We analyzed the impact of gene expression of 7 genes in patient's outcome. Two genes showed p < 0.05. We conducted a Cox regression model in our analysis and only one gene presented p
P value is not adjustable.Unless you will work on the data all over
Dear Pedro,
Yes, you certainly can. Either multiply the probability value by the number of pairs or divide the threshold by the same number. With seven groups there are 21 possible pairs.
Thnaks Abiodun and Peter. But is it necessary? I read works of survival analysis and high impact journals accept univariate and multivariate analyzes in their studies. The Bonferroni adjustment is not common in this type of analysis, or am I mistaken?
Dear Pedro,
I believe it's necessary to reduce false positives i.e. type 1 errors and it's commonly done on univariate tests.
Hope this article helps
"Correction of the p-Value after Multiple Tests in a Cox Proportional Hazard Model"
https://link.springer.com/article/10.1023/A:1020514804325
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Meta-Analysis