Marina Ridchenko | UXR Portfolio

Learn more

Can Cognitive Measures Add Information Beyond Self-Reported ADHD Symptoms?

Objective
ADHD is commonly assessed using self-report symptom questionnaires, but these instruments capture only one aspect of the condition. I wanted to explore whether cognitive and language-related measures provide additional information when predicting ADHD diagnosis and whether combining multiple forms of measurement improves understanding of ADHD-related outcomes.

My Role
I designed the research framework, selected theoretically motivated and validated measures, developed the modeling strategy, and interpreted the findings.

Study Design
To capture different dimensions of ADHD-related functioning, I combined:

  • validated symptom measures (BAARS, PHQ),
  • cognitive measures (working memory, declarative memory, procedural learning),
  • language-learning measures,
  • demographic controls.

Rather than evaluating a single model, I designed a hierarchical comparison framework that progressively introduced additional sources of information to examine their contribution.

Methods
I compared multiple interpretable and predictive approaches (Logistic Regression, Lasso, LDA, QDA, Random Forest, XGBoost) and prioritized model stability and generalization rather than selecting the highest-performing model.

Findings
Models that combined multiple information sources produced stronger predictive performance than cognitive measures alone. Including symptom-report data improved prediction, but explainability analyses suggested that no single variable dominated the outcome.

Conclusion
This project suggested that ADHD-related outcomes may be better understood as emerging from multiple interacting factors rather than one dominant measure. More broadly, it demonstrated the importance of evaluating how different measurement approaches contribute to understanding complex human behavior.

Transferable Skills
Research design · Measurement selection · Experimental design · Quantitative analysis · Explainability · Insight generation · Translating findings into decisions