Table 1.
Meta‐analyses of studies on antecedents of child maltreatment
| Study | Pub year | Maltreatment | Antecedent | Design | k | n | Cohen’s d | 95% CI | Homogeneity | Pub bias | ES trim | N largest study | ES largest study | Quality rating a |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Assink et al. | 2018 | CM, SR | Parent’s CM | Mixed | 84 | >>1,000 | .60 | 0.52, 0.69 | 1 | 3, k = 18 imputed | 0.70 | n.r. | n.r. | 5 |
| Chiesa et al. | 2018 | CM, SR EA | IPV | Mixed | 6 | 5,798 | .47 | n.a. | 1 | 0 | n.r. | 2,508 | 0.20 | 3 |
| Chiesa et al. | 2018 | CM, SR PA | IPV | Mixed | 15 | 8,637 | .35 | n.a. | 1 | 1 | n.r. | 2,508 | 0.10 | 3 |
| Kane et al. | 2018 | CM | Dependency of perpetrators | Mixed | 21 | 1,321 | .36 | n.a. | 0 | 0 | n.r. | 472 | 0.31 | 1 |
| Lo et al. | 2017 | CM | Insecurity | Cross | 10 | 1,090 | .51 | n.a. | 2 | 2 | n.a. | 213 | 0.41 | 3 |
| Lo et al. | 2017 | CAP | Insecurity | Cross | 7 | 740 | .69 | n.a. | 2 | 2 | n.a. | 276 | 0.67 | 2 |
| Madigan et al. | 2019 | CM, SR | Parent’s CM | Mixed | 80 | >>1,000 | .45 | 0.37, 0.54 | 1 | 2 | n.a. | n.r. | n.r. | 4 |
| Mulder et al. | 2018 | CM, SR N | Low SES | Mixed | 28 | >>1,000 | .34 | 0.13, 0.54 | 2 | 3, k = 5 imputed | 0.48 | n.r. | n.r. | 5 |
| Reijman et al. | 2016 | CM, CAP | HR baseline | Cross | 10 | 492 | .24 | 0.03, 0.45 | 2 | 0 | n.r. | 104 | 0.50 | 4 |
| Reijman et al. | 2016 | CM, CAP | ANS reactivity | Cross | 10 | 471 | −.10 | −0.36, 0.16 | 1 | 0 | n.r. | 83 | 0.00 | 4 |
| Seto et al. | 2015 | CM, SA | Parent’s CM | Mixed | 8 | 912 | .31 | 0.15, 0.47 | 1 | 2 | n.a. | n.r. | n.r. | 4 |
n.a., not applicable; n.r., not reported; CM, child maltreatment (officially reported); CAP, child abuse potential; SR, self‐reported maltreatment; EA, emotional abuse; PA, physical abuse; N, neglect; IPV, interpersonal violence; SA, sexual abuse; cross, cross‐sectional.
Quality indicators:
high;
medium;
low.
aQuality rating: overall quality score of the meta‐analysis, including the reporting of intercoder reliabilities for search, moderator coding, and data extraction; higher scores represent higher quality.