For example, patients had their blood pressure measured at two visits 3 months apart 6 months after infection and found no difference between the measurements using a paired test. The researchers then decided to compare whether the blood pressure of patients who had never had the infection differed from the blood pressure of patients who had had the infection at one of the visits.
Yes! In this case a mixed effects model, with patient ID included as a random effect (to account for the repeated measures), and visit and infection status included as fixed effects, would be able to determine this. You would effectively then have p for visit and p for infection status. An interaction factor of visit x infection would answer the final part of your question.
Using this technique means only one statistical model is performed. Post-hoc comparisons can be made using Tukey test. As long as your n number is sufficiently high, the data does not need to be normally distributed so long as the residuals do follow a normal distribution. This can be checked within your statistical software of choice. I hope this helps.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science
- Materials Chemistry
- Polymer Chemistry