What are the weaknesses to this study's design/method?

https://scholar.google.com/scholar?hl=en&as_sdt=0%2C26&q=The+Effect+of+Organized+Versus+Supervised+Recess+on+Elementary+School+Children%E2%80%99s+Participation%2C+Physical+Activity%2C+Play%2C+and+Social+Behavior%3A+A+Cluster+Randomized+Controlled+Trial&btnG=

https://journals.humankinetics.com/view/journals/jpah/15/10/article-p747.xml

["Methods: Participants: Children (N = 281; 144 girls and 137 boys; age: 8-10 y) from 14 second-grade classes of 14 different elementary schools in Flanders (Belgium) volunteered to participate. Class sizes for PE ranged between 14 and 30 children. Depending on the school's schedule, all children received one or two 50-minute PE lessons per week. Informed consent was received from children's parents. Sixteen PE teachers (3 females and 13 males) participated. Their average age was 35 years (range = 27-58), and their average PE teaching experience was 8 years (range = 2-28), of which 6 years (range = 3 mo to 18 y) occurred in the school where the study was conducted. Data were collected over 5 months. Teachers had no prior experience in teaching parkour. The study was approved by the Social and Societal Ethics Committee of the first author's university (approval number G-2015 01 152).

Instruments: Four trained observers used the System for Observing Children's Activity and Relationships during Play (SOCARP)23 to assess children's PA, play, and social behavior during recess. Physical activity was coded as either sedentary (ie, lying down, sitting, and standing), moderate (ie, equal to walking), or vigorous (ie, more energy expenditure than walking). The group size reflects the number of children that the target child is interacting with and was coded as alone, 2 to 4 people, or 5 people or more. The activity type was coded as activity games (eg, parkour, tag games, dancing, and fantasy games), locomotion (eg, transition), and sedentary (eg, sitting and standing). Social behavior and interactions were coded as either prosocial, antisocial, or neutral. Prosocial behaviors consist of positive social interactions with other children (eg, providing encouragement, support, assistance, and feedback). Antisocial JPAH Vol. 15, No. 10, 2018 748 Coolkens et al behaviors consist of negative social interactions with other children (eg, intimidation, provocation, animosity, and disapproval). Neutral behavior was coded when no prosocial or antisocial behaviors were observed during the interval. Momentary time sampling using a 6-second observe and 6-second record interval was used to code SOCARP variables. During the "observe" interval, observers focused continuously on the target child. On the "record" prompt PA, the group size and activity type of the target child was coded. For social behavior, partial interval recording was used, meaning that any social interaction the target child was engaged in during the 6-second observe interval was coded. All observations had an average duration of 20 minutes, and and each target child was observed the full duration of parkour recess.

Statistical Analysis: All data were analyzed using SPSS (SPSS Statistics for Windows, version 20.0; IBM Corp, Armonk, NY). We calculated that, for an individually randomized trial, we would need 64 students in each arm of the study to detect a 10% difference in MVPA with a SD of 10% and a power of 80%. To account for a .04 ICC between classes, 96 students would be needed in each group. To anticipate drop out, we aimed at recruiting at least 120 students in each arm. ICC between classes (schools) was .03, and therefore, analysis was conducted on the individual level. Descriptive statistics were used to report teacher and student characteristics. A chi-squared test was used to assess differences in the proportion of participation between organized and supervised recess and between boys and girls. Kolmogorov-Smirnov testing revealed violation of normality for group size and social interactions. As such, the nonparametric Kruskal-Wallis test was used to analyze differences between organized and supervised recess and between boys and girls. For normally distributed variables, 1-way analysis of variance was used. An alpha level of .05 was set as a criterion for significance." ]