How to label values of independent variables in logistic regression output?

[AbuYusuf]

Hi,

I have categorical variables in the logistic regression and I'd like to label their values in output, so instead of having 1,2,3 etc. I'd like to have names of ethnic groups. Is there a way to do that?

Thank you!

Analysis of Maximum Likelihood Estimates
Parameter	DF	Estimate	Standard	Wald	Pr > ChiSq
				Error	Chi-Square
Intercept		1	-30.9965	220.3	0.0198	0.8881
race	1	1	13.295	186.4	0.0051	0.9431
race	2	1	-3.7821	392.2	0.0001	0.9923
race	3	1	-2.33	514.4	0	0.9964
race	4	1	-2.1724	472.6	0	0.9963
race	5	1	-1.5253	725.1	0	0.9983
race	6	1	-1.9817	580.2	0	0.9973

[PaigeMiller · Accepted solution]

Use a format, something like this:

proc format;
     value $racef. '1'='Martian' '2'='Venusian' '3'='Earthling';
run;

proc logistic data=whatever;
    ...
    format race $racef.;
run;

[AbuYusuf]

Great! Thank you very much!

[PaigeMiller]

Correcting a typographical error — no dot after $racef in PROC FORMAT

proc format;
     value $racef '1'='Martian' '2'='Venusian' '3'='Earthling';
run;

proc logistic data=whatever;
    ...
    format race $racef.;
run;

[AbuYusuf]

Sorry for more trouble, but I have a follow up question. Format worked just fine, in terms of formatting the output. The problem is that when I run the code with format it drops some estimates. Is there a reason why it's doing it?

Here is output without format:

Analysis of Maximum Likelihood Estimates
Parameter	DF	Estimate	Standard	Wald	Pr > ChiSq
				Error	Chi-Square
Intercept		1	-30.9965	220.3	0.0198	0.8881
race	1	1	13.295	186.4	0.0051	0.9431
race	2	1	-3.7821	392.2	0.0001	0.9923
race	3	1	-2.33	514.4	0	0.9964
race	4	1	-2.1724	472.6	0	0.9963
race	5	1	-1.5253	725.1	0	0.9983
race	6	1	-1.9817	580.2	0	0.9973
race	7	1	-2.2975	576.2	0	0.9968
race	8	1	-1.4041	709.9	0	0.9984
race	9	1	-2.2255	849	0	0.9979
race	10	1	-2.6247	567.1	0	0.9963
race	11	1	-3.034	1310.9	0	0.9982
race	12	1	13.7587	186.4	0.0054	0.9412

And here is the output when I add format:

Analysis of Maximum Likelihood Estimates
Parameter	DF	Estimate	Standard	Wald	Pr > ChiSq
				Error	Chi-Square
Intercept		1	-30.2701	918.8	0.0011	0.9737
race	Aboriginal	1	13.5539	115.8	0.0137	0.9068
race	Arab	1	-2.1409	416.6	0	0.9959
race	Black	1	-2.2589	382.7	0	0.9953
race	Chinese	0	0	.	.	.
race	Filipino	0	0	.	.	.
race	Japanese	0	0	.	.	.
race	Korean	0	0	.	.	.
race	Latin American	0	0	.	.	.
race	Other	1	-2.7809	271.4	0.0001	0.9918
race	South Asian	1	-3.7314	300.6	0.0002	0.9901
race	Southeast Asian	0	0	.	.	.
race	West Asian	0	0	.	.	.

thank you,

i.

[PaigeMiller]

We need to see the code you used, and (a portion of) the RAW data.