The critics we reviewed believe that because of the general enthusiasm for college, many young people who do not have the ability or educational preparation to perform well in 4-year colleges nevertheless enroll in them. What do the data tell us about the qualifications of 4-year-college students and how their abilities have changed over time?26
A study by Berkner et al. (1997) sheds some light on the extent of enrollment in 4-year colleges by high school graduates of limited ability. Using National Educational Longitudinal Survey (NELS) data, the authors constructed a scale to measure qualification for admission to 4-year colleges. At the top of the scale were those "very highly qualified" students whose highest value on any of five criteria would put them among the top 10 percent of 4-year college students. The criterion items were high school grade point average, class rank percentile, NELS test percentile, combined Scholastic Aptitude Test (SAT) scores, and/or composite American College Test (ACT) score. At the bottom of the scale were "marginally or not qualified" students for whom all values were in the lowest quartile.
The authors found that a little over one-third (35.5 percent) of the 1992 high school graduates were marginally qualified or not qualified to attend a 4-year college. Of that one-third, 47.9 percent had no postsecondary education as of 1994, 37.3 percent were enrolled in 2-year or other subbaccalaureate institutions, and 14.7 percent were enrolled in 4-year colleges. Hence, most of these poorly prepared high school graduates with postsecondary experience were attending subbaccalaureate institutions, not 4-year colleges.
Viewing the surveyed students from another angle, the authors found that 11.3 percent of the NELS high school graduates who were enrolled in 4-year colleges were marginally or not qualified to attend, based on the criteria used. Another 12.6 percent were "minimally qualified," meaning that they did meet the minimal qualifications for 4-year college but did not exceed them by much. One may question the cut-point on the scale that distinguishes the marginally or not qualified from the minimally qualified. But even if we eliminate the line and say that somewhere between one-ninth and one-fourth of college students have shaky academic qualifications at best, it is not clear that this situation is different from the past or that such proportions are wildly excessive, given the multiplicity of goals in college admissions.
Have the abilities of 4-year college students changed over the last several decades? The American College Tests (ACTs) provide a good direct measure of the cognitive skills of these students. The tests are designed to yield valid and reliable information on the achievement of high school seniors interested in attending college. In addition to producing and administering the tests, the ACT organization conducts annual surveys of 2-year and 4-year colleges participating in its program to collect information from administrative records on freshmen who took the test in high school, usually in the previous school year. A broad range of colleges participates in the ACT Program, but they differ from the universe of American colleges in a number of ways. They are more likely to be located in the Midwest, the Rocky Mountain States, the Great Plains, and the South. They are also more likely to be public colleges and universities, less likely to be elite private universities. Like all surveys, the ACT college surveys have some nonresponse, which may introduce its own biases. Further, not all students in these colleges took the test. Despite these limitations, however, the survey data provide useful information about large numbers of freshmen at a wide range of American colleges, both public and private.27
Until 1989, the ACT measured reasoning abilities in 4 areas?English, math, social studies, and natural sciences. A composite score provided a summary of performance on the 4 tests. In 1989 the tests were rewritten. The social studies test was replaced by a reading test?not as big a change as it might seem, because the earlier social studies test had emphasized reading and problem solving. The other subject areas?English, math, and science?remained the same. Since scores on the new version of the test were higher than those on the old version, we have deflated them in order to make them comparable (see appendix C). Figure 1 shows the mean ACT composite scores of freshmen in the 4-year colleges surveyed.
The ACT composite scores of freshmen dropped between 1970 and 1980. This drop was part of the more general decline observed in a number of different achievement tests in the late 1960s and the 1970s, including the Scholastic Aptitude Test (SAT).28 The ACT scores then increased in the 1980s, and by the mid-1990s they slightly exceeded the levels of the early 1970s. Appendix figure B-1 shows the changes over time in mean math and English scores on the ACT tests. Both show the same basic pattern, except that the English scores in 1994 slightly exceeded those in 1970, and the math scores did not quite reach their 1970 levels.29
Bishop (1989) found a similar pattern in the General Intellectual Ability (GIA) scores that he developed using Iowa Test of Educational Development (ITED) scores, age, and schooling. Bishop estimated the GIA scores of 17-year-olds entering college (among others), based on the proportion of each cohort of seniors enrolling in college.30 Figure 2 shows the changes in these scores from 1948 to 1990.
Bishop reported a marked increase in the estimated mean GIA scores of college freshmen between 1948 and 1966. The quality of students entering college rose partly because of improvements in the achievement levels of high school graduates and partly because colleges were introducing more selective admissions policies. Like the ACT data, Bishop?s estimates then showed a drop in scores between 1967 and 1980, followed by a rebound. By 1990, the estimated GIAs of college freshmen were very close to those of freshmen who had entered college in 1966.
Bishop (1989) also calculated composite test scores for college seniors applying to graduate schools, combining information from the Graduate Record Examination (GRE) achievement and subject matter tests and from the Law School Admissions Test (LSAT), the Graduate Management Admissions Test (GMAT), and the Medical School Admissions Test (MCAT). He reported a test score decline in the late 1960s and in the 1970s, followed by a rebound in the 1980s to levels exceeding those in the mid-1960s. His composite score fell .121 standard deviations between 1966 and 1980 but then rose by .173 standard deviations between 1980 and 1989.
In general, there is not much evidence for the critics' contention that the college movement has swept large numbers of unqualified young people into 4-year schools. By one systematic measure, 11 percent of college students were not qualified to attend and another 13 percent had the minimal qualifications. We don?t know whether these proportions are larger or smaller than in the past. However, using composite ACT scores, we see that the measured abilities of college freshmen fell from the mid-1960s to about 1980, then increased and exceeded their earlier levels. Other data from Bishop (1989) show a similar pattern. Even though enrollments have increased greatly, college students today are at least as able as they were 4 decades ago, by these measures. If there has been growth in the number of lower-ability students moving into postsecondary education, much of it is apparently being absorbed by community colleges, consistent with Rosenbaum's argument.
 [Growth of College] |
![[ Table of Contents ]](/icons/up.gif){kind=icon} |
 [Changes in College Completion Rates] |