topic Proc surveylogistic_help in SAS Programming

Proc surveylogistic_help

joon1 — Fri, 15 Nov 2019 02:30:48 GMT

Hi Madam/Sir,

In the SAS 9.2, the output of "proc surveylogistic" generates wald chi-square statistics. But in the SAS 9.4, it generates t-statistics.

In the SAS 9.4, proc logistic still generates wald ch-square statistics. But I need to use proc surveylogstic to do firm-level clustering in the standard variation of variables. Is there a way to report wald chi-square statistics? Your help will be highly appreciated. Joon.

proc surveylogistic data=s19;
class gvkey;
model gc(descending) = sdd1 starmeasure lev chlev lnage return beta sd_residual zscore lnasset
ploss investment ffinancing ocf reportlag big4 y2001-y2016 id1-id65/stb;
run;

Analysis of Maximum Likelihood EstimatesParameter Estimate StandardError t Value Pr > |t| StandardizedEstimateInterceptsdd1starmeasurelevchlev

-6.6960	0.9303	-7.20	<.0001
1.3863	0.2235	6.20	<.0001	0.2345
1.0422	0.2931	3.56	0.0004	0.2327
0.6787	0.2085	3.26	0.0011	0.1338
0.5872	0.3805	1.54	0.1229	0.0683

Re: Proc surveylogistic_help

unison — Fri, 15 Nov 2019 02:33:39 GMT

PROC SURVEYFREQ?

https://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#statug_surveyfreq_a0000000223.htm

Re: Proc surveylogistic_help

joon1 — Fri, 15 Nov 2019 05:33:23 GMT

Hi Unison,

Thank you for your prompt response
I had a following error message. Could you please help me out to fix this
error?
1427 proc surveylogistic data=s19 ;
1428 tables wchisq;
------
180
ERROR 180-322: Statement is not valid or it is used out of proper order.
1429 class gvkey;
1430 model gc(descending) = sdd1 starmeasure lev chlev lnage return beta
sd_residual zscore
1430! lnasset
1431 ploss investment ffinancing ocf reportlag big4 y2001-y2016
id1-id65/wchisq;
1432 run;

Thanks again

Sincerely,
Joon1

Re: Proc surveylogistic_help

Kurt_Bremser — Fri, 15 Nov 2019 07:47:07 GMT

tables is not a valid statement in proc surceylogistic. I suggest you also study the documentation for the MODEL statement, which includes:

Note that unless you specify the DF=INFINITY option, PROC SURVEYLOGISTIC displays an F test instead of a chi-square test.