Categories or Dimensions? A Critical Systematic Review of Taxometric Evidence in Autism, ADHD, and Language Disorders

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

This article is not in any list yet, why not save it to one of your lists.
Log in to save this article

Abstract

Background. Taxometric analysis is designed to test whether the latent structure underlying clinical conditions is categorical (taxonic) or dimensional. Despite its relevance to ongoing debates on neurodevelopmental conditions, its application to this field remains limited. Methods. We conducted a preregistered systematic review (OSF) searching PsycINFO, PubMed, Scopus, and Web of Science from inception to April 2025, with a supplementary search made in January 2026. We included peer-reviewed empirical studies applying at least one taxometric procedure (e.g., MAMBAC, MAXEIG, MAXCOV, L-Mode) to neurodevelopmental conditions/traits and reporting a categorical versus dimensional conclusion (or sufficient indices to infer it). Two reviewers independently screened and extracted data. Study quality (risk of bias for taxometric inference) was appraised using an ad hoc checklist focused on key threats, especially artificial admixture/compound sampling, and on reporting of indicator validity, nuisance covariance, and skewness. Results. We identified 126 records across the main database search (n=110) and the supplementary search (n=16). Fourteen studies were included. ADHD and language impairment studies largely supported dimensional models, whereas most autism and subtype studies reported taxonic patterns. Artificial admixture was very common in studies reaching taxonic conclusions, especially for autism. In Monte Carlo simulations, admixture consistently generated false-positive taxonic findings from dimensional data. Conclusions. Based on the best available evidence, ADHD and language impairment appear continuous dimensions rather than discrete categories. Evidence for a categorical structure is stronger for autism, but a definitive conclusion is frustrated by sampling-related biases, especially artificial admixture. More population-based, methodologically transparent taxometric studies, especially in relation to autism, are needed.

Article activity feed