Prevalence studies assessing the problem behavior of children and adolescents with and without disabilities often have methodological limitations (e.g., they use one source of information; participants’ responses are subjective), resulting in varying prevalence estimations. The main aim of this study was to pilot a two-step procedure intended to estimate the prevalence of problem behaviors in educational settings in the most accurate manner. Two schools (a special and a regular school) participated in this study, involving a total of 897 students with and without disabilities. Educators, headmasters, and school psychologists were asked to participate in the two-step procedure by answering two questionnaires by interview dealing with their students’ problem behaviors. In the first step, informants identified students exhibiting problem behaviors, and in the second one, the type, frequency, and intensity of the identified behaviors were explored. Results indicated that the two-step piloted procedure helped to overcome wrong estimations of problem behavior prevalence, which were found to be 1.58 and 60% in regular and special education schools, respectively. Additionally, and perhaps more importantly, results stressed a high variability in prevalence rates depending on the informant asked. Implications for future lines of research are outlined.