Dear friends,
If we can not control the levels of a factor which is usually defined a priori such as "time" (e.g. time 1, 2 ,3 and 4), should we consider it a random factor?
Example: We've got an intervention which the number of levels of the factor "time" depends upon the subject's skills. An experienced subject may reach 12 levels while an unexperienced likely stops at the level number four. In such case, "time"should be consider random factor? And how to deal with the missing values?
Best regards,
Paulo