| |
Abstract:
Abstract: Under normal conditions, recognition performance
shows a speed-accuracy trade-off: fast responses are less accurate
than slow responses. Accelerating recognition responses tend to
reduce the number of `old' responses to old items (hits) and to
increase the number of `old' responses to new items (false alarms).
We investigated the effect of accelerating recognition responses in
the Deese-Roediger-McDermott (DRM) paradigm. Subjects studied a
series of lists composed of words semantically related to a
nonpresented critical lure and then performed a recognition test.
In Experiment 1, subjects were asked to respond either as
accurately as possible or as fast as possible. Recognition
responses to targets and unrelated lures showed the standard
speed-accuracy trade-off. In contrast, 'old' responses to critical
lures decreased when responses were accelerated. This result was
replicated in Experiment 2 using a deadline procedure. When allowed
response time (the amount of time between item presentation and the
deadline) was decreased from 1200 to 300 ms, accuracy decreased for
targets and unrelated lures, but increased for critical lures:
subjects were more likely to reject critical lures as recognition
speed increased. We interpret these results as suggesting that the
familiarity feeling that leads to false recognition of critical
lures takes time to accumulate, possibly reflecting the retrieval
of associated words from the studied list.
|
