Statistically characterizing intra- and inter-individual variability in children with Developmental Coordination Disorder.
Swap GLM for random coefficient modeling to see each child’s true motor trend.
01Research in Context
What this study did
The team looked at kids with Developmental Coordination Disorder. They wanted to see how each child’s motor skills changed from day to day.
Instead of the usual GLM stats, they used random coefficient modeling (RCM). This lets every child have their own slope and intercept.
What they found
RCM caught group differences and personal motor trajectories. The old GLM missed both.
In short, RCM showed who was really improving, staying flat, or getting worse.
How this fits with other research
Matson et al. (2013) also tracked motor variability in kids with cerebral palsy. They used a Langevin method, while Mount et al. (2011) used RCM. Both papers agree: classic tests hide individual noise.
Mueser et al. (1991) built an early single-case stat for PTSD. Mount et al. (2011) updates that idea with RCM for motor data.
Prain et al. (2012) warn that percent agreement can fake high reliability. Same warning here: GLM can fake a flat line when real life is bumpy.
Why it matters
If you track fine-motor goals, RCM gives you each learner’s unique trend. You can spot who needs a new prompt or schedule quicker than GLM ever will. Next time you graph dexterity data, ask your software for multilevel or mixed-model output. One click, clearer picture, faster clinical choice.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Re-run last month’s motor counts with a mixed-model in Excel or R and compare slopes.
02At a glance
03Original abstract
Previous research investigating children with Developmental Coordination Disorder (DCD) has consistently reported increased intra- and inter-individual variability during motor skill performance. Statistically characterizing this variability is not only critical for the analysis and interpretation of behavioral data, but also may facilitate our understanding of the processes underlying DCD. Thus, the primary purpose of this research was to demonstrate the utility of a flexible statistical technique, a random coefficient model (RCM), that characterizes the increased intra- and inter-individual variability in children with and without DCD. We analyzed data from a sensorimotor adaptation task during which participants executed discrete aiming movements under conditions of rotated visual feedback. To highlight the advantages of this statistical approach, we contrasted the results from the RCM with those from a traditionally employed general linear model (GLM). The RCM revealed differences between the two groups of children that the GLM did not detect; and, characterized trajectories of change for each individual. The RCM provides researchers an opportunity to probe behavioral deficits at the individual level and may provide new insights into the behavioral heterogeneity in children with DCD.
Research in developmental disabilities, 2011 · doi:10.1016/j.ridd.2010.12.043