Lapse resistance in the verbal letter reporting task

Abstract

Lapses, or misreporting errors, can affect accuracy of threshold measurements. Assumptions about lapse rate, especially in untrained observers, have consequently guided the design of at least one clinical psychophysical test. Lapse rate was assessed using a verbal letter identification paradigm like that used in visual acuity and letter contrast sensitivity testing. Subjects occasionally made slip-of-the tongue errors but spontaneously corrected them. Lapse rate (excluding such errors) was 0-3 errors per 1,536 (average rate of 0.0005). In this common clinical paradigm, in which observers set their reporting pace, and where opportunity to amend responses is available, lapse rate is negligible.

Fig. 1

The Weibull functional form of the psychometric function, showing λ, the lapse rate, as the departure from perfect performance, when stimulus intensity is substantially above threshold. In this example, the probability of a correct guess and the lapse rate are both 0.1.

Fig. 2

Sample (first page) of the sequence of letters used in the experiment. Actual size of the page was 8.5 × 11 in.