Table a1 Statistical power of the optimal test, the LRS test and the Stone's Poisson maximum test at the 5% significance level according to four values of the SIR

From: Incidence of childhood leukaemia in the vicinity of nuclear sites in France, 1990–1998

  

Optimal test

  

H1

ρ 1

α 0

Power

LRS test

Poisson max. test

Linear decrease

 

1.1

4.79

36.98

32.94

19.60

 

1.2

4.95

84.58

77.78

53.48

 

1.5

4.71

100.00

100.00

99.89

 

2

4.99

100.00

100.00

100.00

  

α 0

 

4.91

3.93

Rapid decrease after 5 km

 

1.1

4.03

19.00

18.18

19.07

 

1.2

5.34

47.22

40.34

47.27

 

1.5

4.16

98.40

95.57

98.40

 

2

5.49

100.00

100.00

100.00

Rapid decrease after 10 km

 

1.1

4.52

43.82

33.43

21.13

 

1.2

5.05

91.59

78.42

64.60

 

1.5

4.57

100.00

99.99

99.98

 

2

4.34

100.00

100.00

100.00

Rapid decrease after 15 km

 

1.1

4.88

71.00

52.59

22.40

 

1.2

4.84

99.42

95.41

72.33

 

1.5

4.52

100.00

100.00

100.00

 

2

4.49

100.00

100.00

100.00

  

α 0

 

4.83

4.10

  1. LRS test=linear risk score test; SIR=standardised incidence ratio. ρ1 in the 0–5 km zone under two types of alternative hypothesis H1: a linear decrease in SIR and three forms of rapid decrease in SIR (after 5, 10 or 15 km); α0 is the real value of the probability of type-I error (given by simulation). Note: These calculations are for the 27 site global analysis.