Sessions were conducted within a junior high classroom at a therapeutic day school for children with a primary diagnosis of autism who had been referred to this school for the presence of severe problem behaviors that could not be managed within the public school setting. Four of the five students in the classroom were included in the study due to the specialized programming of one of the students. All students had a significant intellectual disability and an individualized education plan (IEP) stating that they benefit from a small class size (i.e., 6–7 students) and individual support. Bradley was a 14-year-old White boy who engaged in self-injury and physical aggression. Jason was a 13-year-old White boy who engaged in physical aggression, self-injury, and off-task behavior. Andrew was a 13-year-old White boy who engaged in physical aggression and tantrum behaviors. Tara was a 14-year-old Black girl who engaged in self-injury to the head, physical aggression, and emesis.

One licensed special education teacher and four paraprofessionals were included in the study. Staff members included four men and one woman and all had been employed at the day school for 7–29 months (M = 18.8). Staff experience prior to starting at this school varied from no experience to more than five years. The classroom teacher had a bachelor’s degree, one paraeducator had a master’s degree, two paraeducators had a bachelor’s degree and one had a high school diploma. All staff were trained using behavior skills training in behavior management systems and behavior support techniques which included training on student-specific programming, as well as staff expectations in the classroom and school as a whole.

All observations occurred during group reading lessons and consisted of presenting a nonfiction book (differentiated to their reading level, grade one) and follow-up comprehension questions led by the classroom teacher.

Students and staff were observed up to four times daily. At the beginning of each day, the head teacher provided the daily schedule outlining the possible opportunities for observations to the first author. To control for extraneous variables, observation times and teacher–student ratios were selected using a randomized drawing. Once a ratio was chosen, it was removed from that day’s future drawings. Owing to state regulations requiring the presence of a staff member with a special education license, the head teacher participated in all observation sessions. Paraprofessionals participating in the observation were drawn before each session. Five minutes before the observation was to occur, the head teacher and staff were notified of the ratio and the staff participating in the observation.

The school’s board-certified Behavior Analyst (BCBA) and two graduate students trained in the data collection procedure conducted all observations. Observations only occurred on days when all participants were present. One observation of each learning session started ten minutes after a reading lesson had begun. A momentary time-sampling procedure was used to extend previous research and evaluate the effects of varying teacher–student ratios in a special education classroom within a therapeutic day school. The observer scored categories of classroom environment, classroom organization, student activity, and staff activity using a momentary time-sampling procedure. Observers did not interact with staff or students during data collection. Data collection sheets and definitions of targets were provided to observers and are available upon request from the corresponding author. Similar to Grauerholz-Fisher et al. (2019), there was not a specific time limit specified for each observation. When observations began, observers would enter a room and the primary data collector would indicate the student to be observed, scanning right to left. The order of variables observed for each student was: (a) classroom environment, (b) classroom organization, (c) student activity, and (d) staff activity. Once both observers scored each of the variables, the primary data collector would nod to the next student to be observed. This would continue until all students in the classroom had been observed and the observation session concluded.

The classroom environment was scored as the observer scanned the classroom from right to left specifically noting two areas, safety and supplies/materials. Observers first noted the total number of students in the room, followed by the total number of students who were displaying accurate responses. Safety was scored ( +) if the student was seated, has a clear pathway to classroom exit in case of emergency, and all sharp objects are out of reach, including scissors, pens, pencils if not in use. A score ( −) was given if a student was out of their seat, if the pathway to classroom exit was blocked, or if sharp objects were within reach, including scissors, pens, pencils, if not in use. Supplies/Materials score of ( +) was given if all materials were present (i.e. token economy, escalation visual, communication system, and student specific safety materials described in the student’s behavior plan). A ( −) score was given if any of the materials were missing.

All variables within classroom organization were scored as a permanent product. The classroom organization was scored as the observer scanned the classroom from right to left noting mini-schedules and completed expectations. Mini-schedules were scored ( +) if the student’s mini-schedule was present, listed session expectations, and was within the student’s visual field. A score of ( −) was noted if the mini-schedule was absent, did not list session expectations, or was outside of the student’s visual field. Among the completed expectations scores, a ( +) was noted if completed expectations were crossed off of the mini-schedule once completed by student. A ( −) score was noted if completed expectations had not been crossed off of the mini-schedule once completed by the student. Expectations represented on the mini-schedule were the number of pages being read aloud as a group and the number of comprehension questions the student was required to answer.

Student activity was scored as the observer scanned the classroom from right to left noting two areas, appropriate academic task and engaging in target problem behavior Appropriate academic tasks were scored ( +) if the student was engaging in a task/activity that was listed on the mini-schedule. For activities that did not require physical activity, such as reading, student eye gaze must have been on material(s). A score of ( −) was noted if the student was engaging in any task/activity that was not listed with an expectation displayed on the mini-schedule. For activities that did not require physical activity, such as reading, student eye gaze was not on the material(s). Engaging in problem behaviors included a ( +) score when the student was engaging in any target problem behavior listed in the student’s behavior intervention plan. These behaviors included: self-injury (Bradley: head to fixed object or knee to head; Jason: biting; Tara: closed or open fist to head), physical aggression (Bradley: head to person’s body; Jason: biting; Tara: scratching and/or pinching; Andrew: open hand to person’s body), off-task behavior (Jason: not engaging with or attending to assigned tasks), and tantrum behavior (Andrew: crying, whining, other loud vocalizations, flipping desk over, lying on the ground). A score of (–) was noted if the student was not engaging in any target problem behavior listed in the student’s behavior intervention plan.

The classroom environment was scored as the observer scanned the classroom right to left noting three activities, including standing, appropriate student interaction, and behavioral intervention targets. Standing received a ( +) score if the staff was standing and a (–) if they were sitting. Appropriate student interaction included a ( +) score if the staff was interacting with at least one student in regard to an appropriate academic task. A (–) score was noted if the staff was not interacting with at least one student or interacting with students about something other than an appropriate academic task. Behavioral intervention scores included a ( +) if staff was engaging in any reactive behavior intervention protocol outlined in the student’s behavior intervention plan and a ( −) if the staff was not engaging in any reactive behavior intervention protocol outlined in the student’s behavior intervention plan.

The following teacher–student ratios were evaluated; 5:4, 4:4, 3:4, and 2:4. Ratio 5:4 included the classroom teacher, four paraprofessional staff members and four students. Ratio 4:4 included the classroom teacher, three paraprofessional staff members and four students. Ratio 3:4 included the classroom teacher, two paraprofessional staff members and four students. Ratio 2:4 included the classroom teacher, one paraprofessional staff member and four students. Each ratio was observed 10 times across the duration of the study.

For classroom environment, organization, student activity, staff activity, and all subcategories, percentage of observations was calculated by adding the total number of “ + ” responses observed, dividing by the total possible responses, and multiplying by 100. Percentage of observations was calculated for all observed teacher–student ratios.

Interobserver agreement was assessed by having a second observer conduct simultaneous but independent observations for 38% of sessions. During reliability observations, the lead observer would guide the second observer into the classroom and following the signal of a timer both observers would scan the room right to left. Each observer first observed the classroom environment, then classroom organization, followed by student activity, and finished the observation session with staff activity. Once the observers had completed assessing all categories and subcategories they would signal each other with a nod and simultaneously exit the classroom. Interobserver agreement was calculated for all categories by dividing the total number of agreements by the total number of agreements possible (both accuracy and violations). Mean agreement scores were 100%, for classroom environment, 100% for classroom organization, 98% for student activity (range 75–100%), and 97% (range 67–100%) for staff activity.