Rigorous evaluation of social and emotional learning programs requires the use of measures that provide reliable and valid information on the meaningful differences in children's social emotional skills across treatment and control groups, as well as changes over time. In contexts affected by conflict and crisis, few measures can provide the evidence required to support their use in program evaluations, which limits stakeholders' ability to determine whether a program is working, how well it is working, and for whom. The Teacher Observation of Learners' Social Emotional Learning, known as TOOLSEL, holds promise for addressing this gap. TOOLSEL is a teacher-report questionnaire about children's behavior as observed in natural classroom settings. It is used to assess a set of social, emotional, behavioral, and cognitive competencies among primary school-age children in fragile, conflict-affected settings. In this article, using the data from a sample of 3,661 Syrian refugee children who were enrolled in formal Lebanese public schools and had access to a nonformal remedial support program, we report evidence on the psychometric soundness of the TOOLSEL. We provide empirical evidence of the TOOLSEL's reliability and validity, and that the TOOLSEL captured these Syrian refugee children's social and emotional learning skills in ways that were unbiased and comparable across treatment groups, gender, age, and time. We also provide recommendations for using the TOOLSEL, including ways to improve its feasibility, reliability, and validity.