Academic journal article Perception and Psychophysics

On Estimating the Difference Limen in Duration Discrimination Tasks: A Comparison of the 2AFC and the Reminder Task

Academic journal article Perception and Psychophysics

On Estimating the Difference Limen in Duration Discrimination Tasks: A Comparison of the 2AFC and the Reminder Task

Article excerpt

This article assesses whether the two-alternative forced-choice (2AFC) and the reminder tasks (i.e., method of constant stimuli) yield identical estimates of the difference limen (DL). In a series of six experiments, participants discriminated between two duration stimuli. Experiments 1-5 employed auditory stimuli, and Experiment 6 employed visual stimuli. Experiments 1 and 2 combined each of the two tasks with an adaptive and a nonadaptive procedure for threshold estimation. Experiment 3 varied the distribution of the comparison levels, whereas Experiment 4 employed random interstimulus intervals. Experiments 5 and 6 examined the influence of the presentation order of the standard and comparison stimuli. Results indicate that both the adaptive and the nonadaptive procedures yield virtually identical DL estimates; yet, the 2AFC task produces consistently larger DLs than does the reminder task. In addition, DL increases when the standard occurs in the second rather than in the first stimulus position. In order to account for these results, we assume that participants use an internal standard instead of the actually presented standard as a reference for their judgment.

The difference Urnen (DL)-also termed just noticeable difference (JND)-is a traditional and important concept in psychophysical work. It provides a means for quantifying the sensitivity of a sensory system. Although signal detection theory also provides powerful tools to assess the sensitivity of a sensory system (see Macmillan & Creelman, 2005; Wickens, 2002), this theory has not replaced the concept of DL. Indeed, a 2001 special issue of Perception & Psychophysics edited by Klein and Macmillan covered new approaches to estimating DL.

The DL measures the amount of change in a stimulus that is required to produce a JND in the sensation of this stimulus (see, e.g., Gescheider, 1997; Guilford, 1954). For example, on each trial of a duration discrimination task, two tones are successively presented to a participant (e.g., Getty, 1975). One tone is die standard, and its duration is kept constant-say, at 500 msec-throughout the experiment. The second tone is the comparison, and the experimenter varies its duration from trial to trial. The listener may experience a difference between the two durations only when the comparison is at least 554 msec. In this case, the DL would be 54 msec-that is, equal to the duration difference between the comparison and the standard. The smaller the DL, the higher the discrimination performance.

Several psychophysical tasks are available for measuring the DL (for a review, see Gescheider, 1997). Two common tasks are the reminder task and the two-alternative forced-choice (2AFC) task (Macmillan & Creelman, 2005).1 In either instance, the participant is presented with a standard and a comparison stimulus on each trial. Moreover, several comparison durations are employed, and in each trial, one of the comparisons is randomly chosen from this set of durations. In the reminder task, the standard is always presented first, followed by the comparison. On each trial, the magnitude of the comparison can be smaller, equal to, or larger than the magnitude of the standard (e.g., the duration of the comparison may be shorter, equal to, or longer than the duration of the standard). At the end of each trial, the participant is asked to report whether the comparison was longer or shorter than the standard by responding "longer" or "shorter," respectively. The results generated by this reminder task can be displayed in the form of a psychometric function. This function plots the proportion of "longer" responses against comparison duration. Typically, this function resembles the shape of an ogive curve, is O at small comparison values, and approaches 1 for large values. A psychometric function from a certain parametric family (e.g., logistic function) is then fitted to the observed data points. Finally, the DL is estimated as being half the interquartile range of this fitted function-that is, DL = (x^sub . …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.