Return to Website

Number Watch Web Forum

This forum is about wrong numbers in science, politics and the media. It respects good science and good English.

Number Watch Web Forum
Start a New Topic 
Is P

P<5% means there is a 1 in 20 chance the findings of a study are false; but,does that also indicate that there is a probability the findings are off by 5%?

That would be a 5% 'margin of error(MoE)'.

Say a study showed a 1 per 100 increased death rate;but, a 5% MoE would mean that anything less than a 5 per 100 rate would not be valid.

Re: Is P

My command of statistics is sadly less than it ought to be, but I think the answer is no. For instance, if the result is binary, i.e. 1 or 0, yes or no, the result cannot be 5% off.
John has covered it in "Sorry, wrong number", but I don't have my copy to hand.

Re: Is P

I have the book and 'margin of error' is not mentioned.

Re: Is P

You should be careful to distinguish between P and relative risk. The nearest statistical measure to Margin of Error would be Confidence Interval. See our FAQs.