Sinha, V., Trocme, N., Fallon, B., & MacLaurin, B. (2013). Understanding the investigation-stage overrepresentation of First Nations children in the child welfare system: An analysis of the First Nations component of the Canadian Incidence Study of Reported Child Abuse and Neglect 2008. Child Abuse & Neglect, (in press). (6): 740-750.
The overrepresentation of Aboriginal children in child welfare systems is well documented. The present analysis, part of a continuing effort to disentangle factors that contribute to this overrepresentation, uses data from the First Nations component of the Canadian Incidence Study (FNCIS-2008). This analysis starts with overrepresentation at the original referral for investigation, in contrast to previous studies, which have generally only focused on the later agency substantiation stage.
The present sample comprised new maltreatment-related investigations opened between October 1 and December 31 in 2008 (n=3,106 First Nations children; n=12,240 non-Aboriginal children). Cases were drawn from 111 agencies, 89 provincial/territorial (a stratified random sample) plus 22 Aboriginally-governed (a purposive sample, thought to be capable enough for study participation), in each case representing ¼ of the agencies that conduct child welfare investigations in Canada. Variables considered were referral source, focus of investigation, child demographics and functioning issues, caregiver and household risk factors, and case dispositions. Bivariate analyses were performed for each variable, comparing the incidence (rates per 1000 children) for First Nations and non-Aboriginal children, and calculating the “disproportionality ratio” of these incidence rates.
The incidence of investigations in every maltreatment category was found to be significantly higher for First Nations children than for non-Aboriginals, and a higher proportion involved non-professional referrals. The overrepresentation was highest for the neglect category and lowest for physical abuse. This significant disproportionality was found to be associated with highly significant differences in many risk categories: for example, for First Nations and non-Aboriginals respectively, low income was found in 53.6% and 31.8% of investigations, substance abuse in 55.7% and 25.3%, and domestic violence in 43.7% and 31.6%. The authors argue that, if this observed overrepresentation is to be reduced, child welfare agencies must be equipped to help families address these risk factors.
This overrepresentation at the investigation stage was compounded at the later case disposition stage, though the disproportionality here was not as great. Ratios ranged from 1.2 for substantiation of maltreatment (42.5% vs. 35.2%, First Nation and non-Aboriginal cases respectively) to 3.0 for placement in substitute care (9.7% vs. 3.2%). As a result of the compounding effect of investigation overrepresentation and placement overrepresentation, the incidence of placement in the jurisdictions served by the sampled agencies was 13.6 per 1,000 First Nations children, more than 12 times the rate of 1.1 per 1,000 for non-Aboriginal children. The authors suggest that these differences point to the need for complementary research on practice models in Aboriginally governed agencies.
A better understanding of the reasons for this observed overrepresentation is clearly needed, to guide possible future corrective action. However, there appear to have been some methodological issues in this present study that made this difficult. For example, an elaborate weighting and geocoding procedure was described, “appropriate for use in . . . comparisons of Aboriginally-governed and provincial/territorial agencies.” Such a comparison would appear to be important, but apparently was not attempted in this study. Also, the authors point out that the purposive sampling of Aboriginally-governed agencies means that these results cannot be generalized beyond the study sample. A further limitation is the fact that only bivariate analyses were used here; multivariate methods should have made it easier to tease out the relative effects of community risk factors and agency procedures on the observed disproportionality.