The Delta likelihood ratio does not incorporate study power. Author response
We thank the author for his interesting critique of our paper [1]. With regard to the false positive rate being dependent on alpha rather than the P-value, we point out that alpha is set to be equivalent to the P-value; thus, the two equations shown in the letter are identical. It is illogical to have a positive likelihood ratio less than one or a negative likelihood ratio greater than one. The only way to guarantee this does not occur is to set alpha equal to the P-value, given that that from Equation 3 in our paper, power will always be greater than alpha and hence greater than the P-value.
Source: Journal of Clinical Epidemiology - Category: Epidemiology Authors: Nicholas Adams Tags: Letter to the Editor Source Type: research
More News: Epidemiology | Study