Vi-Nhuan Le, RAND Corporation
2002
Vi-Nhuan Le is the author of this 200-page RAND study on behalf of the "Bridge Project" of the Stanford Institute for Higher Education Research. The core question is how well do the high school tests (of English and math) required by these five states line up with tests of the same subjects administered to first-year college students (for admissions and/or course placement, particularly decisions about whether a student needs remediation or is ready for college-level work). The states under review are California, Georgia, Maryland, Oregon and Texas, each of which gets a substantial case study here. The "alignment" being examined, however, has mostly to do with the tests' coverage or content, not their "performance standards." Do high school math achievement tests, for example, probe the same kinds of math and contain the same kinds of exercises as tests administered for college entry or placement? The less-than-electrifying conclusions are apt to hold greater interest for testing experts than education policy watchers, and in several cases the definition of "alignment" strikes me as too relativistic: "In math, there are no instances of misalignments, as discrepancies among assessments reflect variations in test use." In other words, the tests differ because they're used for different purpose but (in the author's conceptual framework) that doesn't mean they're mis-aligned. Much the same conclusion is reached in English, except for "one notable misalignment" when it comes to "the scoring criteria of the writing measures." Perhaps more interesting, "most state achievement tests cannot inform colleges admissions or broad course placement decisions...because they contain, on average, fewer problem-solving, inference, or advanced content items than do college admissions and college placement tests." To my eye, this is a needlessly complacent and passive view of test alignment at a time when many believe that the entire education system would be better served if these various tests were prodded to incorporate both the same content and the same performance standards. If you'd like to see for yourself, you can download a PDF version or order a hard copy for $10 at www.rand.org/publications/MR/MR1530.0/.
