|
|
Actually true |
Actually not
true |
Test total |
|
Test positive |
A 30 |
B 15 |
45 |
|
Test negative |
C 10 |
D 45 |
55 |
|
Actual total |
40 |
60 |
100 |
· true positive
· test positive and actually true
· A
· false positive
· test positive but actually not true
· B
· false negative
· test negative but actually true
· C
· true negative
· test negative and actually not true
· D
· sensitivity
· test positive and actually true
· A/A+C = 30/40
· specificity
· test negative and actually not true
· D/B+D = 45/60
· accuracy
· test positive and actually true and negative and actually not true
· A+D/A+B+C+D = 75/100
· positive predictive value
· likelihood that patient with positive test has disease
· A/A+B = 15/45
· negative predictive value
· likelihood that patient with negative test does not have disease
· D/C+D = 45/55
|
|
study detected difference |
study detected no difference |
|
null hypothesis true |
type 1 error |
confidence |
|
null hypathesis false |
power |
type II error |
· power
· correctly rejects null hypothesis
· recognises that treatment makes difference
· power to change treatment
· confidence
· correctly accepts null hypothesis
· recognises that treatment makes no difference
· confidence to continue current treatment
· type I error
· incorrectly rejects a true null hypothesis
· does not recognise that treatment makes no difference
· type I person likely to think treatment works when it really doesn’t
· type II error
· incorrectly accepts a false null hypothesis
· does not recognise that treatment makes difference
· type II person likely to not recognise treatment works when it really does