The participants were observed during break and classroom time as well as group and single sessions. Observation times ranged from 9.3 to 12.4 hours (Taylor & Oliver, 2008). The observers did not interact with the participating individuals and were careful to remain unnoticed. Observers made use of the ObsWin program which allowed for recording of frequency and duration of behaviours that are of interest. Environmental events of interest were attention of caregivers towards the SMS individuals, demands given by the caregivers, self-injury by the participants, and aggressive outbursts from the participants. Inter-observer reliability was calculated by the authors for 5 hours of the total observation time (Taylor & Oliver, 2008). Cohen’s Kappa was employed and calculated for the environmental events of attention and demand with values of 0.75 and 0.62, respectively. These values were deemed to signify excellent agreement between the observers (Taylor & Oliver, 2008). Data analysis was conducted using the percentages obtained from the ObsWin program and both a co-occurrence analysis and a sequential analysis were conducted between the participant’s behaviours and the external environment
The participants were observed during break and classroom time as well as group and single sessions. Observation times ranged from 9.3 to 12.4 hours (Taylor & Oliver, 2008). The observers did not interact with the participating individuals and were careful to remain unnoticed. Observers made use of the ObsWin program which allowed for recording of frequency and duration of behaviours that are of interest. Environmental events of interest were attention of caregivers towards the SMS individuals, demands given by the caregivers, self-injury by the participants, and aggressive outbursts from the participants. Inter-observer reliability was calculated by the authors for 5 hours of the total observation time (Taylor & Oliver, 2008). Cohen’s Kappa was employed and calculated for the environmental events of attention and demand with values of 0.75 and 0.62, respectively. These values were deemed to signify excellent agreement between the observers (Taylor & Oliver, 2008). Data analysis was conducted using the percentages obtained from the ObsWin program and both a co-occurrence analysis and a sequential analysis were conducted between the participant’s behaviours and the external environment