|
Adam:
At least you are giving careful thought to discrepancies between the model's parameter estimates and a priori expectations of odds. I'd advise that you switch to GENMOD with the DESCENDING option. Your attempts at recoding and specifying ordinal and continuous variables appear to be a distraction. GENMOD's variant of logistic regression tolerates a wider range of data types and should give you estimates very close to those of a correctly-specified LOGISTIC regression.
You may also want to take a look at one of the continuing examples in Hastie et al's Statistical Learning text. They carry over an example of a SA heart disease study that 'features' parameter estimates with the wrong sign.
It takes some time to verify that your model specification reflects what you think that you are estimating. Before GENMOD or LOGISTIC, I wrote a SESUG paper about using 2X2 tables to estimate single predictor parameters of CATMOD models.
Correct specifications will not necessarily eliminate 'wrong signs' or give you meaningful estimates. Search for Cassell in the SAS-L archives to learn more than you want to know about how Statistical Models Go Wrong.
Sig
________________________________
From: owner-sas-l@listserv.uga.edu on behalf of Adam
Sent: Mon 10/10/2005 11:53 AM
To: sas-l@uga.edu
Subject: Re: Dummy variables & proc logistic - what am I missing?
Yes, sorry about that - I was switching back and forth between "ac6"
and "stroke"...here are the results when "stroke" is replaced with
"ac6" in the model (and I ran it both ways again just to make sure; the
parameter estimates for the predictor variables come back the same
whether stroke is coded in ascending or descending order)...sorry for
the confusion...here are the results for with "stroke" instead of
"ac6":
The LOGISTIC Procedure
Analysis of Maximum Likelihood Estimates
Standard Wald
Parameter DF Estimate Error Chi-Square
Pr > ChiSq
Intercept 1 4.0015 0.0256 24358.2853
<.0001
lowincome 1 -0.1333 0.00389 1174.3857
<.0001
SRAGE_P 1 -0.0471 0.000304 24077.9639
<.0001
SRSEX 1 -0.0604 0.00396 232.3938
<.0001
AB29 1 0.7180 0.00434 27401.5473
<.0001
diabetes 1 -0.3277 0.00439 5562.5470
<.0001
AB34 1 0.7701 0.00388 39346.4711
<.0001
AE13 1 0.0856 0.00141 3675.4610
<.0001
smokes 1 -0.8101 0.00488 27585.4492
<.0001
Odds Ratio Estimates
Point 95% Wald
Effect Estimate Confidence Limits
lowincome 0.875 0.869
0.882
SRAGE_P 0.954 0.953
0.955
SRSEX 0.941 0.934
0.949
AB29 2.050 2.033
2.068
diabetes 0.721 0.714
0.727
AB34 2.160 2.144
2.176
AE13 1.089 1.086
1.092
smokes 0.445 0.441
0.449
Association of Predicted Probabilities and Observed
Responses
Percent Concordant 68.9 Somers' D
0.389
Percent Discordant 30.0 Gamma
0.393
Percent Tied 1.1 Tau-a
0.061
Pairs 5931331 c
0.695
| 
