Statistical modelling studies examining the dimensional structure of psychopathology experienced by adults with intellectual disabilities: Systematic review.
Current factor-analytic studies of psychopathology in adults with ID use weak methods—demand better stats next time you review or conduct one.
01Research in Context
What this study did
The authors hunted every paper that used factor analysis or other fancy stats to map mental-health symptoms in adults with intellectual disability.
They read the methods sections like detectives. They asked: Did the researchers check if their math fit the data? Did they test if the same model works for different groups? Most studies failed these checks.
What they found
Almost every study used weak statistics. Models looked neat on paper but fell apart when tested on new people.
Because of this, we still do not know if problem behaviors, mood swings, or anxiety cluster into clear dimensions in this population.
How this fits with other research
van Timmeren et al. (2016) — the very same year — ran a factor analysis and said two clean dimensions explain behavior and mood in adults with ID. The review says that study is one of the shaky ones.
Christopher et al. (1991) found six behavior dimensions in a huge sample, hinting that simple two-factor models miss nuance. The new review backs that warning.
Goodwin et al. (2012) listed ready-made scales for clinicians. The review adds: pick a scale only if its factor structure has been double-checked with strong stats.
Why it matters
Before you trust any questionnaire that promises to sort "internalizing" from "externalizing" in your adult clients with ID, ask to see the cross-validation data. If the authors have not rerun the stats on a fresh sample, treat the cut-offs as guesswork, not gospel.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Email the test publisher and ask for the factor-analysis validation report before you give that symptom checklist again.
02At a glance
03Original abstract
Diagnosing mental ill-health using categorical classification systems has limited validity for clinical practice and research. Dimensions of psychopathology have greater validity than categorical diagnoses in the general population, but dimensional models have not had a significant impact on our understanding of mental ill-health and problem behaviours experienced by adults with intellectual disabilities. This paper systematically reviews the methods and findings from intellectual disabilities studies that use statistical methods to identify dimensions of psychopathology from data collected using structured assessments of psychopathology. The PRISMA framework for systematic review was used to identify studies for inclusion. Study methods were compared to best-practice guidelines on the use of exploratory factor analysis. Data from the 20 studies included suggest that it is possible to use statistical methods to model dimensions of psychopathology experienced by adults with intellectual disabilities. However, none of the studies used methods recommended for the analysis of non-continuous psychopathology data and all 20 studies used statistical methods that produce unstable results that lack reliability. Statistical modelling is a promising methodology to improve our understanding of mental ill-health experienced by adults with intellectual disabilities but future studies should use robust statistical methods to build on the existing evidence base.
Research in developmental disabilities, 2016 · doi:10.1016/j.ridd.2016.01.018