An analysis of the methodological and human error within momentary time sampling data collection

Abstract

Teachers often serve as data collectors for the problem behavior of referred students in their classrooms; yet, the accuracy of teacher data collection has rarely been directly assessed. Momentary time sampling (MTS) may be a potentially useful option for teacher data collection because it does not require continuous monitoring, but rather requires the teacher to score the occurrence or non-occurrence of targeted behaviors at given instances. Research has shown that the smaller the interval between observations, the less methodological error will be introduced into MTS. However, the use of short-interval windows requires additional effort on the part of the teacher, and data collection becomes potentially more susceptible to competition with the teacher's other responsibilities. It is not clear based upon previous research to what extent human error influences the accuracy of MTS data. Therefore, the purpose of the current study was to evaluate the amount of methodological, human, and total error introduced during MTS data collection within two highly controlled experimental contexts, in which the duration of occurrence was determined. In highly controlled settings, results demonstrated that the amount of methodological and total error tended to increase as the MTS interval became longer and that human error was observed to be low across all MTS intervals.