The long reach of child maltreatment: Evidence for intergenerational transmission
Widom, C. S., Czaja, S. & DuMont, K. (2015) Intergenerational transmission of child abuse and neglect: Real or detection bias? Science, 347(6229), 1480-1484.
Do abused and neglected children grow up to abuse and neglect their children? For many years, clinicians and researchers alike have largely assumed that intergenerational transmission of maltreatment was a common, albeit not inevitable, consequence of maltreatment. Over the years, increasing evidence demonstrated a fairly strong connection between maltreatment experienced in childhood, and maltreating children as an adult. These studies were typically weakly designed as cross-sectional or retrospective surveys, and thus, their conclusions were limited. To answer the question of whether or not maltreating is intergenerational, a study design requires longitudinal data with a control or comparison group.
Widom and her colleagues set out to begin to answer this question in 1986, collecting data over thirty years, and following a simple model of intergeneration transmission [Generation 1 Generation 2 Generation 3]. First, they collected data on abused and/or neglected children (N=908) who were matched with non-maltreated children of similar socio-economic status (N=667). In this study, these children were called the second generation, or G2. G2 child cases of abuse and neglect were identified from court records, containing some parent information (or G1 information), from a metro area of the Midwestern United States during 1967 to 1971. Non-maltreated G2 children (the comparison group) were sampled based on birth records and school records that matched, as closely as possible, the demographic characteristics of G2 children who were abused or neglected. All identified G2 children were under 12 years old during the sampling timeframe. Second, they followed these G2 children to adulthood, noting their appearance in subsequent criminal records. Third, they conducted detailed in-person interviews with 1,196 of the G2 sample from 1989 to1995, and they re-contacted and re-interviewed 896 of those respondents from 2000 to 2002. In 2009, they began data collection for this study on G3, or the offspring of G2. CPS records were checked for all known children of G2 sample members (N=999) from 2011 to 2013, and 70% of those children were interviewed (n=697) in 2009. The mean age for the G3 offspring at the time of interview was 22.8 years.
As a result of their analyses from 2011 to 2013, Widom et al. found that G2 adults who had experienced abuse and neglect as children (as documented through court records) were found to be significantly more likely to be reported to CPS as parents, compared to G2 adults who had not experienced maltreatment as children (AOR=2.01, p<.001). Notably, these groups did not differ significantly on CPS reports related to physical abuse, or self-reported maltreatment. G2 adults who had experienced documented maltreatment as children were also more likely to have a child placed in custody of the court (4.8%, AOR=3.77, p<.05, compared to 1.3% of the comparison group). The authors also found, in comparing self-report to CPS reports, suggestions of surveillance bias. The G3 generation who reported being abused and neglected (through self-report) and who had parents who were maltreated were over twice as likely to have an official CPS report (AOR=2.28, p<.003), compared to G3 individuals who reported abuse but whose parents did not have documented histories of abuse.
This study is the first to rigorously test the intergenerational transmission of maltreatment. It strongly suggests that the predicted pattern of intergenerational transmission of maltreatment is more complex than previously theorized or observed. First, the authors did not find any evidence to show intergenerational transmission in cases of physical abuse. These findings were consistent across data sources: G2 self-report, G3 self-report, and CPS reports on G3. Second, the authors observed that parents with documented histories of experiencing neglect or sexual abuse in childhood were significantly more likely to maltreat their children. Third, the authors underscored the need for multiple data sources when examining intergenerational transmission. They suggest that over-reliance on parental self-report may skew estimates. In calling for more research on the subject, the authors highlighted the fact that it cannot be determined if families who are subject to CPS intervention are more “dysfunctional,” or if they are simply surveilled at a higher rate. The findings suggest that CPS is not detecting a large number of children who experience abuse or neglect in a family with a non-maltreated parent.
There are multiple strengths of this study, nonetheless, some caution must be taken by the reader in interpreting the findings. First, the cases of the maltreating G2 individuals were drawn from court records, and likely represent only severe child abuse and neglect cases. They do not represent less severe cases, or cases that were not detected by CPS. Second, the matching procedure is rough and is not statistically rigorous, calling into question any causal inferences. Third, because the abused and/or neglected G2 children were matched to non-maltreated children of similar SES backgrounds, the sample overall represents a disadvantaged SES group, limiting the generalizability.