topic Re: Count variables analysis in Statistical Procedures

Count variables analysis

Miracle — Fri, 28 Feb 2014 04:05:23 GMT

Dear all,

Can I please ask how do I go about analyse a data with 3 variables as such

- a 2-block time variable with nth categories ie 12am-2am, 2.01am-4.00am and etc (time).

- 2 count variables, one being the total number of calls made for the ith time category(n) and another one being the cases resolved for the ith time category(x).

I wish to find out the adjusted proportion of cases resolved for each time categories taking into account the number of calls made at the same time. I would think the more calls made the more cases resolved and more weighting for that time slot. Is my proc genmod below correct? Or am i completely off the right track? Any input is greatly appreciated.

proc genmod data=foo;

class time/ param=ref;

model x=time / offset=log_n dist=poisson link=log;

exact time / joint estimate;

run;

Re: Count variables analysis

StatDave — Fri, 28 Feb 2014 15:58:42 GMT

It sounds to me like this is a simple logistic model applied to aggregated data. So, you could use PROC LOGISTIC and events/trials syntax as below. Unless the data set is small and sparse, the EXACT statement isn't necessary and would probably take too much time.

proc logistic;

class time;

model x/n = time;

run;

Re: Count variables analysis

SteveDenham — Fri, 28 Feb 2014 18:16:33 GMT

I agree, especially if the denominator is large. However, if the denominator is less than 50, then the genmod code would be my preference.

Steve Denham