1

I want to study the effect of 2 drugs (used singly and in combination) compared to a placebo on a disease. What is the best statistical method? Kindly provide full details.

DrWho
  • 799
  • 4
  • 12
  • 23

1 Answers1

1

What you are looking to compare is to see if the mean outcome is improved or not by using the mean. You dont want to do this with ordinary t-tests as the tests are not independent and the probabilities would no longer be accurate (you would have a larger type 1 error).

Therefore you want to test the means jointly using for example an ANOVA test that does not suffer from the inflated type 1 error.

robinsa
  • 383
  • 2
  • 10
  • Thank you. Can you provide any additional special steps to be taken using t ANOVA? – DrWho Sep 22 '16 at 12:00
  • The procedure is a little tedious but here is a good explanation of how to do it step by step in great detail: http://sphweb.bumc.bu.edu/otlt/MPH-Modules/BS/BS704_HypothesisTesting-ANOVA/BS704_HypothesisTesting-Anova_print.html Almost all statistical software will do this automatically for you in a built in function. I hope this helps but if not I am sure someone else can provide a more detailed answer. – robinsa Sep 23 '16 at 09:01
  • Thank you. I have statistical software. The reference is quite useful. – DrWho Sep 24 '16 at 06:20