Research on Hierarchical Agglomerative Cluster Analysis as a Trial Method Revealing Developmental Coordination Disorder (DCD) Subtypes

The present study investigated the nature of possible cognitive-motor profiles of DCD using clustering
methods. Dependent variables were selected on the basis of the characteristics of children with DCD
and the specific difficulties observed in cognitive- motor domain according to the DCD literature. For
the purpose of the study we adopted “PASS” neurocognitive theory (Planning, Attention,
Simultaneous, Successive) and the norm-referenced Cognitive Assessment System. Based on this
hierarchical agglomerative cluster analysis six (6) statistical sub-groups emerged with number of
participants ranged from 5-43 students with or without DCD. Internal and external validity of the
clustering solution was controlled by different clustering methods (Wards method analysis, Complete
Linkage method, Centroid method, K-Means iterative partitioning method and split-sample
replication), as well as other parametric methods (MANOVA, ANOVA and discriminant analyses). The
impact of different DCD profiles may provide larger benefits for alternative and effective instructional
methods and early intervention programs in order to avoid motor learning disabilities and low
academic achievement. Future research in evaluating and designing intervention programs may be
need to focus on the individual profiles of children across a broad range of areas (motor, cognitive,
social and emotional), looking at their unique strengths and weaknesses.