Libby and colleagues1 recently wrote about predictors of caregivers’ receiving services for mental health, emotional, or substance use problems. Their key finding was that American Indian caregivers are less likely than Hispanic caregivers to receive such services. This research is important for at least 2 reasons. First, there has been a dearth of research on disparities experienced by American Indian families in the child welfare system; most research in this area has focused on Black families.2 Second, it is rare for such research to focus on the services received by caregivers; most research has instead focused on the children.3
It appears that in their logistic regression analysis Libby et al. made a critical decision to treat American Indian caregivers as the reference group. This decision was somewhat unconventional. In disparity research, the usual approach is for the dominant group (most often, in the United States, Whites) to be the reference group.4 In addition, it is usually advisable to avoid using a numerically small group as the reference group, as such an approach results in unreliable estimates of odds ratios.5
This statistical decision was also consequential. A weakness of logistic regression with dummy variables is that it precludes comparisons that do not involve the reference group.5 Thus, the authors’ decision allowed comparisons only between American Indian caregivers and other caregivers and precluded comparisons between White care-givers and other caregivers. The use of dummy variables is also fraught with interpretational dangers. Although the authors assert that “Hispanic caregivers fared the best,”1(p630) the data do not support this conclusion. No comparisons were made between Hispanic and White or Black caregivers.
The authors’ unconventional approach to the dummy variables appears to have allowed findings that would not have been obtained otherwise; however, it may have obscured other important disparities. For example, a similar disparity between Black and White caregivers is likely to exist, but this is impossible to confirm with American Indians as the reference group. Given the data suggesting that Black children are less likely than other children to receive adequate services from the child welfare system,3,4 it would be useful to know whether this is also the experience of Black parents. If the authors were to rerun the analyses with either White or Black care-givers as the reference group, I would be interested in seeing the results.
Acknowledgments
The author’s work is funded by the Administration for Children and Families (grant 90-CA-1747).
The author would like to thank Angiela Zielinski and Richard H. Calica for suggestions.
References
- 1.Libby AM, Orton HD, Barth RP, et al. Alcohol, drug, and mental health specialty treatment services and race/ethnicity: a national study of children and families involved with child welfare. Am J Public Health. 2006;96:628–631. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Burns BJ, Phillips SD, Wagner HR, et al. Mental health need and access to mental health services by youths involved with child welfare: a national survey. J Am Acad Child Adolesc Psychiatry. 2004;43:960–970. [DOI] [PubMed] [Google Scholar]
- 3.Garland AF, Lau AS, Yeh M, McCabe KM, Hough RL, Landsverk JA. Racial and ethnic differences in utilization of mental health services among high-risk youths. Am J Psychiatry. 2005;162:1336–1343. [DOI] [PubMed] [Google Scholar]
- 4.Gaviria-Soto JL, Castro-Morera M. Beyond over-representation: the problem of bias in the inclusion of minority group students in special education programs. Quality and Quantity. 2005;39:537–558. [Google Scholar]
- 5.Hardy MA. Regression With Dummy Variables. Thousand Oaks, Calif: Sage Publications; 1993.