Transcript Type I and Type II errors PPT

Type I error (alpha error)




Occurs when an experimenter thinks
she/he has a significant result, but it is
really due to chance
Analogous to a “false positive” on a drug
test.
Risk of a Type I error is the same as the
significance level, e.g., p < .05
Solutions: avoid internal validity errors
(such as confounding variables), use a
more stringent significance level, use
replication
Jump to first page
Type II error (beta error)




Occurs when a researcher fails to find a
significant result when, in fact, there was
something significant going on.
Analogous to a “false positive” on a drug
test.
Must be calculated with a test of
statistical “power,” e.g., given the sample
size, how big would an effect have to be
in order to detect it?
Solutions: increase sample size, use
more sensitive precise measures, use
replication
Jump to first page
Type I and Type II errors
Based on sample,
H0 is supported
In the larger population, In the larger population
H0 is correct
H1 is correct
Correct Decision
Incorrect Decision:
Type II Error
Base on sample, H1 is
accepted
Incorrect Decision:
Type I Error
Correct Decision
Jump to first page
Implications

To some extant, Type I and Type II errors
trade off with one another


Decreasing the chance of a Type I error
may increase the chance of a Type II
error.
A Type I error is the more egregious of
the two
Type I entails shouting “Eureka” when
you haven’t really found it.
 Scientific skepticism makes Type II errors
more palatable

Jump to first page