Correct-go response variability is associated with trial- and sequence-level accuracy in an independent task-switching task

Reaction-time variability may carry information about how accurately a person performs beyond the task in which it is measured. We examined this possibility by relating correct-Go reaction-time (RT) variability in a Go/No-go task to accuracy in an independent task-switching paradigm. One hundred sixty-one undergraduates completed both tasks. Twelve RT-variability indices were summarized by a first principal component (PC1), with equal-weight and single-index scores examined as alternatives. PC1 explained 84.6% of indicator variance and showed high within-session reliability (Spearman-Brown = 0.952). Greater variability was associated with lower condition-balanced trial-level accuracy (Spearman ρ = -0.520) and all-responses-correct sequence-level accuracy (ρ = -0.525). These accuracy scores were strongly related (ρ = 0.981); variability explained an additional 0.95% of sequence-score variance after trial accuracy and demographics were considered (cross-validated ΔQ² = 0.010). Negative associations were observed across switch and repetition conditions and across error locations, with no reliable variability-by-condition interactions. An equal-weight score was almost identical to PC1 (r = .99996) and performed comparably in internal cross-validation. Task-switching RT variability also correlated with correct-Go variability, although their adjusted association weakened after mean speed and accuracy were included. Correct-Go variability thus captured a reproducible within-sample association with accuracy across these two paradigms. Its psychological source and suitability for practical assessment require further validation.