Psychological Reports

A multiple-alternative item type having the basic characteristics of true-false items but differing from the conventional true-false type item in several respects was described. Evidence was presented supporting the contention that the item type is suited to the task of measurement in the areas of diagnosis and hypothesis formulation. Further, following the construction of a key with the aid of item analysis, the correlation between the test and a performance test criterion measure was .60 for a hold-out group used for purposes of cross-validation.

Taylor has suggested that where a negative relationship exists between intelligence and scores on her anxiety questionnaire, it is due to the ability of the more intelligent Ss to “outguess” the anxiety scale and present themselves in a favorable light. 84 naval cadets took the Taylor scale under instructions to give the “best” answer. Under these conditions a statistically significant correlation of —.29 between the Taylor scores and ACE scores was obtained. Where 319 Ss were given the scale under “normal” circumstances, the correlation was only .06. These results are interpreted as supporting Taylor's position.

One object was presented alone for two trials and then paired with a second object for six trials. Next, the second object was paired with a third object for six trials, and then sometimes the third object was paired with a fourth object for six trials. Four such sequences were given to seven test-wise rhesus monkeys on each of five days. Retention tests involving pairing of the dicrimination objects with new objects occurred on three days. When a chosen correct or incorrect object was paired with a new object, there was high positive transfer. When an unchosen negative object was paired with a new object, there was a tendency to choose the former. Although there was somewhat better retention of positive than negative objects, this may be attributed to unequal selection in the learning trials.

231 cadets who had failed or were withdrawing from the Naval Flight Training Program were classified into “anxiety” or “no-anxiety” groups. The “anxiety” group had a significantly higher score on the Heineman forced-choice form of the Taylor Manifest Anxiety Scale than did the “no-anxiety” group.

55 middle class American males and 57 females were tested for number of shifts made in level-of-aspiration experiments with a stylus maze and a series of digit-symbol substitutions. The difference between the sex groups did not reach statistical significance in the case of either task.

Methodological difficulties are encountered in the construction of a frequently used problem-solving rigidity scale. Specifically, lack of reliability data, sample loss due to arithmetical error and indirect rather than direct solution to control problems, and a lack of or an insufficient number of control problems (without use of control groups) are prevalent with use of such scales. Part I of this paper reports an attempt to resolve these difficulties. A split-half reliability coefficient, significant at the .01 level of confidence, was obtained for control and critical problems. 4 control and 5 critical problems were suggested as optimum. No differences were found between direct and indirect solvers on the basis of age, sex, or major course of study. A model rigidity scale applicable to groups was presented. In Part II, this scale was administered to 97 college students. A sample loss of 9% and indirect scores to critical problems sufficient to establish a rigidity scale were found. The females responded to Einstellung effect significantly differently from chance but the males were not differentiated with reference to this effect.

89 Air Force tactical instructors were rated for performance on team problems, and by basic trainees under their supervision, by supervisors, by other tactical instructors, and by themselves for overall teaching effectiveness. Intercorrelations of ratings indicated that: (a) none of the ratings of teaching effectiveness was significantly correlated with performance test ratings; (b) the teaching effectiveness ratings showed fairly high intercorrelations; and (c) supervisor, self, and peer ratings were not significantly correlated with student ratings.

This investigation studied the relationship between anxiety level and transfer of training in a paired-associate learning situation. On the basis of conventional anxiety theory, it was predicted that high-anxiety Ss would show greater positive and greater negative transfer than low-anxiety Ss. To test this prediction 84 undergraduates were selected from the extremes of the anxiety continuum as measured by the Taylor Scale of Manifest Anxiety. These Ss were divided into six groups of 14 each. The major high- and low-anxiety groups were divided into sub-groups which served in a positive transfer, negative transfer, and control condition. Each group, containing an equal number of men and women, learned two paired-associate nonsense syllable lists which differed in stimulus and response terms or both, in order to produce, theoretically, positive, negative, or no transfer. The results of the experiment yielded no statistically significant differences among the groups in terms of initial learning. In the transfer task the results conformed to the usual transfer of training paradigm. However, when the results were analyzed in terms of sex difference and anxiety level, high-anxiety male Ss behaved according to the original prediction while women behaved in an almost opposite manner. The low-anxiety female Ss performed similarly to the high-anxiety men. These results, therefore, did confirm the traditional conception of transfer, while the differences in performance between high-anxiety men and high-anxiety women indicated possible shortcomings inherent either in the methodology of the experiment, or in the theoretical assumptions underlying anxiety as a motivational variable. Several possibilities were considered.

There is general agreement that intellectually gifted children should be encouraged to develop their potentialities. However, a great deal of evidence indicates that currently used intelligence tests are not equally efficient in detecting high levels of ability in the various social classes. It is proposed that the highest in each class should be selected for special training.

Ss were run in a “two-choice” experiment in which the stimulus series consisted of equally-likely alternatives arranged in a random sequence. Half the Ss were run under non-involving instructions and half under involving instructions. Half the Ss in each of these groups were told that the sequence of alternatives was random and half were told that the sequence had a definite structure. Telling an S that the problem was soluble significantly reduced his stereotypy and increased his tendency to alternate an incorrect response. Involvement had no effect on an S who believed the problem to be insoluble, but had a significant effect on Ss who believed the problem to be soluble. Ss in the involved-structured group grossly underevaluated their performance and had the highest level of aspiration relative to Ss in the other groups.

Experimental studies of ego involvement are reviewed with respect to definitions of the concept, procedures and designs used, and effects of these on retention, perception, motor responses, problem solving, preferences, evaluative judgments, set-shifting tasks, level of aspiration, and test responses. Generalizing from their review, the writers propose the hypothesis that varying reactions to ego involving conditions reflect, at least in part, a continuum of motivation. The relationship between degree of ego involvement and efficacy of response appears to be a curvilinear one, inhibitive or disruptive at the extremes and optimally facilitative near the middle. There also appears to be a common personality dimension which is responsive to ego-involving conditions.

Test-retest reliability coefficients were computed for personality data using Q-sorts by 10 Ss. Coefficients of correlation were high (between .93 and .97), suggesting that the technique yields reliable measures.

It is hypothesized that the cancer patient has, more often than chance alone would permit, lost a major cathexis some time before the development of the neoplasm. Certain predictions as to relative cancer mortality rates in different groups are made from this hypothesis. An examination of the statistical literature in the cancer field bears out these predictions and offers additional evidence. The possible validity of this hypothesis suggests that further investigation would be highly desirable.

A comparison of abilities and personality attributes of nursing students on the basis of pre-entrance testing data is made for a successful graduating group and an unsuccessful drop-out group. 89 students admitted to the college of nursing in 1949, 1950, and 1951 were studied. 49 graduated; 40 dropped out. Of these 40, 18 were dismissed for “scholastic failure”; 6 dropped out reportedly because they “disliked nursing.” For the remainder, marriage, personal reasons, college attendance, and illness are the reasons listed in the college records. 51% of the 40 drop-out students were identified as “poor risks” in the recommendation of the psychologist, while only 4% of the 49 who graduated were similarly identified. Since academic success at college level is essential for remaining in training, it is not surprising that L or the T percentile of the ACE is the most effective single index in predicting success or failure. The CRs for the mean percentile difference are 3.07 and 2.96 for the L and T factors respectively, significant at the .01 level of confidence. Although group differences on the Kuder Preference Record are not statistically significant, the CRs for three scales, Literary, Social Service, and Clerical, approached significance (1.97, 1.87, and 1.76). Mean scores for all scales on the MMPI were almost identical for the two groups. Careful screening through various means for identification of gross personality defects before admission is suggested as a reason for negative findings.

The Morgan Test of Logical Reasoning, which is entirely verbal in form, was given to 133 males and 35 females. All Ss were in their 20s and applicants for government employment. All had the BA or BS degree but no other, and none had had any training in formal logic. No significant difference was found between the mean scores of males and females on this test, nor between the distribution and variation of scores.

McCreary and Bendig found a reduction in the mean Taylor score on retest. They hypothesized that fatigue reduces the anxiety scores and, therefore, that the scale imbedded in the full length MMPI would yield lower scores than when given separately. To test this hypothesis three groups of naval cadets (N=106, 134, 113) were given both the scale alone and the scale imbedded in the MMPI. The reduction in anxiety score on retest was confirmed but this reduction occurred even after an 18-hr. rest period. Further, there were no significant differences between forms. These results were interpreted as indicating that fatigue does not affect the anxiety scores in a consistent way, and that the two forms of the Taylor scale are comparable.

In the present study, albino rats were given discrimination training in a modified Skinner Box. During this training several schedules of S^D—S^R pairings were presented; differential effects of this training were measured during extinction, which also involved a further scheduling of the R—S^r pairings. The results of the study showed that the animals were discriminating, and that S^D functioned as S^r during extinction. However, there was no consistent evidence that a moderate degree of intermittence in reinforcement during either discrimination or extinction phases had favorable effects on secondary reinforcing properties. There remains the possibility that there are real but small effects which could be detected only by much more extensive experimentation.

3 rhesus monkeys were given various concentrations of sucrose for lever pressing on a variable interval schedule of reinforcement. 7 sucrose concentrations were studied at 2 levels of food deprivation. The response rates accelerated rapidly with increasing concentrations, and then declined after reaching a maximum, generally between 15 and 30% sucrose concentration. The decline was attributed to a satiation effect. The higher level of food deprivation tended to increase the response rate at all but the extreme high and low concentrations.

Rats (N=80) were taught a conditioned avoidance response of the Miller type and were subsequently extinguished by lack of reinforcement. Following habit extinction selected animals were subjected to carbon disulfide inhalation, metrazol injection, or exposure to high frequency sound, and replaced in the modified Miller box. It was found that all agents used were effective in reinstating the extinguished response, but that the form of the recovered habit differed depending on the agent used to elicit recovery. It was suggested that the methodology employed in the study might prove useful in throwing light on bio-mechanisms involved in the acquisition of different types of habit.

The purpose of this research was to determine the relationship between reactions of cadets during simulated high altitude in a decompression chamber and the later development of anxiety toward flying as reported in terminal interviews. The results for 1540 cadets indicated that significantly more of those who withdrew because of anxiety toward flying had anxiety reactions in the decompression chamber than of those who completed the Naval Air Training Program.

The purpose of this study, modified from an earlier experiment by Tolman, Ritchie, and Kalish, was to test Hull's theory of response learning and Tolman's theory of place learning. 10 female albino rats were trained to run an elevated T-maze. Preliminary training consisted of 6 trials per day for 4 days, emphasizing the alternation of left and right turns on successive trials. The experiment proper commenced on the fifth day and continued until each rat had reached the criterion of 11 out of 12 successive errorless trials of turning to the right, with food serving as the reinforcement. After reaching the criterion, each rat was given a test trial which consisted of a run from a new starting point, located exactly 180° from the original starting point, to either the previously rewarded goal box now located to the animal's left or to the previously unrewarded goal box now located to the animal's right. It was found that on the test trials 8 of the rats ran to the previously rewarded place and only 2 made the previously rewarded response of turning to the right. This proportion is different from chance at the .055 level. It was concluded that in situations in which there are marked intra-maze cues and marked extra-maze cues there is some evidence that place learning is probably simpler than response learning. The evidence points consistently, however, to the interpretation that the animals learned something more than merely the place where the food was to be found.

Intuitive processes have their place in the clinical judgments of the therapist. They are also necessary in the scientific analysis of personality. But in diagnostic classification for the purpose of deciding the treatments to be given to patients, intuitive processes are ineffective and should be replaced by objective procedures whose efficiency has been established empirically.

This is a critical examination of the evidence and issues involved in actuarial and clinical methods of psychodiagnostic evaluation. After demonstrating that the evidence to date is inadequate and largely irrelevant to the problem, consideration is given to the basic issues involved. The basic difficulty in arriving at adequate criterion variables is stressed. The answer to the general question of which methods are superior under specified conditions is seen to lie in an area of interest to all psychology: the psychology of prediction. In the field of psychodiagnosis, two types of predictions are specified, prediction of carefully specified behavioral events and prediction of carefully specified behavioral processes. In the first, clinical judgment is required in assigning weights to prediction variables according to individual patterns. In the latter, contingency predictions about emergent behavior are required. While improvement of both actuarial and clinical methods is undoubtedly necessary, it is maintained that not actuarial methods but “human judgement, let us hope based on sufficient evidence, will always have the last word.”

Three groups of animals, a 15 ECS group (N of 9), a 15 ether-ECS group (N of 11), and a control group (N of 9) were exposed to a series of “timidity box” tests 27, 75, 110, and 163 days after last ECS. The “timidity box” was found to provide a sensitive measure of relatively long lasting emotional changes as a function of the series of 15 ECS. The ECS group proved to be more “timid” than the control group. The ether-ECS animals in this experiment exhibited minimal convulsions and evidenced changes in behavior similar to those seen by the ECS group. The degree of significance of differences tended to decrease as the post-ECS time increased, i. e., the “timid” behavior induced by ECS decreased as a function of time.

In a study of the correlations between the factors measured by MMPI and STDCR, 52 college freshmen and sophomores completed both tests. Correlations between the factors in the two tests revealed that: (a) a large number of MMPI factors seem to test traits not tested by STDCR; and (b) factors S and R of STDCR seem to measure in trait areas similar to factors D, Ma, and Si of MMPI; and factor D of STDCR overlaps to some degree with factors D and Si of MMPI.

This experiment was designed to evaluate the effects on retention of introducing non-optimal muscular tension in the legs during the course of learning 40 words. The tension was introduced during the recitation of the second 10 words. There was a tension group and a no-tension group, each with 30 Ss, in which there was no inter-trial “rest.” There were two corresponding groups in which there was a 1-min. interval between word lists. Analysis of the immediate recall data supports the conclusion that non-optimal tension during learning is directly related to recall scores. This suggests a general hypothesis that tension during learning may bring about poorer recall scores but will not improve recall.

Three groups of 16 rats were exposed to no shock, 200 microamperes, or 400 microamperes, respectively, for 50 min. The next day, half of each group were trained to press a bar to turn off a 200 μa shock and half were trained to turn off a 400 μa shock. The higher the current during the first (exposure) session, the lower and more variable was the rate of pressing during the second (training) session. The higher the current during training, the higher and less variable was the rate of pressing. The persistence of the effects of prior exposure after 24 hr. suggests that some kind of learned behavior is involved.

Recent developments in the study of stimulus generalization by means of operant conditioning methods require a re-examination of the problem of stimulus generalization and a review of the role of this concept in description and analysis of behavior. No final view of the outlines of learning theory can be achieved until more complete and fundamental information about generalization has been obtained. It has been established that in the pigeon orderly and reproducible bidirectional gradients can be obtained for individual Ss along the light-wavelength continuum. A dilemma is posed by the finding that these gradients are orderly on a physical scale; but not on an equal-discriminability scale. The form of the generalization function raises questions as the interpretation and meaning of the just-noticeable-difference concept. The problem of multidimensional stimulus generalization is discussed. The understanding of stimulus relationships in the generalization phenomenon is proposed as an issue of greater significance and difficulty than that of the control of generalization by such conventionally more prominent variables as drive and reinforcement.

The Pensacola Z Scale was adopted as a measure of authoritarianism and administered to 761 men undergoing training for submarine duty and 407 naval retrainees confined for “delinquent” behavior. Significantly higher levels of authoritarianism were found among confined Ss. However, the difference was not significant when scholastic aptitude (GCT) was held constant. The results suggest that the frequently observed relationship between authoritarianism and confinement for “delinquent” behavior may be mediated by scholastic aptitude.

In a previous experiment (2) it was shown that rats will choose that arm of a maze the brightness of which has been changed between an exposure trial and a free-choice trial. The present experiment tests the hypothesis, derivable in part from Glanzer's satiation theory, that if both arms are changed, rats will choose the arm representing the greater change. Ss were 28 male albino rats satiated for food and water. Four stimulus conditions were used, a condition being composed of two trials: on Trial 1 rats were exposed to 2 arms of a Y-maze for 3 min., but were prevented from entering either by means of glass partitions; for Trial 2 the brightnesses of both arms were changed and the partitions removed. In Condition I, the Trial 1 configuration was left-black, right-grey; on Trial 2 both arms were white. In Condition II, on Trial 1 the configuration was left-grey, right-black, and again on Trial 2 both arms were white. Conditions III and IV were analogous to I and II with white and grey on Trial 1, and both arms black on Trial 2. Of the 28 Ss, 20 made the predicted response, significant by the sign test (one tail) at better than the .02 level. Errors were distributed equally between Conditions I and II combined and III and IV combined, indicating that the grey arm was discriminable from both the white and black arms.

Two investigations are described in which series of verbal classification problems were used to study two hypotheses concerning the nature of set or direction in problem solving. The evidence is considered to support the hypotheses, which state that direction is a function of priority of stimulus activation of response systems under certain conditions and of activation of strong and pervasive response systems previously acquired by Ss under other conditions.

A technique has been described whereby preferences of animals for a large number of stimuli can be studied simultaneously, and the responses automatically recorded. Groups of Ss are presented with a number of colored stimuli on the sides of an enclosure. Pecks are recorded on electrical counters. Data are presented showing that differential responses to colored objects do occur in chicks (N=200) and ducklings (N=100) that have had very limited visual experience and no opportunity for directed learning. It must therefore be supposed that these responses are innately organized. For the chicks there is a bimodal preference to color, with one peak occurring in the orange region of the spectrum and a second peak in the blue region. The ducklings, on the other hand, have a narrower range of preference with a single sharp peak within the green and yellowish-green region. The color preference curves of chicks and ducklings are obviously quite dissimilar. The dissimilarity suggests some functional determinant, but the nature of the determinant is as yet unknown.

Using 30 Ss and 10 words for which the total possible number of decompositions were obtained, factors most associated with success in finding valid decompositions were maximum number of available decompositions, length of the stimulus word, and S's IQ. The capacity to make visual and phonetic associations among decompositions also appeared to play a substantial but secondary role.

Several studies are described in which relevant patterns of verbal associations, set up by learning in the first stage of the experiment, are shown to be associated with the frequencies of certain types of solution in the Maier two string and hat rack problems. It is also shown, in the case of a simple verbal problem, that problem solution requiring choice of members of a chain of free associations is affected by prior reinforcement of one member of the chain. These investigations are interpreted as giving support to and indicating the fruitfulness of a conception of set or direction in problem solving as consisting of complex response systems or habit families.