Analyses of Binary outcomes Flashcards
1
Q
Notation: Sample proportion
A
p=r/n (T or C)
2
Q
Notation: Population odds
A
pi/1-pi (T or C)
3
Q
Notation: Sample odds
A
q=r/(n-r) (T or C)
4
Q
Notation: Risk Difference
A
RD; Pi_T-Pi_C
5
Q
Notation: Risk Ratio
A
RR= Pi_T/Pi_C
6
Q
Notation: Odds ratio
A
OR= [Pi_T(1-Pi_C)]/[Pi_C8(1-Pi_T)]
7
Q
The number need to treat formula
A
NNT= 1/RD
8
Q
assumptions for using Z test
A
subjects independent
np, and n(1-p) greater than 5 for T and C
9
Q
assumptions for Chi-Squared test
A
80% of expected frequencies are above or equal to 5
10
Q
advantages of using Risk Difference over Z test?
A
Can easily calculate Confidence intervals
The EAsily identify the direction of effect
Can calculate easily 1 sided test if appropriate
