Type I and Type II errors Flashcards
how to control type I error
we control type one error when we choose the signifcance level
What is a type one error
Rejecting the null when it is true, this occurs when p<signficance level but the null is remains true. This can occur bc in a 95% confidance interval we are 95% sure that our value will fall within that interval however there is a 5% chance that the value falls outside of this value
What is type II error
Fail to reject the null when it is false , This may occur as there are times when the null is false however we do not have the evidance to spport it
what causes type II error
our sample size and study design, a larger sample size should limit type II error
what is POWER
1-P(typeII error) High powers make it easier to reject a null, low powers make it more challnaging
mutplie comparisons
Suppose the null is true. We still have a 5% chance of rejecting a true null, and therefore a 95% chance of correctly not rejecting it.
If we have two nulls, both of which are true, then the probability that we correctly don’t reject either is 0.95*0.95 = 0.9025. This means we have a 9.75% chance of rejecting at least one of them, even though they’re both true.
The multiple comparisons problem states that, when we check a lot of p-values, the Type 1 Error increases.