Brown Logo

The News Service
38 Brown Street / Box R
Providence RI 02912

401 863-2476
Fax 863-9595

Distributed June 29, 2004
Contact Kristen Cole

Study: Chance to retake the high-stakes high school exit exams is critical

Researchers at Brown and Harvard universities say that the opportunity for high school students to retake high-stakes exit exams will likely impact both the size and diversity of the group that will eventually pass and obtain a high school diploma. These and related findings are in the August 2004 issue of Economics and Education Review.

PROVIDENCE, R.I. — Without the option to retake the high school exit exams, the ranks of students who obtain high school diplomas would be substantially different, according to a new study by researchers at Brown and Harvard universities.

Using 10 years of data on the General Educational Development (GED) test, researchers found the pass rate increased most dramatically for black test-takers who retook the exam. For black test-takers, the pass rate increased from 52 percent to 73; for white test-takers, the pass rate increased from 78 percent to 90 percent.

“Like state exit exams, the GED exams are high-stakes tests – in both cases the results matter to the futures of the individuals taking the tests,” according to the authors. “The GED candidates we looked at in this study are very much like the students with whom we are often most concerned when considering the effects of state exit exams – relatively low-skilled students who are on the margin of dropping out of school.”

Many states now require students to achieve passing scores on standardized exams in order to obtain high school diplomas. But implementing those exams is difficult. State officials struggle to determine the design, the minimum passing scores, and rules regarding options for students who fail, according to the researchers.

John H. Tyler, associate professor of education, public policy, and economics at Brown University, led the study with Richard J. Murnane and John B. Willett, professors of education in the Harvard Graduate School of Education. Their findings appear in the August 2004 issue of Economics and Education Review

Their paper, titled “The devil’s in the details: evidence from the GED on large effects of small differences in high-stakes exams,” looks at high school dropouts in Florida and Texas who were between the ages of 16 to 19 when they first attempted the GED from 1991 to 2001.

In Florida, the percentage of students who obtained the GED credential on their second try included, 21 percent of blacks, 17 percent of Hispanics, and 13 percent non-Hispanic whites. In Texas, 18 percent of blacks, 14 percent of Hispanics, and 10 percent of non-Hispanic whites, obtained the GED on their second try.

It is unclear whether these statistics would also characterize a situation where students are allowed only one chance to take a high-stakes exit exam, researchers said. Given a single chance to pass any high-stakes exam, it is possible that more test-takers would better prepare to pass the exam, and would pass it on the first try.

The GED also provides an opportunity to examine the relationship between different subject areas and groups of test-takers who pass or fail. For example, females who failed the GED scored lowest on the math exam; males who failed scored lowest on the writing exam.

Researchers attempted to determine whether introducing high school exit examinations would increase the number of students who drop out of school. Using information on test-takers in Texas, which raised its passing standard in 1997, the researchers found support for a temporary “dropout” effect of high-stakes testing. Raising the GED passing standard in 1997 resulted in a short-term decline in the size of the test-taking pool, but the size of the pool recovered over the next five years.

Of the expected half-million individuals who will drop out of school this year, about one in three will eventually earn a GED, according to the researchers. To obtain a GED, individuals must take five tests that require more than seven hours to complete; the tests cover math, writing, science, social sciences, and reading.

The study was supported by the National Center for the Study of Adult Learning and Literacy (NCSALL), and the Russell Sage, Smith Richardson, and Spencer Foundations.


News Service Home  |  Top of File  |  e-Subscribe  |  Brown Home Page