Skip to main content

Table 1 Minimum detectable OR from J controls/case, for example minimum detectable OR for 1 control/case, based on a derivation valid for small associations (see Supplement). Note that the percent reduction in the OR only depends on the number of controls per case, subsuming power, α, number of cases, or control marker frequency

From: Increase in power by obtaining 10 or more controls per case when type-1 error is small in large-scale association studies

Controls/case (J)

Minimum detectable odds-ratio (OR) for 1 control/case

% Reduction towards null vs J = 1 ORa

% Reduction towards null vs J = 4 ORb

1.05

1.1

1.2

1.3

1.5

1.75

2

1

1.05

1.1

1.2

1.3

1.5

1.75

2

0.0%

-26.5%

2

1.043

1.087

1.17

1.26

1.43

1.65

1.87

13.4%

-9.5%

3

1.041

1.082

1.16

1.25

1.41

1.61

1.82

18.4%

-3.3%

4

1.04

1.079

1.16

1.24

1.40

1.59

1.79

20.9%

0.0%

5

1.039

1.077

1.16

1.23

1.39

1.58

1.78

22.5%

2.0%

10

1.037

1.074

1.15

1.22

1.37

1.56

1.74

25.8%

6.2%

20

1.036

1.072

1.15

1.22

1.36

1.54

1.73

27.5%

8.3%

50

1.036

1.071

1.14

1.21

1.36

1.54

1.71

28.6%

9.7%

100

1.036

1.071

1.14

1.21

1.36

1.53

1.71

28.9%

10.1%

1000

1.035

1.071

1.14

1.21

1.35

1.53

1.71

29.3%

10.5%

  1. a Fractional reduction towards the null for minimum detectable odds-ratio for J controls/case (ORJ) vs 1 control/case (OR1) equals (OR1 – ORJ)/(OR1 – 1)
  2. b Fractional reduction towards the null for minimum detectable odds-ratio for J controls/case (ORJ) vs 4 controls/case (OR4) equals (OR4 – ORJ)/(OR4 – 1)