The method used by The Seattle Times to estimate graduation rates for Seattle-area high schools offers important advantages over the way graduation and dropout rates have been reported in the past.

• It tracks students over four years. Until a few years ago, Washington state's graduation rate was based on how many students made it from fall to spring of their senior year. Students who left earlier (as many dropouts do) weren't part of the equation.

• It uses readily available enrollment and graduate data. The enrollment data, in particular, are subject to careful scrutiny because districts receive money based on their student counts.

• It relies on a simple, common-sense definition of who is an on-time graduate — someone who leaves high school with a diploma after four years. Students who earn a General Educational Development diploma (GED) are not considered graduates, for example. And students whose status is unknown are not taken out of the picture.

Until recently, Washington schools did not have to report unknowns or GED recipients as dropouts. In nearly every year in the past decade, the unknowns outnumbered the confirmed dropouts.

The state's official method relies on dropout data reported by school districts. This year, virtually every district — out of habit or sloppiness, state officials say — failed to report students who had dropped out before their senior year.

The numerous definitions and approaches left the mistaken impression that many more students graduated from high school on time than now appears to be the case.

For example, Everett High's principal, until a few years ago, usually cited his school's one-year graduation rate, which is more than 90 percent. That's the number of seniors who start each fall and graduate in the spring.

When the state Office of the Superintendent of Public Instruction (OSPI) changed to reporting a four-year rate, Principal Pat Sullivan said the school figured it was in the low 80s, which his district office later revised to the mid-70s.

The Times' estimate for that class is 53 percent, a figure Sullivan disputes.

Some educators, by contrast, have long thought the state's numbers didn't provide an honest look at what was going on. Highline schools have long counted their "unknowns" as dropouts. "I'm not going to let us off the hook," said Superintendent Joseph McGeehan.

Critics of the approach used by The Times make two main arguments:

• It presumes that changes in a school's overall population are identical to changes in the size of the class of 2002.

But classes can vary from year to year for many reasons: changes in school boundaries, grade configuration or school size; a magnet program that affects the school's draw with students; a high number of students who transfer in and out; and school policies that retain students at particular grade levels.

The latter is one reason Everett High and Renton High believe The Times' numbers are too low.

Jay Greene, who developed this method for the Manhattan Institute for Policy Research, argues that it's reasonable to expect that growth or decline in successive classes will be similar.

• The method penalizes schools for students who need more time: those who finish over the summer, spend an extra semester or two in high school, or who leave a regular school to graduate from an alternative school.

Everett High, for example, has about 50 students in their fifth year this year, Sullivan said. And Renton school officials note that nearly 20 percent of high-school students there are in special education or learning English as a second language and often need more than four years to graduate.

But Greene said fifth-year seniors are often among the diploma recipients in any given year. As long as the number of fifth-year graduates is about the same every year, it wouldn't affect the method.

Pete Bylsma, research director for OSPI, agrees the objections raised by some educators don't invalidate the approach.

In fact, if Washington districts don't improve their dropout reporting, the state may start using an approach similar to the one developed by the Manhattan Institute, Bylsma said.

"I don't want to scare folks," he said, "but we do need to have a better sense of what's happening."