Development of a Brief Rating Scale for the Formative Assessment of Positive Behaviors

• Main Author: Cressey, James
• Format: Others
• Published: ScholarWorks@UMass Amherst 2010
• Subjects: Behavioral Assessment; Formative Assessment; General Outcome Measure; Positive Behavior Support; Response to Intervention; Screening; Psychology
• Online Access: https://scholarworks.umass.edu/open_access_dissertations/248; https://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1208&context=open_access_dissertations

In order to provide effective social, emotional, and behavioral supports to all students, there is a need for formative assessment tools that can help determine the responsiveness of students to intervention. Schoolwide positive behavior support (SWPBS) is one framework that can provide evidence-based intervention within a 3-tiered model to reach students at all levels of risk. This dissertation begins the process of developing a brief, teacher-completed rating scale, intended to be used with students in grades K-8 for the formative assessment of positive classroom behavior. An item pool of 93 positively worded rating scale items was drawn from or adapted from existing rating scales. Teachers (n = 142) rated the importance of each item to their concept of "positive classroom behavior." This survey yielded 30 positively worded items for inclusion on the pilot rating scale. The pilot scale was used by teachers to rate students in two samples drawn from general education K-8 classrooms: a universal tier group of randomly selected students (n = 80) and a targeted tier group of students with mild to moderate behavior problems (n = 82). Pilot scale ratings were significantly higher in the universal group than the targeted group by about one standard deviation, with no significant group by gender interaction. Strong results were found for the split-half reliability (.94) and the internal consistency (.98) of the pilot scale. All but two items showed medium to large item-total correlations (> .5). Factor analysis indicated a unidimensional factor structure, with 59.87% of the variance accounted for by a single factor, and high item loadings (> .4) from 26 of the 30 factors. The unidimensional factor structure of the rating scale indicates its promise for potential use as a general outcome measure (GOM), with items reflecting a range of social, emotional, and behavioral competencies. Future research is suggested in order to continue development and revision of the rating scale with a larger, more diverse sample, and to begin exploring its suitability for screening and formative assessment purposes.