Thomas L. Hilton

Sex differences in mathematics achievement: A longitudinal study.

AbstractWith the objective of investigating sex-typed interests as possible causes of difference in mathematics achievement between the sexes, this study made use of longitudinal data from the Growth Study, begun at Educational Testing Service (ETS) in 1961. Growth in mathematics achievement as measured by STEP Math and SCAT-Q was compared with changing interest patterns as reflected in certain biographical questionnaire responses. At grade 5 there were no differences in achievement but thereafter the boys pulled ahead, while parallel differences emerged in the percentage perceiving mathematics as interesting and as likely to be helpful in earning a living.

An Investigation of Item Bias

a concomitant increase in concern about the applicability of widely used tests in different cultural groups. The College Entrance Examination Board, for example, has become concerned about the appropriateness of the Scholastic Aptitude Test (SAT) and the Preliminary Scholastic Aptitude Test (PSAT) for some subgroups of the population, particularly Negro Americans. It is often difficult to determine what is meant by the word &dquo;bias&dquo; when it is used in reference to tests. Test &dquo;bias&dquo; is explored here in terms of individual test items. An item of a test is said to

Negro-White Differences in Adolescent Educational Growth.

The past 50 years have seen considerable consistency in published research concerning Negro and white differences. The results generally indicate that average scores made by Negro groups on intelligence tests are consistently and significantly lower than those made by white groups in the same communities (e.g., Coleman et at 1966; Tyler, 1956). Almost all of these investigations, however, have been cross-sectional in nature, an approach involving testing samples of subjects at each grade or age level under concern at one point in time. A recent nationwide study of this type was conducted by Coleman in 1966. He reports that whites scored higher than Negroes in each region of the United States and at each grade level investigated. In addition, he found that test score differences between Negro and white students were greater at grade 12 than those initially present at grade 1. These results are fairly typical of cross-sectional findings and

Career Decision-Making

Career decision-making is an imperfect process at best. Precisely what determines its course has received little direct attention, even though the larger problem of career development has been widely studied. In this paper, five possible conceptions of the process will be summarized and a theoretical model resulting from exploratory research by the author will be presented. The model is descriptive, not normative. We are concerned at this point with how individuals typically make career decisions, not with how they might optimally make them. The five different models are not an exhaustive inventory of available models described in the literature; the author has omitted those which represent integrations of the approaches which will be described. Two examples are the excellent models by Blau et al. (1956) and Ginzberg et al. (1951).

Demography, Demagoguery, and Test Performance

I have been asked to speak on demography and testing and, specifically, on the impact of demographic variables on the use of tests in schools. What testing should we be doing, knowing the demographic characteristics of our students? I presume that this request derives from a number of recent events, the first of which is the concern for the condition of schooling in this country. According to a recent poll “education has shot to the top of what Americans see as the important issues of the day” (Hart, 1983). This concern is fueled in part by the dramatic decline in Scholastic Aptitude Test (SAT) scores over the last twenty years. The latest evidence is that the decline may have leveled off. For the second consecutive year the average combined mathematical and verbal score has remained at 893 (College Entrance Examination Board, 1983). Nevertheless, the decline from 1963 to 1981 was substantial and real and may resume. Two colleagues at the Educational Testing Service (Em) and I demonstrated that a large share of the decline from 1960 to 1972, possibly as much as twothirds, was attributable to a broadening of the population of students taking the SAT, from a select group of students headed for the elite colleges to a more representative group of high school seniors applying to a broad range of postsecondary schools, including two-year colleges and vocational schools (Beaton, Hilton, & Schrader, 1977). But that leaves at least one-third of the decline from 1960 to 1972 unexplained and, because there is no evidence since 1972 of a significant change in the population of young people taking the SAT, it leaves all of the decline since 1972 unexplained. A number of plausible hypotheses have been advanced to explain the decline including a demographic hypothesis that I will discuss shortlybut to my knowledge, none of the hypotheses has been confirmed in any rigorous way, although some suggestive studies have been reported. A second reason for interest in demography and testing stems from what appears to be a growing reliance on testing to assist the schools in achieving higher standards of academic performance. An increasing number of states are adopting minimum competency testing. According to the poll mentioned earlier, the majority of Americans supports competency testing. Any such increase in the use of tests properly should trigger a heightened concern for their effect on the student population.