. Author manuscript; available in PMC: 2022 Sep 11.

Published in final edited form as: Psychol Bull. 2021 Nov;147(11):1184–1214. doi: 10.1037/bul0000344

Table 3.

Participant Moderator Analyses

Moderator	j	k	β	95% CI	P	R ₍₂₎ ²	R ₍₃₎ ²	Q	τ₍₂₎²	τ₍₃₎²

Age (linear)	224	943	−0.007	[−0.01, −0.005]	<.0001	.000	.179	3368.9^*	0.08^*	0.09^*
Age (quadratic)	—	—	0.0002	[0.00001, 0.0003]	.041	.000	.200	—	0.08^*	0.09^*
Sex	220	901	−0.17	[−0.51, 0.17]	.324	.002	.000	3289.5^*	0.09^*	0.11^*
Education	136	563	0.05	[0.03, 0.07]	<.001	.000	.163	2049.1^*	0.08^*	0.10^*
Premorbid IQ	55	243	0.01	[−0.005, 0.02]	.333	.001	.024	841.3^*	0.08^*	0.07^*
Diagnostic status	236	995	−0.30	[−0.40, −0.21]	<.001	.000	.134	3524.9^*	0.09^*	0.09^*
Depression status	236	995	−0.06	[−0.18, 0.07]	.387	.000	.002	3524.9^*	0.09^*	0.11^*
Matched age	216	911	−0.13	[−0.27, 0.02]	.096	.000	.022	3226.7^*	0.09^*	0.11^*
Matched education	177	713	0.06	[−0.11, 0.22]	.504	.001	.001	2682.5^*	0.10^*	0.11^*
Medication status	166	661	−0.24	[−0.39, −0.09]	.002	.001	.091	2340.7^*	0.08^*	0.11^*
Symptom severity—MADRS	27	115	−0.01	[−0.02, 0.001]	.101	.000	.213	307.6^*	0.0003	0.05^*
Symptom severity—HDRS	83	327	−0.01	[−0.02, 0.0002]	.053	.001	.046	972.8^*	0.04^*	0.09^*
Symptom severity—BDI	102	438	−0.01	[−0.02, −0.004]	.009	.000	.191	1735.8^*	0.18^*	0.05^*
Symptom severity (HDRS scale)	179	752	−0.01	[−0.02, −0.01]	<.001	.000	.097	2683.6^*	0.10^*	0.07^*
Episode (first vs. recurrent)	75	298	−0.03	[−0.32, 0.26]	.844	.0003	.0002	1147.7^*	0.06^*	0.11^*
Time since diagnosis	67	296	0.0001	[−0.001, 0.001]	.851	.000	.002	1278.3^*	0.08^*	0.12^*
Age of depression onset	67	296	−0.01	[−0.02, −0.01]	<.001	.000	.228	1278.3^*	0.08^*	0.09^*
Early- vs. late-onset	29	116	−0.25	[−0.71, 0.21]	.301	.005	.057	357.8^*	0.01	0.14^*
Recruitment site	183	772	−0.36	[−0.46, −0.25]	<.001	.004	.269	3068.0^*	0.11^*	0.08^*

Mood-congruency effects (interactions with valence)
Age	221	936	—	—	—	.075	.255	3320.2^*	0.08^*	0.08^*
Age			−0.01	[−0.01, −0.003]	<.001
Valence			0.33	[0.05, 0.61]	.029
Age × Valence			−0.001	[−0.01, 0.01]	.722
Depression status	232	984	—	—	—	.076	.173	3470.7^*	0.08^*	0.09^*
Depression status			−0.08	[−0.21, 0.05]	.251
Valence			0.30	[−0.07, 0.66]	.153
Depression status × Valence			0.03	[−0.36, 0.41]	.901
Diagnostic status	232	984	—	—	—	.083	.256	3470.7^*	0.08^*	0.08^*
Diagnostic status			−0.29	[−0.39, −0.19]	<.001
Valence			0.18	[0.04, 0.31]	.020
Diagnostic status × Valence			0.19	[0.01, 0.40]	.071

Note. j = number of studies/clusters. k = number of effect sizes. β coefficients are from metaregression analyses where continuous moderators were entered in the models as predictors, or categorical moderators (in the case of matched age and matched education) with two levels were dummy coded and entered in the models as predictors. 95% CI corresponds to the β for moderators. p corresponds to the β for moderators. R₍₂₎² = proportion of estimated heterogeneity explained by the predictors at Level 2. R₍₃₎² = proportion of estimated heterogeneity explained by the predictors at Level 3. Q = Q statistic on the homogeneity of effect sizes. τ₍₂₎² = heterogeneity of effects due to differences between measures. τ₍₃₎² = heterogeneity of effect sizes across studies after controlling for the different types of measures at Level 2. MADRS = Montgomery–Åsberg Depression Rating Scale; HDRS = Hamilton Depression Rating Scale; BDI = Beck Depression Inventory.

p < .05.