Measures commonly used in admissions and scholarship selection in higher education are discussed. Alternative measures that may improve assessments for the increasingly diverse applicants to colleges and universities are evaluated.
Although we have a 100-year history of interest in admissions testing in higher education through the College Entrance Examination Board, the perspective of decision makers in exploring a variety of measures for admission and scholarship selection has been very limited. Part of the reason for this has been that things seemed to be moving along as expected through most of the last century. The Scholastic Aptitude Test (now the SAT I) and its derivatives began in 1926, and the American College Test (ACT) started to compete in 1959. The Graduate Record Examination (GRE) began in 1949 for use with graduate school applicants. Test developers, educators, parents, and students seemed to accept things fairly well until the 1960s. For many different reasons--statistical, political, social, and educational--criticism reached a threshold in the 1960s (Crouse & Trusheim, 1988) and has been increasing steadily since then.
The biggest problem with the standardized tests that are used in higher education is that they measure a limited range of attributes, namely, verbal and quantitative ability. Throughout most of the last century, test developers have been refining available measures, with some success, but there are limits to the ability to develop such measures. Even if the population of potential students might show measurable differences on the constructs of verbal and math ability, current methods of test development may not show those differences. Thus, there may be no more practical variance to assess in the verbal and math areas. Although these dimensions have been shown to be useful for most of the applicants to higher education in the last century, they do not seem to adequately assess the potential of the much wider range of applicants currently seeking access to colleges and universities in the United States.
It is well documented that U.S. college and university students have been increasing in diversity on dimensions of race, gender, sexual orientation, religion, age, disability, and many other attributes (McTighe Musil et al., 1999). These changes will likely have a profound impact on the kinds of attributes that should be measured to predict student success.
Some measures of those attributes have already been developed and validated on many student groups, but they have not been as widely used as they might be. One reason for this might be what has been called the "Three Musketeers" problem (Sedlacek, 1994).
The rallying cry of "all for one and one for all" is one that is used often in developing what can be thought of as fair and equitable admissions measures. Commonly, the interpretation of how to handle diversity is to hone and fine-tune measures so they yield equally useful assessments for everyone (Berk, 1982; Helms, 1992; Sackett, Schmidt, Ellingson, & Kabin, 2001; Sedlacek, 1986). However, if different groups have different experiences and different ways of presenting their capabilities, it is unlikely that a single measure, test item, and so forth could be developed that would result in equally fair outcomes for all. If there is a concentration on results rather than on intentions, it could be concluded that it is important to do an equally good job of selection for each group, not that the same measures should necessarily be used for all to accomplish that goal. Equality of results, not process, should be the goal. Therefore, the variance that exists across diverse groups in assessments should be retaine d rather than eliminated.
MEASURES OF INTELLIGENCE
Recent work in assessing intelligence may help in understanding the Three Musketeers problem better. …