PG851: GLM--ANOVA

This blog is for the use of students enrolled in either section of PG851. It is intended to be a forum for asking, discussing, clarifying, and helping.

Monday, February 21, 2011

Alpha correction question

When utilizing a Bonferroni correction factor, does it matter what type of t-test (paired sample or independent) you run? More specifically, if you run one paired sample and two independent t-tests, would you need to correct Bonferroni based on 3 t-tests?

Thanks!
Jenn

posted by Jennifer @ 10:17 AM  

2 Comments:

  • At 8:59 PM, Anonymous Nikki said…

    I am not sure that I completely understand your question so please let us know if you need more information.

    If you have a mixed design or repeated measures design, you need to include dependency in the model so you would use a paired samples t-test to determine which levels are significantly different. The independent t-test is used for randomized groups, which should be independent of each other.

    A Bonferroni correction is done by dividing the probability value (usually .05) by the number of tests being performed.

    Does that help?

     

  • At 6:50 AM, Anonymous Nikki said…

    To answer your question more specifically, yes, you would need to divide by 3 if you ran 3 tests for post-hoc comparisons.

     

Post a Comment

<< Home