Eye-tracking was used to compare problem-solving strategies (Constructive Matching [CM] vs. Answer Elimination [AE]) during Raven's Standard Progressive Matrices (RSPM) in 38 adolescents/young adults with ADHD vs. 41 typically developing (TD) controls (mean age ~19 years), with CFT-assessed intelligence as a covariate across three difficulty levels.
ADHD participants showed broadly normal CM gaze patterns and similar adaptation to increasing task difficulty as TD controls, despite slightly lower RSPM accuracy (61% vs. 68%; trend p=.052, d=0.37) and significantly lower CFT IQ (103 vs. 118, p<.001). Critically, higher CFT intelligence predicted greater CM use in ADHD (ηp²=0.07–0.11) but not in TD controls — an effect limited to the easy-to-medium difficulty transition and absent from medium-to-hard items.
Small 9-item RSPM short form reduces reliability vs. full version; ADHD subtypes were not documented; working memory was not directly measured, leaving the proposed mechanism inferential.
Children and young adults with ADHD may use normal visual reasoning strategies despite lower accuracy, suggesting strategy coaching alone may be insufficient — the quality of cognitive execution matters. Higher intelligence appears to buffer ADHD-related processing inefficiency only up to moderate task complexity, which clinicians should keep in mind when interpreting high-IQ ADHD presentations.
