Table 2 Performance matrix for different methods of estimating variance.

From: Differential impacts of regenerative agriculture practices on soil organic carbon: a meta-analysis of studies from India

Method

Estimate

SE

CI.lb

CI.ub

\(\tau ^2\)

\(I^2\) (%)

REML

3.11

0.22

2.68

3.54

4.58

75.6

DL

3.07

0.21

2.65

3.48

4.17

73.9

ML

3.10

0.22

2.67

3.53

4.49

75.3

SJ

4.26

0.55

3.20

5.33

38.58

96.3

HS

3.06

0.21

2.64

3.48

4.11

73.6

HE

4.62

0.72

3.20

6.03

70.23

97.9

HSk

3.06

0.21

2.65

3.48

4.15

73.8

EB

3.85

0.40

3.08

4.63

18.98

92.8

PM

3.85

0.40

3.08

4.63

18.98

92.8

PMM

3.86

0.40

3.08

4.64

19.18

92.9

  1. REML restricted maximum likelihood, ML maximum likelihood, DL DerSimonian–Laird, SJ Sidik–Jonkman, HS Hunter–Schmidt, HE Hedges, HSk Hunter–Schmidt (corrected), EB empirical Bayes, PM Paule–Mandel, PMM median-unbiased Paule-Mandel.
Back to article page