By Jay Mathews
Washington Post Staff Writer
Tuesday, May 23, 2006
Economist Larry Mishel was troubled by high school graduation statistics that contradicted what he thought was good research. That was particularly true of data used by many politicians and pundits to bemoan a 30 percent dropout rate in American high schools.
"This picture was radically different from what I knew from labor market data I regularly examined in my studies of wage and job trends," said Mishel, president of the Economic Policy Institute, a Washington-based think tank. His research indicated that only about 12 percent of the workforce lacked a high school diploma or its equivalent, so how could the dropout rate be so large?
Political scientist Jay P. Greene also had trouble with the data, but for a different reason. He found many school systems were claiming low dropout rates, even though their ninth grades were bulging with restless students eager to be elsewhere and many had disappeared by graduation time. Working as a senior fellow at the Manhattan Institute and as head of the education reform department at the University of Arkansas, Greene reported that graduation rates seemed to be worse than many people thought, as low as 50 percent in low-income urban neighborhoods.
A collision of those two views by prominent scholars was inevitable, and in the past several weeks it has hit the education policy world in an explosion of articles, e-mails and public debates, some quite heated. Experts disagree over who is right, and some say the truth may be somewhere in between. But the argument has aggravated a widespread feeling that information on how many children are disappearing from public schools is not nearly as accurate as it should be.
"Jay Greene and Larry Mishel have performed the valuable service of exposing the huge inadequacies in the way we measure the percent of students who achieve a regular high school diploma -- inadequacies not attended to in over two decades of education reform," said Paul E. Barton, senior associate in the Educational Testing Service's Policy Information Center.
Such congratulatory words have not ended the scholarly strife. Mishel and Greene continue their sometimes testy exchanges, and the argument has broken into disputes over lost diplomas, growth computation mistakes, uncounted immigrants and other issues loved only by people whose livelihoods depend on population data.
The major event has been the publication of a book by Mishel and Economic Policy Institute economist Joydeep Roy, "Rethinking High School Graduation Rates and Trends." It is only 100 pages, many of them full of charts, but it takes a big swing at powerful forces, particularly the National Governors Association and its recent report that said high schools are in crisis.
"About a third of our students are not graduating from high school," the association declared in a 2005 report by a task force that used Greene's data. "About three-fourths of white students graduate from high school, but only half of African American and Hispanic students do."
Mishel and Roy say that is wrong. Using U.S. Education Department data that follow student experiences and results of Census Bureau household surveys, they get very different numbers: an overall high school graduation rate with a regular diploma of 80 to 83 percent, a black student graduation rate of 69 to 75 percent and a Hispanic graduation rate of 61 to 74 percent.
They say that in the past 40 years, the high school completion rate, including graduates and those passing General Educational Development diploma tests, has gone up substantially and that the black-white gap has shrunk, except in the past 10 years, when there has been little improvement. Only graduation among Hispanics increased during the past 10 years.
Greene and Manhattan Institute research associate Marcus A. Winters have quickly counterattacked. They say the Mishel-Roy book is too dependent on Education Department longitudinal studies that follow a representative sample of students over several years and on census surveys that depend on people telling the truth about their success in school. If, for example, there were as many high school graduates in 2003 as Mishel and Roy said, they would number 476,442 more than the number of students school systems reported that year, Greene and Winters said.
Russell Rumberger, a University of California at Santa Barbara education professor, said he carefully checked the longitudinal survey used by Mishel and Roy and found that it appeared to "generate very accurate population estimates confirmed by published data." Greene struck back with a political analogy. He said the "assertion that we should believe the results of a survey over population counts is a little bit like the people who asserted that Kerry really won the 2004 election because the exit polls showed him winning even though the vote count gave the victory to Bush."
Daniel J. Losen, senior education law and policy associate at the Harvard University Civil Rights Project, said he agreed with Greene that the dropout problem is severe. "There is a consensus that this crisis is real and particularly severe for Latinos, African Americans and Native Americans," he said.
Researchers say this is not just an academic question; there are consequences for many children. "If Larry Mishel is right that the graduation rates have been improving, then some of the radical reforms for high schools being proposed may be misguided or dangerous," said Richard Rothstein, a former New York Times columnist and a research associate at Mishel's think tank.
"It may seem that we are talking about just a few percentage points here and there," said John Robert Warren, an associate professor of sociology at the University of Minnesota who agrees with Greene, but "five percentage points would be 175,000 young people annually."
No matter who is right, Barton said, it is embarrassing for educational research to have scholars as reputable as Mishel and Greene be so dubious about the value of major sources of dropout data.
Barton said census officials told him that there had been no field or validity studies of the census question on high school completion rates -- so experts cannot be as confident about that data. By contrast, he said, "tens of millions of dollars have gone into getting the questions right in that survey that gives the monthly unemployment rate."