Table 5.

Separate Weighted OLS Regressions with each Moderator predicting Gender Differences in Major Depression (unshaded) and Gender Differences in Depression Symptoms (shaded)

Continuous Moderators	β	Exp(β)	k	Qmodel	Qresidual	R2
Year of data collection	0.14*	1.01	300	6.25*	316.48	0.02
Year of data collection	0.00		413	0.66	405.64	0.00
Age@			184	82.47**	221.35*	0.27
Linear	−0.45**	0.99
Quadratic	0.35**	1.00
Age@			324	40.06**	319.83	0.11
Linear	−0.37**
Quadratic	0.12*
Nation-level economic indicators
Income inequality	0.09	1.01	234	2.19	252.70	0.01
Nation-level economic indicators
Income inequality	−0.11*		316	3.73*	308.77	0.01
Nation-level gender equity indicators
Contraceptive prevalence	0.22**	1.01	294	16.14**	308.79	0.05
Executive positions	0.10	1.00	256	2.65	271.35	0.01
Literacy ratio	0.18**	1.84	297	11.08**	313.62	0.03
Nation-level gender equity indicators
Contraceptive prevalence	0.03		369	0.36	362.30	0.00
Executive positions	−0.06		376	1.24	369.63	0.00
Literacy ratio	−0.03		404	0.28	397.68	0.00

Note. Each moderator was run in a separate regression, except for the age variables which were run in a multiple regression. β = Beta. Negative values indicate that there are larger reported gender differences at lower levels of the moderator and smaller reported gender differences at higher levels of the moderator. k = number of effect sizes for each age. 95% CI = 95% confidence interval for β. Exp(β) = Exponent of Beta. Significance for Exp(β) is not indicated because it is the same as β. Q_model = Significant values indicate that there is significant variability accounted for by the moderator. Q_residual = Significant values indicate that studies are still heterogeneous after accounting for the moderator variable. R² = Amount of variance accounted for by the moderator(s).

⁺p <.10.

^*p <.05.

^**p <.001.

^@= For age as a moderator, lifetime depression was excluded and samples with large age ranges (e.g., 18–64) were excluded to more sensitively test developmental effects.