Being silenced, loneliness and being heard: understanding pathways to intimate partner violence & abuse in young adults. a mixed-methods study

Table 4 Mediation of association between maltreatment and IPVA, by loneliness (2127 females, 1145 males, imputed data*)

	Odds ratio for the association of child maltreatment with IPVA in young adult relationships between 18–21 years old; before (total effect) and after (direct effects) accounting for mediation by loneliness (95% confidence interval)
	Females (adjusted)*	Males (adjusted)*
Total effect of child maltreatment** on IPVA	1.27 (1.05 to 1.54) P = 0.06	1.46 (1.09 to 1.94) P = 0.01
Direct effects; i.e. the association between child maltreatment* and IPVA after accounting for****:
Loneliness at school at age 14	1.27 (1.02 to 1.49) P = 0.03	1.42 (1.07 to 1.89) P = 0.02
Loneliness within peer group at age 13	1.26 (1.05 to 1.52) P = 0.02	1.47 (1.10 to 1.95) P = 0.01
Overall loneliness at age 13	1.26 (1.04 to 1.52) P = 0.02	1.46 (1.10 to 1.93) P = 0.01
All three loneliness measures	1.24 (1.03 to 1.50) P = 0.03	1.43 (1.07 to 1.91) P = 0.02

^*Pooled beta coefficients and standard errors were derived across imputed datasets using Rubin’s rules. Mean p-values across the imputed datasets are reported
^**Maltreatment = emotional abuse, emotional neglect, physical abuse, sexual abuse, domestic violence (violence between ‘parents’), bullying
^*** Adjusted models include the following covariates for females: maternal age at delivery, maternal parity at delivery, maternal education, whether the child lives in a single parent household. For males, the same set of covariates was used apart from single parent household, which could not be included due to perfect prediction in one or more of the imputed datasets
^****The ‘natural direct effect’

ISSN: 1471-2458