Background: Observation studies of students with attention-deficit hyperactivity disorder (ADHD) problems in natural classroom situations are costly and relatively rare.
Aims: The study enquired how teacher ratings are anchored in actual student classroom behaviours, and how the behaviour of children with ADHD problems differs from their classmates. The authors attempted to broaden the usual focus on disruptive and inattentive behaviours to elucidate the role of various on-task behaviours, as well as considering differences between classroom contexts.
Sample: DSM-III-R criteria were used in conjunction with a teacher rating scale to select a sample of 55 students with ADHD problems, and 55 matched controls from a population of 569 primary school students.
Method: Students were observed in their natural classrooms using the Munich Observation of Attention Inventory (MAI; Helmke, 1988). Correlations between teacher reports and observation codes were computed, and systematic differences between students with ADHD problems and controls in different classroom contexts were examined using a generalized linear mixed model (GLMM).
Results: Global teacher reports showed moderate to strong correlations with observed student behaviours. Expected on-task behaviour demonstrated the strongest relationship (r>-.70) with teacher reports. As hypothesized, the children with ADHD were more disruptive and inattentive than their matched peers. They were also less often inconspicuous on-task as expected by their teachers. However, their behaviour was assigned to two other on-task categories more often than their peers, and this raised their total on-task behaviour to over 66%. Situational differences were found for all codes as well, which mostly affected all students in a similar way, not just students with ADHD.
Conclusions: ADHD related behaviours are pervasive across the classroom situations coded. Teachers appear to distinguish between desirable and undesirable on-task behaviours. Nevertheless, assisting students with ADHD problems requires shaping both. Future studies need to include more differentiated codes for various types of on-task behaviours and also need to code the lesson context concurrently.