Protective Factor Screening for Social Emotional Instruction: Sensitivity and Specificity of the Devereux Student Strengths Assessment–Mini (DESSA-mini)

• In Event: Building Infrastructure for Social Emotional Learning in Routine Practice: Assessing Student Competence and Instruction Fidelity

Abstract

Objective: The World Health Organization screening principles reflect a tension between the compelling and simple idea of early detection and intervention, and the actual challenge of wide-scale screening as a practice that results in an intervention for those with previously undetected needs, and avoids harm to those who are not in need. The DESSA-Mini was designed to overcome practical obstacles to universal prevention screening, but it is unclear as to whether an entirely strength-based, 8-item screening instrument can achieve technical accuracy in routine practice in schools.

In both the academic and behavioral realms, schools often use screening instruments in multi-tiered instructional models that have untested or inadequate performance against traditional metrics. It is desirable to have all metrics (sensitivity, specificity, PPV, and NPV) above 75% (Glover & Albers, 2007). Given the current technical shortcomings of most broad-based behavioral health screening tools, Glascoe (2005) advised practitioners wishing to minimize excessive referrals that acceptable standards for screening instruments should be sensitivities of 50% and specificities of 80%, and PPV rates in the 30-50% range.

Method: Data come from a Social Emotional Learning (SEL) initiative designed to promote students’ social-emotional competence. All district students, kindergarten through Grade 8, were screened using the DESSA-Mini. A random 5% of students (n=1,960) received additional assessment using the full DESSA. Concurrent and predictive test-criterion studies are conducted using the full DESSA as well as administrative records of serious disciplinary infractions.

Results: The DESSA-Mini showed excellent internal reliability, exceeding .90. Negligible to small differences were found between scores on the DESSA-Mini screen and the DESSA full assessment. Prediction consistency between the DESSA-Mini and the DESSA was high (87-94%), with sensitivity and specificity estimates exceeding Glascoe’s practice standards and nearing Glover’s ideals. Predictive test-criterion was reliable; students screened as having an exceptional Need for SEL Instruction at the beginning of the year were 4.5 times more likely to have a disciplinary infraction record at the end of the school year compared to those who were not identified (p<0.001).

Conclusion: These findings compare quite favorably to other instruments used in schools to screen entire student populations, and are consistent with a practice preference of identifying, but not over-identifying, students for intensive SEL interventions.

Authors

• Valerie Shapiro, UC Berkeley

• B. K. Elizabeth Kim, University of California - Berkeley

• Jennifer Fleming Robitaille, Devereux Center For Resilient Children

• Paul Augustine LeBuffe, Devereux Center for Resilient Children