Andrea Maliachi, Medical specialist at the National Cancer Institute, shared on X:
“Not shiny new, still a gem.
The p value, its limitations and misleading use in medical research.
‘p is the outcome of a test that tests the probability of rejecting H0. The smaller P, the stronger the evidence for rejecting H.’
So… p is a measure of probability, not a binary outcome.
The P Value – Donts and Dos:
Don’t conclude ‘p<0.05 statistically significant’
Report p value as continuous number.
Remember this value has nothing to do with the ‘significance’ of presented results.
Don’t base results/conclusions solely on p value.
Take into account: Sample size, Effect size, Power, uncertainty, design and methods, CIs, pre-planned vs posthoc, other data and previous research as well.”
Additional information.
Source: Andrea Maliachi/X