Letter

In Reply: We appreciate Dr. McGregor's concerns and share his desire to see more comprehensive research addressing juveniles incarcerated in adult prisons. When we began our research, no empirical studies had addressed the mental health needs of this group. Thus we faced a dilemma common to researchers approaching any understudied area: Should we wait until we can conduct "the perfect study," or should we begin with a broad "screening" that might identify where further inquiry is needed? This research dilemma is analogous to the clinical dilemma common in the juvenile justice system: Should we attempt a comprehensive assessment of every youth, or should we conduct a broad screening to identify which youths—or which concerns for a particular youth—warrant further assessment? Often, the only practical option in either context is to begin with broad screening.

Because our goal was a broad overview of treatment needs among youths in adult prisons—particularly as compared with youths in juvenile justice facilities—we selected the only measure that allows for comparison to a nationally representative, ethnically diverse sample of youths in juvenile justice facilities ( 1 ). The MAYSI-2 ( 2 ), more than other measures, allows us to compare "apples to apples" when investigating self-reported symptoms among juveniles across various justice contexts.

As Dr. McGregor notes, the MAYSI-2 does not provide formal diagnoses and cannot rule out symptom exaggeration or symptom minimization. Participants may have reported some symptoms for secondary gain—a possibility in any study of incarcerated juveniles or adults. It was perhaps less likely in our sample of youths, who had already been adjudicated and could not leave the Youthful Offender Program until they reached a certain age, which minimized the opportunity for secondary gain. Most of the limitations of MAYSI-2 apply to other self-report measures, and as we noted in our article, we must accept these limitations if our results are to remain in "a common metric" for comparison to juvenile justice studies. Might we learn additional detail from studies that use clinical interviews and structured personality measures? Absolutely! We hope that we—or Dr. McGregor or others—can create opportunities for research that uses such comprehensive assessments. But just as clinicians administer a screening measure to form a general impression of a patient's functioning and where further assessment is needed, we administered a screening measure to form a general impression of a population's functioning and the need for further assessment. Our study revealed many self-reported symptoms, so we certainly agree with Dr. McGregor that further research using other assessment strategies is essential.

Finally, Dr. McGregor suggests that we report results by race-ethnicity. We certainly agree with the spirit of his comment. Indeed, one of us (GMV) addressed this important question in the MAYSI-2 national normative data ( 1 ). In our study, we matched our sample and the comparison sample for race-ethnicity, but we chose not to present race-ethnicity analyses because small cell sizes for some groups (eight white participants) might have yielded misleading results. Disparities, too, warrant further assessment in this population, but only with subsamples large enough for race-ethnicity analyses to be meaningful. Instead, we refer readers to the large MAYSI-2 normative data study, which found that race effects on the MAYSI-2 scales were usually negligible.