School administrators mailed a cover letter and consent form home to all parents of students in the elementary school division (n = 24). Parents who wished to have their child participate in the study were asked to return signed consent forms to the principal investigator. Teachers were consented and asked to complete a baseline booklet of questionnaires for each participating student. Students were shown a social story about the study and told how they could withdraw their participation verbally or nonverbally. Thirteen of the 24 invited students enrolled in this study, representing a response rate of 54%. This study was approved by the Institutional Review Board at the University of Massachusetts Amherst prior to recruitment and data collection.
The present study spanned six days across a 16-day period in March 2017. On each of the six days of data collection, students participated in their regularly scheduled, structured jogging period for 20 min (9:30–9:50). Students from all classrooms in the elementary school division jogged together in a hallway alongside their teachers. Students ran laps around cones placed at each end of the hallway. A video camera stood behind each cone. Students went to homeroom from 9:50 to 10:25. From 10:25 to 10:45, children enrolled in the study participated in a new, unstructured jogging period for 20 min. During this time, students were asked to jog around a circle of cones in the school gym while teachers encouraged them verbally from a distance. A video camera was placed on both sides of the gym. The purpose of the audio and visual recording was to confirm attendance for each student and to observe if any students left the room or displayed any behaviors that affected their participation. The recordings were also used to count the number of verbal prompts given to the students, to ensure that level of prompting was consistent across music conditions. Students wore accelerometers on their waistband on each of the six data collection days. Teachers put the accelerometers on students after they changed into their uniforms in the morning (approximately 9:15) and removed them after the unstructured jogging period (approximately 10:45). Two practice days were completed prior to data collection to habituate students to the accelerometers and video cameras.
Each of the music conditions (i.e., no music, slow music, and fast music) was implemented on two days of data collection for a total of six days of data collection. The order of the music conditions was randomized, but all conditions appeared once before a condition could be repeated. Music conditions appeared in the following order: slow (day 1), fast (day 2), none (day 3), fast (day 4), none (day 5), slow (day 6). The music condition was held constant on a given day to avoid contamination effects. In other words, if students heard fast music during the structured exercise period from 9:30 to 9:50, they heard fast music again during the unstructured exercise period from 10:25 to 10:45.
Playlists of slow and fast music were created for use during the 20-min exercise periods. Target heart rates for vigorous-intensity exercise were calculated based on guidelines from the Centers for Disease Control and Prevention (CDC). For vigorous intensity physical activity, a person’s heart rate should be 70–85% of their maximum heart rate. Using the average participant age (age 9), the estimated age-related maximum heart rate was calculated as 220 − 9 years = 211 bpm. The 70% and 85% levels were calculated as 211 × 0.7 = 147.70 bpm and 211 × 0.85 = 179.35 bpm, respectively. In other words, music ranging from 147.70 to 179.35 bpm would match the average age participant’s target heart rate for vigorous activity. The fast music playlist was composed of five songs with 144 to 160 bpm to reflect this interval. A slow music playlist was composed of six songs with bpm significantly below this interval (60–80 bpm). Songs on each playlist were shuffled randomly and burned to CDs. Students heard the same playlist in the same order on a given day of data collection (e.g., during the structured and unstructured exercise periods), but the order of the songs was shuffled on the second day of data collection in the same music condition.
During the structured jogging period, teachers used verbal and physical prompts when necessary to keep each student jogging. They started by using verbal prompts that address the group as a whole, and then moved to verbal prompts that address specific students. Verbal prompts included phrases like “keep it up everyone”, “let’s go”, and “come on”. After verbal prompts, they began using physical prompts, starting with the least physical. These prompts included gestures, pointing, touch, physical guidance, and hand-over-hand. This prompting protocol was established by the school long before this study. Teachers were asked to continue using this protocol during the structured jogging period. By contrast, however, teachers were asked to refrain from physical prompts during the unstructured jogging period, unless it was necessary for the safety of a student. Students did not participate in either the structured or unstructured exercise periods if they were not in good physical health (e.g., had a fever, cold, or other significant illness requiring the nurse’s attention), in line with school protocol.
To ensure that the level of verbal prompting was similar across music conditions, a research assistant blind to the study’s hypotheses counted the number of verbal prompts for each video recording. There were two video cameras recording during each exercise period, each from different angles. A second research assistant blind to the study’s hypotheses coded five randomly selected video recordings. The number of prompts coded by the first research assistant was averaged across cameras and compared by music condition.
2.4. Analytic Plan
To address the first research question about the impact of music on exercise intensity, a series of one-way repeated measures ANOVAs with Bonferroni post-hoc tests were conducted with music condition (i.e., no music, slow music, and fast music) as the independent variable and exercise intensity as the dependent variable. Parallel analyses were conducted with average METs and the percentage of intervals of vigorous-intensity activity as the dependent variables. Analyses were conducted separately for data collected during the structured and unstructured exercise periods.
To address the second research question about potential moderators of this relationship, the analyses described above were repeated with student characteristics (i.e., age, gender, BMI, adaptive behavior, maladaptive behavior, autism symptom severity) as covariates. In addition to main effects, the interaction of music condition by each of the student characteristics was examined (e.g., music × gender, music × age).