hen666
07-21-2009, 08:43 AM
Hello
I am estimating a three level logistic model (city, neighborhood and individual) with Alternating Logistic Regression.
I am using the Genmod procedure and specifying the Log Odds Ratio Model with NEST1, using neighborhoods as subclusters.
In the SAS output it is stated that Alpha1 represents the log odds ratio for within subclusters while alpha2 represents he log odds ratio between subclusters.
Alpha1 represents the clustering within neighborhoods while Alpha2 represent the clustering within cities - correct?
However, studying the model it is stated that log {PWOR(Yijk, Yilm)} = a0+a1Zijklm - this means that within city log odds ratio ought to be represented by Alpha0 and within neighborhood to be represented by Alpha0+Alpha1? Is it just the SAS output that confuses me, or am I doing something wrong?
Best
Lisa
I am estimating a three level logistic model (city, neighborhood and individual) with Alternating Logistic Regression.
I am using the Genmod procedure and specifying the Log Odds Ratio Model with NEST1, using neighborhoods as subclusters.
In the SAS output it is stated that Alpha1 represents the log odds ratio for within subclusters while alpha2 represents he log odds ratio between subclusters.
Alpha1 represents the clustering within neighborhoods while Alpha2 represent the clustering within cities - correct?
However, studying the model it is stated that log {PWOR(Yijk, Yilm)} = a0+a1Zijklm - this means that within city log odds ratio ought to be represented by Alpha0 and within neighborhood to be represented by Alpha0+Alpha1? Is it just the SAS output that confuses me, or am I doing something wrong?
Best
Lisa