I'm doing research on the influence of one kind of pill on the heart rate. I have two groups: control group and medication group. The control group has 10 subjects while the medication group has 20 different subjects.

Firstly, all 30 subjects' heart rate data were recorded every 15 seconds. Then, the subjects in control group were given a kind of placebo and the subjects in medication group were given pills (they didn't know whether they were given pills or placebo). After that, the heart rate were recorded again every 15 seconds.

In case the description is too complicated, I attached a plot to intuitively show the design, in which the real data were replaced with "xxx".

enter image description here

Now I want to test whether the heart rate of medication group is statistically different from that of control group. What kind of test should I use? Can I use hierarchical ANOVA?

  • $\begingroup$ How many measurements were there per patient? When was the pill administered? Is it's effect supposed to be instantaneous? Usually, pills need to be digested & the active agent absorbed into the bloodstream, etc., such that you wouldn't expect an effect w/i 15 seconds. Do you know anything about the lag that theory suggests you should see? $\endgroup$ – gung - Reinstate Monica Jun 18 '18 at 14:04
  • $\begingroup$ @gung Yes, the effect is instantaneous. Actually, the "pill" is not something virtual, it's a kind of visual intervention. $\endgroup$ – Yujian Jun 18 '18 at 18:27
  • $\begingroup$ How many measurements per patient before & after intervention? Do you expect a single increase (decrease) that otherwise stays constant, or for the effect to ramp up continuously, or taper off, etc? Can you post a small example dataset? $\endgroup$ – gung - Reinstate Monica Jun 18 '18 at 18:40
  • $\begingroup$ @gung Sorry for the long delay. There're 40 measurements for both before and after intervention. I expect the HR to decrease. By the way, I just picked some random data, how could I upload the excel file here? $\endgroup$ – Yujian Jun 19 '18 at 2:30
  • $\begingroup$ Your dataset looks to be rather large. You could try making a reproducible example, or you could upload a datafile to a site that lets you share data (dropbox, github, Google drive, etc) & post a link to it. $\endgroup$ – gung - Reinstate Monica Jun 20 '18 at 18:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.