I’m trying to track cohort survival through a training program. One issue I am struggling with is the impact of the cohort size on how that displays in the data. The course is quite demanding so there are always going to be losses.
for example:
cohort 1 has 100 students and at a point, 1 student fails which makes the survival 99% at that point
cohort 2 has 10 students and at the same point, 1 student fails which makes the survival 90%
so this is being read that the course for cohort 1 was 99% effective but only 90% for cohort 2. This is misleading as the impact on the smaller cohort is greater by loosing the one student.
im trying to work out how I might insert some sort of weighting so loosing students in the smaller cohorts is lessened.
i hope that makes sense?
for example:
cohort 1 has 100 students and at a point, 1 student fails which makes the survival 99% at that point
cohort 2 has 10 students and at the same point, 1 student fails which makes the survival 90%
so this is being read that the course for cohort 1 was 99% effective but only 90% for cohort 2. This is misleading as the impact on the smaller cohort is greater by loosing the one student.
im trying to work out how I might insert some sort of weighting so loosing students in the smaller cohorts is lessened.
i hope that makes sense?
