Bonferroni correction (significance level and CI)

#1
Hi there,

I need another advice.

I'm dealing with the problem of multiple testing and I was planning to use Bonferroni correction. Because I made 15 separate tests, I adjusted the alpha value to 0,0003 (0,05/15).

In the results section, I am reporting p-values and confidence intervals. Since I adjusted the significance level, do I have to adjust confidence interval also? Before Bonferroni correction, I was planning to use 95 % confidence intervals (I also use them on the graphical presentation of my data). Do I have to change that to 99,7 % CI?

Thanks for your help in advance :)
 

hlsmith

Omega Contributor
#2
Yes, this is the appropriate action to take. Also, if you do not adjust both values which are calculated using alpha level, then their values will conflict.