No abstract

T HIS investigation deals with the effect of individual differences in strength of achievement motive on risk taking behavior. It focuses upon (a) goal setting as inferred from choices individuals make among tasks that differ in difficulty (see Lewin, Dembo, Festinger, & Scars, 1944), and (&) the preferences of individuals in an imaginary gambling situation for options that are equated for expected monetary value but differ in the stated probability of winning (see Edwards, 1953).The senior author (1957, 1958b) has presented a theoretical model that extends some ideas developed earlier by Escalona and Festinger in the resultant valence theory of level of aspiration (see Lewin, et al., 1944). It accounts for the association between achievement motivation and entrepreneurial activity involving risk taking highlighted in the writings of McClelland (1955McClelland ( , 1958. The theoretical model was initially formulated to explain the results of two experiments dealing with different aspects of risk taking behavior. Atkinson (1958a) found that the level of performance of female college students was significantly higher when the probability of winning a small monetary prize by getting a high score was 1/2 than when the expectancy of winning was either very high (3/4) or very tow (1/20). McClelland (1958) found that children in kindergarten and second grade who were judged to be highly motivated to achieve on a graphic expression measure of n Achievement preferred an intermediate degree of risk (or difficulty) more frequently in several different competitive games than children presumed to be low in n Achievement. The latter group more often than the high n Achievement group favored playing it very safe or being very speculative,

No abstract

1"Spontaneous allernalion" is a descriptive term which has been applied lo the often observed tendency of rats to behave in an orderly rather than a random fashion when presented with a Iwo-choice alternative more than once in close succession. Empirically, if an animal turns left in a T maze on its first trial, and if it is immediately returned to the starling point, the probability is quite high that it will turn right on the second trial.This systematic tendency to alternate choices has assumed considerable theoretical importance. Solomon (7) related it to Hull's ''reactive inhibition" (I K ). Montgomery (4, 5, 6) found the phenomenon related to the stimulus present and attempted to explain it as a manifestation of exploratory behavior. Glanzer (1, 2) independently verified Montgomery's findings and explained the behavior in terms of ''stimulus satiation."The present, experiment stems from the logic and methods used by Montgomery and Glanzer. Both experimenters performed, among others, an experiment based on this kind of reasoning. Suppose we have a T maze with one arm painted white and one painted black. If an animal is placed in such a maze and turns left to the white stimulus, and on a second trial alternates and thus turns right to the black stimulus, it has alternated both the response (L, R) and the stimulus (w, b). It can be seen that the stimulus and the response in such a situation are confounded.If one asks the question, "Is the animal alternating with respect to the stimulus or to the response?" it is necessary to arrange a situation in which the animal can do either one but not both.Both Montgomery and Glanzer achieved this objective by using a +-shapecl maze. This was accomplished by adding an additional starting stem apposed to the original one. Now if an animal turns left to the while

Krcchevsky (6, 1) has shown lhat under certain conditions normal rats tend to select the more variable of two paths to similar goals. In experiments by Hebb and Mahut (5) and Havelka (4) some rats prefer solving a problem in order to reach food to reaching food by a simple, direct route. Butler (2), Montgomery (9), and Myers and Miller (10) have demonstrated that animals will learn to perform acts instrumental in providing them with interesting environments.Experiments of the above sort point to the importance of what may be called "stimulus complexity" in directing behavior. In the present-experiments rats are given access to two contiguous circular paths. The walls of one path arc lined with a simple pattern; the pattern in the adjacent path is relatively more complex. The ,5' s can move freely within and between paths. Our theory enables us to predict-the direction of any change in path preference.Informally stated, the theory postulates lhat both individuals and stimuli have on any attribute a measure of "complexity." Given free choice, an individual will respond to, or pay attention to, a stimulus in inverse proportion to the difference between its complexity measure and his complexity measure plus some increment. Stimuli with complexity measures greater by this increment, than the individual's measure arc called "'pacers." As a result of commerce with a pacer, the individual's measure increases. Thus, as the individual's complexity measure increases, he will attend to stimuli of greater complexity.In the present experiments neither the measures of the stimuli nor of the Ss were known beforehand. However, by generalizing on some data by Atlneave (1), we have as-' The theory, apparatus, and design for llicsc experiments arc a product of the joint efforts of Dember and Karl. Their names are listed alphabetically. Earl ran Exp. 1. Experiment 2 was carried out by Mr. Paradise.