gla < 3,  2,  cond(gla < 10, 3,  4))) if gla < .

That way, the if condition peels off observations with missing on gla and the others are classified with the rules

if <= 1 return 1
otherwise if < 3 return 2
otherwise if <10 return 3
otherwise return 4.

The recipe is the important detail here. Feel free to change < to <= or vice versa and cut-off values according to circumstance.