Skip to main content
. 2020 Jul 7;27(6):1374–1382. doi: 10.3758/s13423-020-01772-1

Table 2.

Summary of tests for publication bias

Adjusted effect Test for publication bias
All cognitive measures: k = 38, Δuc = -0.13 (p = 0.011)
 1. Rank correlationa r = -.24 (p = .036)
 2. Peter et al.’s (2006) regression test Δ = 0.03 (p = .604) B = -13.43 (p = .009)
 3. PET/PEESEb Δ = 0.04 (p = .513) B = -3.29 (p = .002)
 4. Modified PET/ PEESEc Δ = 0.03 (p = .604) B = -3.36 (p = .009)
 5. Selection modelsd Δ = 0.03 (p = .419) χ2 = 29.53 (p < .001)
 6. Modified selection modelsc Δ = 0.01 (p = .702) χ2 = 26.48 (p < .001)
 7. puniform*e Δ = -0.02 (p = .678) Lpb = 15.49 (p < .001)
Anagram tests: k = 12, Δuc = -0.06 (p = 0.184)
 1. Rank correlationa r = -.27 (p = .250)
 2. Peter et al.’s (2006) regression test Δ = -0.01 (p = .900) B = -9.28 (p = .095)
 3. PET/PEESEb Δ = 0.00 (p = .983) B = -2.31 (p = .069)
 4. Modified PET/PEESEc Δ = -0.01 (p = .900) B = -2.32 (p = .095)
 5. Selection modelsd Δ = -0.01 (p = .785) χ2 = 2.76 (p = .331)
 6. Modified selection modelsc Δ = -0.02 (p = .678) χ2 = 2.49 (p = .114)
 7. puniform*e Δ = -0.04 (p = .422) Lpb = 0.13 (p = .936)
Reasoning tests: k = 17, Δuc = -0.34 (p = .016)
 1. Rank correlationa r = -.43 (p = .017)
 2. Peter et al.’s (2006) regression test Δ = 0.17 (p = .324) B = -23.44 (p = .021)
 3. PET/PEESEb Δ = 0.14 (p = .333) B = -5.06 (p = .009)
 4. Modified PET/PEESEc Δ = 0.17 (p = .324) B = -5.86 (p = .021)
 5. Selection modelsd Δ = 0.08 (p = .175) χ2 = 29.50 (p = .033)
 6. Modified selection modelsc Δ = 0.05 (p = .444) χ2 = 23.92 (p < .001)
 7. puniform*e Δ = 0.03 (p = .704) Lpb = 6.78 (p = .034)
Knowledge tests: k = 10, Δuc = -0.04 (p = .598)
 1. Rank correlationa r = .29 (p = .291)
 2. Peter et al.’s (2006) regression test Δ = -0.21 (p = .200) B = 19.80 (p = 252)
 3. PET/PEESEb Δ = -0.17 (p = .242) B = 3.72 (p = .308)
 4. Modified PET/PEESEc Δ = -0.21 (p = .200) B = 4.95 (p = .252)
 5. Selection modelsd Δ = -0.02 (p = .813) χ2 = 0.80 (p = .371)
 6. Modified selection modelsc Δ = -0.02 (p = .806) χ2 = 0.79 (p = .375)
 7. puniform*e Δ = -0.01 (p = .850) Lpb = 0.86 (p = .649)

Note. All analyses are based on the within-sample averaged sample sizes (see Supplementary Online Material, Section C)

Δ indicates pooled effect adjusted for publication bias, Δuc indicates unadjusted pooled effect, k number of independent effect sizes, B unstandardized regression weight, χ2 test statistic for publication bias with 1 degree of freedom, Lpb test statistic for publication bias

aBegg and Madzumdar (1994)

bStanley (2017)

cSubstitutes the sampling variances of d with a function of the sample size (Pustejovsky & Rodner, 2019)

dVevea and Woods (2005)

evan Aert and van Assen (2020)