Overpowered and underpowered chi-squared tests – or are they?

This is a very quick thought for teaching, in passing. People often talk about chi-squared tests being overpowered when n is large. It occurs to me that a good way to broach this concept in an intuitive way is to point out that they are no different to t-tests and the like, but do not provide a meaningful point estimate. When you see the mean difference in blood pressure from drug X is 0.3mmHg, with p<0.001, you know it is clinically meaningless. When you see X2=3.89, nobody knows what to think. So perhaps the best thing to do is to mention this alongside non-parametric rank-based procedures, when you explain that they don’t give you an estimate or confidence interval.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s