Home > Statistics Every Writer Should Know > The Stats Board > Discusssion
Is there any problem arise when we test the same hypothesis more than once?
Message posted by Kathy on August 8, 2000 at 12:00 AM (ET)
I want to know if i want to test the hypothesis H0: U(control group) = U(experimental group). Moreover, I have 6 questions in the survey also want to test this hypothesis, is any problem arise regards this? If yes, what is the name of this problem and how to solve it? Thanks for your answering.
READERS RESPOND:
(In chronological order. Most recent at the bottom.)
Re: Is there any problem arise when we test the same hypothesis more than once?
Message posted by Chad Allen on August 9, 2000 at 12:00 AM (ET)
If you have only one independent variable (exp. vs. control group) and 6 dependent variables, I would not recommend running mulitple tests (t for example). This proceedure could result in using an alpha level over the long run which is actually larger than what you intended to set it at (probably .05). This inflation in alpha is called "familywise error rate." Instead, run what is called a MANOVA. This test allows you to look at the effects of one or more IVs on several DVs without inflating familywise error rate. This can be a somewhat complicated statistic to interpret. If you need any assistance, feel free to contact me.Good luck
Chad Allen
lewingca@tribe.nlu.edu
Your $5 contribution helps cover part the $500 annual cost of keeping this site online.
