Article

“Mirror” Language AI Models of Depression are Criterion-Contaminated

2025-08-07

Abstract excerpt

<p>A growing number of studies show near-perfect LLM language-based prediction of depression assessment scores (up to R2 of .70). However, many are developed from language responses directly to the depression assessments. These “Mirror models” suffer from “criterion contamination,” which arises when a predicted score depends in part on the predictors themselves. This causes artificial effect size inflation which r...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work: a92ab810-3403-5de2-b2f1-7640659b67af
DOI: 10.31234/osf.io/fe2z9_v1

Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest

“Mirror” Language AI Models of Depression are Criterion-ContaminatedDOI 10.31234/osf.io/fe2z9_v1Read

Using a Fine-tuned Large Language Model for Symptom-based Depression EvaluationSemantic proximity Read

Multi-Criteria Validation of LLM-Inferred Depression Severity from Outpatient Psychiatry NotesSemantic proximity Read

Exploring the Generalizability and Explainability of LLMs in Detecting Suicidal Ideation: The Impact of Data HeterogeneitySemantic proximity Read

Digital shadows in mental health map how LLMs simulate depression, anxiety, and stress through language and psychometricsSemantic proximity Read

Digital shadows in mental health map how LLMs simulate depression, anxiety, and stress through language and psychometricsSemantic proximity Read

Digital shadows in mental health map how LLMs simulate depression, anxiety, and stress through language and psychometricsSemantic proximity Read

Mental Health Benchmarks for Large Language Models: A Systematic Scoping ReviewSemantic proximity Read

Prediction of depression risk based on AI art promptsSemantic proximity Read

Inter-rater Reliability of an LLM in Predicting Depression Among Indian AdultsSemantic proximity Read

The threat of synthetic respondents extends to clinical mental health screeningSemantic proximity Read

Language-Based AI Modeling of Personality Traits and Pathology from Life Narrative InterviewsSemantic proximity Read

Research Letter: Application of GPT-4 to select next-step antidepressant treatment in major depressionSemantic proximity Read