Skip to main content

Table 4 Impact of sample size on type I error rates for the case-only design

From: The case-only design is a powerful approach to detect interactions but should be used with caution

 

Disease prevalence

Number of cases

\({\varvec{\upalpha}}\)

1.0%

1.5%

2.0%

2.5%

3.0%

3.5%

4.0%

4.5%

5.0%

10.0%

15.0%

20.0%

2,500

0.001

0.000983

0.000958

0.000942

0.001019

0.000959

0.000911

0.000962

0.000956

0.001024

0.001057

0.001097

0.001216

0.05

0.049557

0.049826

0.049762

0.050058

0.050021

0.049818

0.050325

0.050030

0.050102

0.051262

0.053111

0.054364

5,000

0.001

0.000996

0.001050

0.000976

0.000976

0.000972

0.001000

0.000949

0.001010

0.001060

0.001153

0.001275

0.001413

0.05

0.049919

0.050110

0.049596

0.050317

0.050346

0.050453

0.050565

0.050731

0.050874

0.053727

0.056344

0.059515

10,000

0.001

0.000994

0.001004

0.001058

0.001037

0.001057

0.001106

0.001115

0.001068

0.001097

0.001357

0.001731

0.001966

0.05

0.050155

0.050387

0.050449

0.050325

0.051138

0.050999

0.051155

0.051520

0.052183

0.057479

0.063428

0.070083

20,000

0.001

0.000980

0.000998

0.001012

0.001052

0.000992

0.001114

0.001162

0.001186

0.001267

0.001838

0.002487

0.003475

0.05

0.049976

0.050759

0.050546

0.051070

0.051843

0.052863

0.053234

0.053472

0.054632

0.065503

0.078536

0.090791

50,000

0.001

0.001034

0.000988

0.001076

0.001160

0.001208

0.001294

0.001373

0.001544

0.001647

0.003449

0.005998

0.008934

0.05

0.050299

0.051092

0.051977

0.053040

0.054681

0.055883

0.058031

0.059773

0.061797

0.089890

0.122479

0.154596

  1. Data were generated under the null with no interaction but two main effects [\({\beta }_{G}=\ln(1.2)\) and \({\beta }_{E}=\ln(2)\)]. The MAF for the genetic variant is 0.2 and the frequency of the environmental exposure is 0.1. Type I error is calculated as the proportion of instances that the null hypothesis is rejected among all 1,000,000 replicates. Type I errors exceeding 5% of the alpha level are highlighted in bold in the table