Skip to main content
. Author manuscript; available in PMC: 2022 Feb 1.
Published in final edited form as: Psychol Bull. 2020 Oct 29;147(2):134–168. doi: 10.1037/bul0000307

Table 2.

Multi-level Model Results for Overall and Moderator Analyses

Moderator F(df1, df2) Between-study variance
(Level-1)
Within-study variance
(Level-2)
QE(df) I2
Overall -- .01 .02 7797.56(749)** 90.39%
Gender 1.19(1, 88) .02 .01 494.51(88)** 82.00%
Race 0.28(1, 174) .003 .01 609.50(174)** 71.29%
Continent 1.56(5, 735) .01 .02 7354.65(735)** 89.94%
Grade level 6.64(5, 741)** .01 .02 6582.82(741)** 88.67%
Teachers 0.59(1, 745) .01 .02 7755.88(745)** 90.38%
Low math ability 6.29(1, 745)* .01 .02 7732.83(745)** 90.35%
Child math anxiety scale 1.26(5, 149) .01 .02 735.33(149)** 79.06%
Adolescent/Adult math anxiety scale 6.87(9, 502)** .004 .02 5233.87(502)** 90.24%
Anxiety scale by topic 12.14(1, 745)** .01 .02 7199.63(745)** 89.64%
Components of math anxiety 3.02(5, 236)* .01 .01 1273.77(236)** 81.08%
Math assessment 2.34(4, 742)* .01 .02 7398.40(742)** 89.92%
Math content area 1.40(7, 327) .01 .02 1176.00(327)** 71.60%

Note.

*

p < .05

**

p < .001

F = omnibus test; df = degrees of freedom; QE = Residual Heterogeneity; I2 = Heterogeneity Percentage.