Table 7. Results of parameter estimation of the distribution of effect sizes and measures of data peeking as a function of population effect size (δ, ρ_F), population heterogeneity (τ), and data peeking, for the simulated data.

Results are based on all p-values 0–1, p-values ≤.05, and ≤.00125.

			τ = 0				τ = .15
	p-values		δ = 0ρF = 0	δ = .2ρF = .099	δ = .5ρF = .247	δ = .8ρF = .390	δ = 0ρF = 0	δ = .2ρF = .099	δ = .5ρF = .247	δ = .8ρF = .390
Without data peeking	0–1	${\hat{\rho }}_F$	0	0.103	0.258	0.413	0	0.103	0.258	0.413
		${\hat{\tau }}_{{\rho }_F}$	0	0	0	0	0.077	0.077	0.077	0.077
	0–.05	${\hat{\rho }}_F$	0	0.103	0.258	0.413	0	0.103	0.258	0.413
		${\hat{\tau }}_{{\rho }_F}$	0	0	0	0.001	0.077	0.077	0.077	0.077
		Misfit χ2	0	0	0	0	0	0	0	0
	0–.00125	${\hat{\rho }}_F$	0	0.103	0.258	0.413	0.1	0.107	0.259	0.413
		${\hat{\tau }}_{{\rho }_F}$	0	0	0	0.001	0.025	0.076	0.077	0.077
		Misfit χ2	0	0	0	0	0	0	0	0
		D	1	1	1	1	1.205	1.006	1.003	1.001
With data peeking	0–.05	${\hat{\rho }}_F$	0	0	0.117	0.345	0	0	0.075	0.360
		${\hat{\tau }}_{{\rho }_F}$	0	0	0	0.038	0	0.055	0.137	0.091
		Misfit χ2	126,267.4	50,298.4	696.6	101.6	14,867.6	1,209.5	576.3	340.6
		N	759,812	811,296	936,517	994,974	434,660	525,023	707,650	889,681
	0–.00125	${\hat{\rho }}_F$	0	0.075	0.218	0.366	0.066	0.161	0.283	0.402
		${\hat{\tau }}_{{\rho }_F}$	0	0	0	0	0.036	0	0	0.012
		Misfit χ2	6.9	3.2	7.1	11.8	2	1.9	2.6	2.1
		N	9,729	21,576	95,615	350,482	14,791	34,530	124,991	366,875
		D	1.977	1.976	1.835	1.166	1.628	1.620	1.472	1.164

Notes.

${\hat{\rho }}_F$, estimated population effect; ${\hat{\tau }}_{{\rho }_F}$, estimated population heterogeneity; misfit 0–.05; misfit of estimates based on p-values 0–.05, misfit 0–.00125, misfit of estimates based on p-values 0–.00125 (bold indicates p < .05); N, number of results included in estimation; D, comparison of observed- and expected p-value frequencies.