Trial-count matching reduced the go/no-go criterion contrast with half-count correction in 158 ds005270 adults

We tested whether age was related to the criterion difference between frequent-go and rare-go blocks. We used 158 ds005270 adults, aged 20 to 86 years. A criterion is a score for the tendency to respond. The main adjusted age slope was 0.0033 per decade, with a 95% bootstrap interval of -0.0204 to 0.0261. This result did not establish an age effect or an absence of one. The mean block contrast was -0.5838 with half-count correction. It was -0.2657 after exact trial-count matching. Both intervals excluded zero. A zero-error example gave a contrast of -0.4047 from unequal counts alone. Thus, unequal counts can affect the size of this score contrast. Count matching changed the score summary, not the task records. The package includes source bytes, provenance, an exact matching method, code, result tables, and bootstrap draws. An independent calculation and a fresh-download replay agreed with the results. The study used accuracy labels to infer responses under a binary task rule. A recorded-response check gave a similar age estimate. The study did not test a brain mechanism or a causal effect of age.

go/no-go task · response criterion · signal detection · trial-count matching · cognitive control · adult lifespan