Scientific Opinion of the PPR Panel on the follow‐up of the findings of the External Scientific Report ‘Literature review of epidemiological studies linking exposure to pesticides and health effects’

. 2017 Oct 31;15(10):e05007. doi: 10.2903/j.efsa.2017.5007

Results of power analysis for a one‐sided, two‐sample proportions test (α = 0.05)^a
N_control	N_exposed	Proportion control^b	Proportion exposed	Odds Ratio	Power
1,018	238	0.1194	0.1399	1.2	0.2279
1,018	238	0.1194	0.1689	1.5	0.647
1,018	238	0.1194	0.2133	2.0	0.9693
1,018	238	0.1194	0.2891	3.0	1
1,018	238	0.2387	0.2734	1.2	0.3047
1,018	238	0.2387	0.3199	1.5	0.8149
1,018	238	0.2387	0.3854	2.0	0.9971
1,018	238	0.2387	0.4847	3.0	1
1,018	238	0.4774	0.523	1.2	0.3522
1,018	238	0.4774	0.5781	1.5	0.8779
1,018	238	0.4774	0.6463	2.0	0.9992
1,018	238	0.4774	0.7327	3.0	1

Stata code used to generate the above results: power two‐proportions ('= 0.5 * 243/1018' '= 243/1018' '= 2 * 243/1018'), test(chi2 ) OR (1.2 1.5 2.0 3.0) n1(1,018) n2(238) one‐side table(N1:''N control'' N2:''N exposed'' p1:''proportion control'' p2''proportion exposed'' OR:''odds ratio'' power:''power'').

One‐sided test α = 0.05 Ho: p2 = p1 vs Ha: p2 > p1; N_controls = 1,018, N_exposed = 238, Number of iterations = 1,000 (data sets).

Representing 1/2x‐, 1x‐ and 2x‐ the observed background rate of lung cancer of 243/1018 in Waddell et al. (2001). Highlighted, bolded region in table above represents power associated with this 1x observed background rate of NHL in cited study.