Type I & Type II Errors
What are Type I & Type II errors?
- There are four possible outcomes of a hypothesis test
- Two good outcomes
- H0 was false and H0 got rejected
- H0 was true and H0 was not rejected
- Two bad outcomes (errors)
- H0 was true but H0 was rejected (a Type I error)
- H0 was false yet H0 was not rejected (a Type II error)
- Two good outcomes
- Type I errors occur when a hypothesis test gives sufficient evidence to reject H0 despite it being true
- This is sometimes called a “false positive”
- In a court case this would be when the defendant is found guilty despite being innocent
- Type II errors are when a hypothesis test gives insufficient evidence to reject H0 despite it being false
- This is sometimes called a “false negative”
- In a court case this would be when the defendant is found innocent despite being guilty
How do I find the probability of a Type I or Type II error?
- The probability of a Type I error is the probability of being in the critical region (rejecting H0) given H0 was true
- P(Type I error) = P(in the critical region | H0 is true)
- The critical region itself was set up based on a significance level, α%, which assumed H0 was true
- So for continuous distributions (normal, )
- P(Type I error) = α%
- It's exactly equal to the significance level
- You can often write this down with no calculation
- For discrete distributions (binomial, Poisson, Geometric)
- It's as close as you can get to α% whilst still being critical
- So P(Type I error) is the actual significance level (≤ α%)
- So for continuous distributions (normal, )
- The probability of a Type II error is the probability of not being in the critical region (not rejecting H0) given H0 was false
- You need to be given the actual population parameter to find this
- For example, H0 assumed but actually
- This is more helpful than just saying (though, in practice, harder to know)
- P(Type II error) = P(not in the critical region | actual population parameter)
- You need to be given the actual population parameter to find this
- Either error is not desirable in a hypothesis test
- Ideally you want to know these probabilities before doing a test
Can I reduce the probabilities of making a Type I or Type II error?
- You can reduce the probability of a Type I error by reducing the significance level
- However this will increase the probability of a Type II error
- You can reduce the probability of a Type II error by increasing the significance level
- However this will increase the probability of a Type I error
- The only way to reduce both probabilities is by increasing the size of the sample
Worked example
Lucy can hit the target 70% of the time when she throws an axe with her right hand. She claims that the proportion, p, of her throws that hit the target is higher than 70% when she uses her left hand. Lucy uses the hypotheses and to test her claim. Lucy makes 100 throws and will reject the null hypothesis if the axe hits the target more than 77 times.