Table 5.
Moderation Analyses for the Cross-lagged Regression Between PE at Time 1 and Depression at Time 2
| Categorical moderators | Category | β | Category | β | Category | β | Q |
|---|---|---|---|---|---|---|---|
| Operationalization of PE | |||||||
| Type of PE measureb | Extraversion (k = 21) | −.07* | Positive affect (k = 30) | −.08* | Behavioral activation (k = 6) | −.05 | 1.75 |
| Time span of PE measure | State (k = 22) | −.09* | Trait (k = 36) | −.07* | 1.43 | ||
| Demographics and other sample characteristics | |||||||
| Sample ageb | Adults (k = 42) | −.09* | Children/Adolescents (k = 15) | −.04* | 8.45* | ||
| Sample sexb | More than 70% female (k = 15) | −.09* | Less than 70% female (k = 42) | −.07* | 0.33 | ||
| Type of sample I | Clinical (k = 7) | −.09* | Nonclinical (k = 51) | −.07* | 0.20 | ||
| Type of sample II | Clinical and physical health (k = 14) | −.10* | Other samples (k = 44) | −.07* | 1.45 | ||
| Restricted range I | Depressed/anxious-only samples (k = 6) | −.12* | Other samples (k = 52) | −.07* | 3.64 | ||
| Restricted range II | Samples excluding baseline depression (k = 6) | −.13* | Other samples (k = 52) | −.07* | 2.28 | ||
| Likely inter-assessment change | Yes (k = 17) | −.11* | No (k = 41) | −.06* | 8.46* | ||
| Special population | Yes (k = 11) | −.11* | No (k = 47) | −.07* | 3.20 | ||
| Study measures | |||||||
| Depression measure report | Self-report (k = 52) | −.07* | Other-report (k = 6) | −.10* | 0.55 | ||
| Depression measure distribution | Continuous (k = 56) | −.07* | Originally dichotomous (k = 2) | −.12* | 1.86 | ||
| Methodological considerations | |||||||
| Study location | U.S. and western Europe (k = 51) | −.08* | Other locations (k = 7) | −.05* | 1.92 | ||
| Data publication status | Data published (k = 21) | −.08* | Data requested (k = 37) | −.07* | 0.02 | ||
|
| |||||||
| Continuous moderators | β | Q | |||||
|
| |||||||
| Sample size | <.001 | 0.27 | |||||
| Publication year | −.002 | 0.58 | |||||
| Total number of study waves | −.001 | 0.08 | |||||
| Percent attrition | .054 | 1.00 | |||||
Note. Computations utilized a random effects model. β = standardized regression coefficient; Q = test of moderation.
b
Analysis excluded one study that did not include enough information for coding moderators or did not fall into one of the moderator categories. Additional information can be found in Supplementary Table 1.
*
p < .01. No p values were < .05.