[size=4]It is inconvenient to run BE studies with a large number of subjects so often the subjects are split into two groups. I created a data set to evaluate in Phoenix. I have a question about the code you need and how you enter it. The code shown below for fixed effects and random comes from Helmut so it is most probably correct fixed: Group+Sequence+Sequence(Group)+Period(Group)+Treatment random: Subject(Sequence*Group) [/size] I have tried to get the Program to accept the code, but the text turns red. I have tried several options for the check boxes, but was not successful in getting the program to accept the code. I attach the file, Any thought are welcome, Post script I have now got the file to run using group as a classification variable and including subject in set up mapping (see link below) http://screencast.com/t/EpJVIvdBr Also new file is attached name=LARGE_STUDY_DATA_WITH_GROUPS.phxproj size=155639]Certara | Drug Development Solutions (152 KB)LARGE_STUDY_DATA.phxproj (67.6 KB)
Hi Angus, I just had a chance to look at this, trying it with my own data and I didn’t have a problem. Glad to see that you got it worked out. Regards, Linda Hughes
Linda; Thanks for the note yes; it works well. The fly in the ointment with this model appears to be the risk of false positives. I did not realize that GROUP is a CLASSIFICATION variable. Your own data would be better than my data: it seems that I have more problems with faking data as is opposed to using genuine data, Angus