Intercorrelations of measures of learning in the albino rat.

Abstract

A brief statement of the community of function between certain learning situations, as applied to rats in the Stanford psychology laboratory, is given in a recent paper by Commins, McNemar, and Stone (2). One of the chief contributions of that paper is the fact that a multiple-T alley maze and two diverse patterns of the Miles types of elevated maze (13) repeatedly yielded intercorrelations of approximately .6, whereas each of these instruments yielded only zero coefficients when correlated with a multiple light discrimination apparatus ( 19). The latter fact, confirming in part some unpublished data collected by Stone and Nyswander in 1927, in connection with their study of maze reliability (18), and likewise somewhat similar findings independently obtained by Williams (28), would seem to be a discovery of capital importance for students of individual differences in animals because it clearly establishes the fact that two fairly reliable instruments may have no community of function.We look upon this discovery as one of capital importance because now-a-days many problems in animal psychology are attacked by laboratorymethods masquerading under the same name, but differing from each other in many essential details. So long as this continues to be the case, and so long as we lack accurate information as to the community of function of these diverse techniques of experimentation, just so long shall we continue to grope in the dark while seeking plausible explanations of the divergent results obtained on fundamental problems. Naturally, with the mere demonstration of community of function or absence of the same in different experimental settings we should not stop experimentation, for this demonstration is merely a first step 73

Correlation between VTE's on a maze and on a visual discrimination apparatus.

“…First, there is the fact that there was no correlation between errors on the two apparatuses (-.10). This type of result has also been found by most previous investigators, for example: Williams (20), Cummins, McNemar and Stone (5), Tomilin and Stone (14), Vaughan (17), and Anderson (1). It appears, in short, that maze learning and visual discrimination learning are two relatively independent functions.…”

Mapping Individual Variations in Learning Capacity

“…The factor analytic frameworks developed to describe individual differences in human intelligence have also been used to measure variation in the learning abilities of non-humans (Anastasi, Fuller, Scott, & Schmitt, 1955;Banerjee et al, 2009;Dunlap, 1933;Harrington, 1968Harrington, , 1988Herrmann, Hernandez-Lloreda, Call, Hare, & Tomasello, 2010;McCullock, 1935;Tomlin & Stone, 1934;Van Steenberg, 1939). Early studies with rats found little evidence of general intelligence comparable to that seen in human intelligence tests (Campbell, 1935;Commins, McNemar, & Stone, 1932;Livesey, 1970;Loevinger, 1938;Searle, 1949;Tomlin & Stone, 1934;Tryon, 1942;Wahlsten, 1978;Warren, 1961). For example, it was impossible to predict which rats would learn a discrimination task the fastest based on their performance in a maze task (Rajalakshi & Jeeves, 1968;Tolman & Ritchie, 1943;Tryon, 1942).…”

Interrelationships in Children's Learning

“…To the writers' knowledge there have been no studies of the interrelationship of performance on different types of learning tasks with children, although there have been frequent studies of this type with animal Ss (Commins, McNemar, & Stone, 1932;Dunlap, 1933;Geier, Levin, & Tolman, 1941;Tomilin & Stone, 1934;Thorndike, 1935;Vaughn, 1937;Anderson, 1938). The results of the studies with animals (rats and chickens) revealed moderately high correlations (typically around .60) between performance on similar types of learning tasks, such as two types of mazes or two types of discrimination problems, but very low correlations (typically around .10 or .20) between performance on different types of tasks, such as mazes and discrimination problems.…”