Better Business Decisions from Data

CHAPTER 12

Types of Error

How Wrong Can You Be?

Whenever a significance level is quoted, there is a chance that the stated result is incorrect. If the null hypothesis is rejected when, in fact, it is correct, the error is referred to as a Type I error. So, if our null hypothesis is that there is no significant difference between the mean marks from the boys’ results and the girls’ results in the same examination, we may decide that there is a difference, at the 5% level, say. If in fact there is no difference, and our result is simply due to the random effect embodied in our 1-in-20 chance of being wrong, then a Type I error has occurred.

Alternatively, our result may show no significant difference, and we would accept the null hypothesis. If ...