Effects of truncation on reaction time analysis.

Abstract

Many reaction time (RT) researchers truncate their data sets, excluding as spurious all RTs falling outside a prespecified range. Such truncation can introduce bias because extreme but valid RTs may be excluded. This article examines biasing effects of truncation under various assumptions about the underlying distributions of valid and spurious RTs. For the mean, median, standard deviation, and skewness of RT, truncation bias is larger than some often-studied experimental effects. Truncation can also seriously distort linear relations between RT and an independent variable, additive RT patterns in factorial designs, and hazard functions, but it has little effect on statistical power. The authors report a promising maximum likelihood procedure for estimating properties of an untruncated distribution from a truncated sample and present in an appendix a set of procedures to control for truncation biases when testing hypotheses.

Comparison of Absolute Thresholds Derived from an Adaptive Forced-Choice Procedure and from Reaction Probabilities and Reaction Times in a Simple Reaction Time Paradigm

“…However, because our goal requires measures of detection and of reaction thresholds that are as comparable as possible, we used the 0.561 quantile rather than the median. The mean RT may also be a good measure of the location of an RT distribution, although it has the disadvantage of being very sensitive to Boutliers,p articularly to late reactions (see Ulrich and Miller 1994). However, a mean cannot be computed from all trials when there are misses.…”

The impact of storage on processing: How is information maintained in working memory?

“…"The median, however, is inappropriate for this purpose because it is not, in general, additive" (see Sternberg, 1969, p. 286). Our main hypothesis is concerned with the examination of linear relations between RT and memory load for which Ulrich and Miller (1994) showed that "using truncated RTs can seriously distort linear relations between RT and an independent variable" (p. 34). This document is copyrighted by the American Psychological Association or one of its allied publishers.…”

Enhancing the learning of new words using an errorless learning procedure: Evidence from typical adults

“…prior to analysis erroneous responses were excluded and an inverse transformation was used to reduce the effect of remaining outliers (see Ratcliff, 1993;Ulrich & Miller, 1994). For the Stop Signal Reaction Time task, SSRTs were estimated using the mean method (mean of inhibition function subtracted from the mean of RT distribution) as outlined by Band et al (2003) and Verbruggen and Logan (2009).…”