Content area
Full Text
Throughput is a corrected response rate measure giving the number of successes per unit of discretionary time. It is a simple but general index applicable to psychomotor, behavioral, and cognitive tasks in which response times are measured. This measure has several attractive features: (1) It allows comparisons to be made across various tasks in which speed and accuracy are meaningful and measurable, independently of temporal differences in hardware, software, and procedures; (2) under conditions in which both speed and accuracy decline (or improve), throughput will be a more sensitive index of performance than either alone will be; and (3) in those tasks in which the speed-accuracy trade-off phenomenon operates, throughput will tend to be relatively less variable than either component alone will be. The measure allows both behavioral and information-processing interpretations of data and may be useful as a simple composite index, a measure of effectiveness or of cognitive consistency in studies investigating performance degradation or enhancement.
(ProQuest-CSA LLC: ... denotes formulae omitted.)
Under conditions common to many psychological studies, the behavioral measures that are most sensitive to experimental manipulation are also the ones most sensitive to extraneous variables. If many of these extraneous variables are uncontrolled or uncontrollable, the data generated with such measures will be noisy. Conversely, measures that are stable across conditions or time tend to be relatively insensitive to random variables, but also to the independent variables of interest. Sensitivity and variability are usually inseparable.
A composite response measure (herein, called throughput) was originally devised for reasons of convenience specific to a particular set of experiments in our laboratory in which a computerized psychological test battery was used (Thorne, Genser, Sing, & Hegge, 1985). This measure subsequently proved to be more orderly than was anticipated, revealing effects that were larger and less variable than those shown by more conventional measures. Some of the reasons for this superiority have now been identified, and they suggest that the measure may be useful in a number of frequently encountered experimental situations.
Definition and Interpretation No. 1
Numerically, throughput is equal to the number of correct responses on a task, divided by the cumulative reaction times (RTs; both correct and incorrect). Conceptually, throughput may be viewed as a corrected rate measure giving the...