138 PART 3 Getting Down and Dirty with Data

The same kind of correspondence is true for other confidence levels and

significance levels. For example, a 90 percent confidence level corresponds to

the α = 0.10 significance level, and a 99 percent confidence level corresponds

to the α = 0.01 significance level, and so on.

So you have two different but related ways to estimate if an effect you see in your

sample is a true effect. You can use significance tests, or else you can use CIs.

Which one is better? Even though the two methods are consistent with one

another, in biostatistics, we are encouraged for ethical reasons to report the CIs

rather than the result of significant tests.»

» ^{The CI around the mean effect clearly shows you the observed effect size, as}

well as the size of the actual interval (indicating your level of uncertainty about

the effect size estimate). It tells you not only whether the effect is statistically

significant, but also can give you an intuitive sense of whether the effect is

clinically important, also known as clinically significant.»

» ^{In contrast, the p value is the result of the complex interplay between the}

observed effect size, the sample size, and the size of random fluctuations.

These are all boiled down into a single p value that doesn’t tell you whether

the effect was large or small, or whether it’s clinically significant or negligible.