Hello Community,
I am trying to calculate model fit for a full and reduced model with the code below using the loglikelihood ratio generated from Proc glimx.
%let numParmsFull = 4;
%let numParmsReduced = 3;
%let LLfull = 4206.91
%let LLreduced = 4088.56
data _null_;
set FitStatistics;
if Model='Full' then call symput('LLfull',value);
if Model='Reduced' then call symput('LLreduced',value);
run;
data _null_;
df=&numParmsFull-&numParmsReduced;
/* numParmsFull and numParmsReduced should be the number of parameters in the full and reduced models respectively */
testStat = &LLfull - &LLreduced;
pvalue = 1 - probchisq(testStat,df);
run;
When i run the code i get this ERROR The function PROBCHISQ is unknown, or cannot be accessed.
Also, can you also help to determine model fit for two model using other method?
Can anyone help with how i can get the p-value.
Thanks,
Timothy
The function that computes the probability from a chi-square distribution is called PROBCHI, so use
pvalue = 1 - probchi(testStat,df);
In general, I avoid the PROB* functions and use the CDF function instead. So I would use
pvalue = 1 - cdf("chisq", testStat, df);
or you could use the SDF (survival) function, which has the computation "1 minus CDF" built in:
pvalue = sdf("chisq", testStat, df);
The function that computes the probability from a chi-square distribution is called PROBCHI, so use
pvalue = 1 - probchi(testStat,df);
In general, I avoid the PROB* functions and use the CDF function instead. So I would use
pvalue = 1 - cdf("chisq", testStat, df);
or you could use the SDF (survival) function, which has the computation "1 minus CDF" built in:
pvalue = sdf("chisq", testStat, df);
Thank you
Build your skills. Make connections. Enjoy creative freedom. Maybe change the world. Registration is now open through August 30th. Visit the SAS Hackathon homepage.
Register today!Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.