Rogers Elliott

Choosing and Leaving Science in Highly Selective Institutions.

This study sought to discover some of the causes of initial interset in and atrition from the natural sciences and engineering among the students (N=5320) who entered four highly selective institutions in 1988, with particular attention to possible special causes for the disproportionate attrition of women from science. Though a smaller proportion of women (35 percent) than men (49 percent) were initially interested in science, gender added little to the prediction of such initial choice when preadmission measures of developed abilities were taken into account in regression analysis. Of the group of 2,276 students initially interested in science, 40 percent did not finally concentrate in science, and smaller proportions of women (48 percent) than of men (66 percent) persisted. The most significant cognitive, factor predicting these losses was low grades earned in science courses taken during the first two years of study. With grades held equal, gender was not a significant predictor of persistence in engineering and biology; gender added strongly to grades, however, as a factor associated with unusually large losses of women from a category that included the physical sciences and mathematics. Responses to a questionnaire administered in the fall of 1991 showed that science majors regarded their instruction as too competitive, with too few opportunities to ask questions, taught by professors who were relatively unresponsive, not dedicated, and not motivating. Students who defected from science did so largely because of the attraction of other fields, but many shared the criticism of overcompetitiveness and inferior instruction, along with the view that the work was too difficult. Several items were about elements of classroom instruction and atmosphere thought to be especially difficult for women (i.e., the chilly climate), but except for perceived competitiveness, women did not rate their classroom experiences as being more unpleasant than did men.

Breaking the Cigarette Habit: Effects of a Technique Involving Threatened Loss of Money

Participants in a program to break the cigarette habit posted money for a period of either 12 weeks (1965–66 groups) or 16 weeks (1866–1967 group). Participants forfeited their money to the group if they smoked, but earned their money back by completing specified periods of abstinence. Of 25 participants in a college community, 21 abstained for the duration of the program, and the success rates at follow-up periods of more than a year (for the 1965–66 groups) and of 3 months (for the 1966–67 group) from the end of the program are 38 % and 36%, respectively.

The modeling of sharing: Effects associated with vicarious reinforcement, symbolization, age, and generalization

Abstract Four groups of young children matched for age, sex, and the sharing of candy were shown either a model sharing, a model sharing and being rewarded a model sharing, being rewarded, and told why he was rewarded, or no model. Posttests on similar and different kinds of sharing resulted in significant effects of modeling, with the reinforcement-with-instruction proving most powerful. Rewarded modeling was no more effective than modeling without reward. Sharing was significantly and directly associated with age. The sexes were the same in total amount shared but different with respect to which resource they preferred to share. Little generalization of modeling effects to other kinds of altruism occurred.

Simple reaction time: Effects associated with age, preparatory interval, incentive-shift, and mode of presentation

Abstract Children ( N = 216) aged 5–13 and young adults ( N = 72) performed in a simple auditory reaction time task for two sessions. The experimental conditions were low incentive in the first session followed by high incentive; high followed by low; and low followed by low. The preparatory intervals were 1, 2, 4, 8, and 16 seconds, and both the regular and irregular modes of presentation were used. In the Low-High groups the incentive effects were highly reliable in all age groups, but four to six times as great in the youngest children as in the adults, and greater on the harder tasks than the easier ones. A tendency for improvement in speed to be accompanied by a decreased dependence of the reaction time on the preparatory interval mimicked the same strong trend accompanying increasing age. There was no “depression” effect when incentives were shifted down in the High-Low groups. The form of the effect of the preparatory interval of the preceding trial on the reaction time of any current trial was about the same in children as in adults.

The role of ethnicity in choosing and leaving science in highly selective institutions

This study sought to assess the role of ethnicity in both initial choice of, and persistence in, science majors. Standardized test scores, high school records, initial concentration preference, college grades, and final majors of all the white, Asian, black, and Hispanic students who enrolled in 1988 at four highly selective institutions provided the database. Despite relative deficits in scores on measures of preparation and developed ability, blacks entered college with a strong interest in majoring in science. Black students interested in science also suffered the highest attrition from it; Asians were lowest, with whites and Hispanics near the average attrition of 40%. Ethnicity did not add significantly to ability and achievement variables in predicting attrition from science. The results are discussed in terms of two main issues: first, the effect of different standards of selection for the various groups on their success in science curricula; and second, the relevance of various well-known intervention strategies to the problems of minority attrition in science in highly selective institutions.

Expert testimony about eyewitness identification

Recent research concerning eyewitness identification is surveyed with respect to its adequacy (reliability and validity) to support expert testimony. The conclusion is that the scientific basis is generally inadequate and that the more we have learned about various aspects of eyewitness identification, the most inadequate it appears. The argument is made that presentation in policy settings requires greater circumspection than would be required in presentation to scientific audiences. Prudential rules are suggested. They would have the effect of sharply reducing participation in the adversarial system.

Social science data and the APA

The wisdom of the American Psychological Associations submitting amicus briefs to affect social or legal policy is questioned by an analysis of the brief claimed to be a strong example of the effective use of social science data in the public policy arena; namely, the APA brief (Bersoff & Ogden, 1987) inLockhart v. McCree (1986). The data relied upon in the brief do not appear to support the assertions based upon them, and other data are adduced to develop the critique. It is concluded that it is mischievous for the Association to address itself to the courts by generalizing a data base well beyond its useful limits: The adversarial and scientific methods of establishing truth are in several respects antithetical.

Simple visual and simple auditory reaction time: A comparison

In two experiments the classically reported difference of about 40 msec between simple RTs to high intensity tone and light was reproduced using 1° and 3° targets, but was reduced to 24 msec by illuminating virtually the whole retinal surface. In a third experiment, the indicated relations between target size and RT were repeated.

Sex differences in the educational and occupational aspirations of rural youth

Abstract Rural girls (N = 926) and boys (N = 861) in grades 10, 11, and 12 answered questions concerning educational and occupational aspirations. The young women showed significantly higher educational aspiration, the same or higher occupational aspiration, and equal ranges of job choice. The results are discussed with respect to previous findings, and to status attainment models.

EFFECTS OF SPECIFIC TRAINING ON FRAME DEPENDENCE

Forty-four normal adults were screened on the rod and frame test and two groups, each consisting of seven persons who made relatively large errors, were given training specifically designed to improve their performances. The first group received instructions on how to resist the influence of the tilted frame and how to use bodily cues. The second group received the same training, plus feedback and discussion of each judgment. The first group did not improve. The second did, but the improvement was transient and disappeared upon later retesting under standard conditions. Test-retest correlations within both groups were high.