0

I'm not a statistician by training so I'm looking for some suggested tests to determine differences in population distribution. I have the data presented below. A very long list of individuals with an assigned IMD score (1 to 5) on the X and an assigned medical clinical ID on the Y.

The 'value' (apologies, I haven't changed the key title) are the number of patients with that IMD score and that particular clinical ID.

I would like a pairwise comparison between IMD scores to see whether the patient visits across clinic IDs are significantly different e.g., is IMD 3 different to IMD 2?

enter image description here

Anthony Nash
  • 595
  • This is a weird graph, also your x axis is not ordered. So you have a number of clinics, each clinic has a number of patients, and each clinic received a number of scores from 1 to 5. Now you want to check whether there is a different number of patients in the IMD variable for each clinic? – user2974951 Jul 23 '19 at 07:36
  • This was just a screen dump from R-studio. The ordering doesn't matter and would be rectified with more time. The reason why this looks 'weird' is that it is actually a massive heatmap with a lot of missing Clinic IDs. Your summary is correct. – Anthony Nash Jul 23 '19 at 08:51
  • Looks like a job for linear regression with ordinal predictors. – user2974951 Jul 23 '19 at 10:58

0 Answers0