STALT stands for Smaller Than A Lowest Threshold

This one is so simple. It shouldn't really be a test! But it's my life, so there you go. Either way, we've found more problems with it in a month than we did with the other tests in a year.

A lot of papers report the outcome of statistical tests as 'p<.05', or 'p<.001' -- so the question you should always ask is 'oh, it's smaller is it, so HOW MUCH SMALLER?'

Would it change your mind if the p value 'smaller than .05' was actually 0.0000000005? You'd ask questions. So, you simply back-calculate the relevant statistical test where possible.

Some additional fun points:

• STALT is just back-calculation in context. It's very very easy to understand.

• STALT reveals something I've always said, which is p-values are annoying.

Go back Home