On Test Variance and the Dimensions of the Measurement Situation

Pitfalls in Assessing Test Speededness

“…Designated tau, this estimate is based on comparing the performance of the same subjects in timed and untimed administrations of parallel tests. A test is then seen as speeded "to the extent that true standard scores in the group would be changed if more time were allowed" (Morrison, 1960). As can be seen from the equation below, tau is the square of a corrected-for-attenuation estimate of the correlation between timed and untimed administrations: ,n,r,4,B, tau (6) rA,n, rA, n, where A and B are parallel tests and t and p identify time limit and power conditions respectively.…”

“…On the other hand, in a pure power test, rate of work has no necessary impact on scores.In the early history of group tests, meager evidence was gathered to support an assumption that speed and power tests served as interchangeable measures of the same construct. Such a notion implied to test constructors that it made no difference whether abilities were measured "against a scale of difficulty, a scale of speed, or some combination of the two" (Morrison, 1960).Subsequent psychometric nonchalance regarding effects of time limits on scores has been seen as resulting "partly through confusion in terms and partly through misinterpretation of... experimental evidence" (Davidson and Carroll, 1945). The explanation of this misconception continued:The most frequent error [has been] that of identifying time limit scores as speed scores and then proceeding to cite the correlation between time limit scores and scores obtained [on the same test] in unlimited time.…”

“…In the early history of group tests, meager evidence was gathered to support an assumption that speed and power tests served as interchangeable measures of the same construct. Such a notion implied to test constructors that it made no difference whether abilities were measured "against a scale of difficulty, a scale of speed, or some combination of the two" (Morrison, 1960).…”

Validity Issues in Test Speededness

“…The advantage of this approach is that it is free of assumptions about “the nature of the new factors introduced with the time limit” and “allows for the possibility that there may be several types of speededness, depending on other characteristics of the test and upon the subjects” (Morrison, 1960, p. 236). Also, multiple test administrations under timed and untimed conditions provide realistic information on how extra time could change the score distributions.…”

Testing and the Use of Test Results

“…Wesman (1960) agreed that speed is sometimes part of the criterion variable that a test is designed to measure; only in such circumstances did he favor using speeded tests. Morrison (1960) discussed the effects of speed conditions on the performance of examinees and provided an excellent bibliography on this topic. Curtis and Kropp (1961) compared the scores of examinees on the widely used School and College Ability Test, Form 3A, when the test was given normally and when it was given by projecting the items on a simulated television screen.…”