4

The research question: Do vision test results differ among 3 groups (of eyes) separated on the basis of retinal health: Normal: Mild Disease; Moderate disease. Problem: Both eyes of each person are tested and we would like to use data from both eyes. For some people, both eyes are in the same group (ie. have the same retinal health, so there are 2 measures from a single person in a group), and for other people, the two eyes have different retinal health and fall into separate groups. Thus, there are independence problems within and between groups.

Question: Can we include both eyes in our analyses to determine whether the 3 groups of eyes differ on the vision measures, and if so, how?

MES
  • 41
  • 2
    You definitely should include both eyes, and hierarchical regression modeling would be a natural technique to accomplish this. Is it at all possible not to categorize the clinical assessment of retinal health? Can you use retinal imaging to quantify this assessment—maybe even along several different continuous 'dimensions' of retinal health? – David C. Norris Sep 13 '17 at 21:18
  • 1
    Unfortunately, no. We do not have a continuous measure to use. – MES Sep 14 '17 at 17:39

0 Answers0