1

I've got a dataset with two outcome measures

  • Count of submissions made by an individual. At least one submission was made by all individuals in the dataset.
  • Count of those submissions that were accepted. Some individuals received zero acceptances.

My hypothesis is that certain groups will have a better return on investment and their rate of acceptance will be higher. (i.e., certain groups can make fewer submissions and have those submissions accepted)

What test would be appropriate?

Thanks.

Cadence
  • 11
  • 1

1 Answers1

0

From what I understand, you're trying to predict which "groups" (sub-samples of your training set) have higher acceptance/submission ratios than other groups. Are these groups clearly defined in your training data, or are you trying to create that distinction? If the distinction is already there, have you calculated the mean acceptance/submission ratio for each one? If not, I'd try maybe a bayes classifier, depending on how much extra data you have.

Daniel C
  • 51
  • The groups would already be defined by the time I measure this, but I haven't started the analysis yet as I'm still in the planning stages. I'm defining groups through a latent class analysis. I can classify group membership as most likely membership (0,1) or by the probability the individual would be in said group. – Cadence Jan 23 '20 at 19:41