Of the more than 98,000 students who began their freshman year in Virginia's high school Class of 2004 -- the first required to pass state Standards of Learning exams to graduate -- fewer than 70,000 are expected to receive their diplomas this month. A child-advocacy group charged yesterday that the exams are partly to blame for the gap.

JustChildren, an arm of the Charlottesville-based Legal Aid Justice Center, held a news conference in Richmond to ask state educators to investigate graduation rates and their connection to the SOL exams.

"We support high standards, but we do not support punishing children," said Debra Grant, a Virginia Beach mother of three who spoke at the conference.

In recent years, the number of seniors who graduated in Virginia has been about 25 percent less than the number of freshmen who entered the class four years earlier. This year, according to a survey of school districts in the state, the decrease is projected to be about 29 percent.

Andy Block, legal director for JustChildren, said he suspects that the pressure of the tests discourages students and prompts them to drop out. "Given the significant percentage of students who end up without diplomas, we hope that figuring out why this is going on becomes an urgent priority," he said.

State education officials said a study has been commissioned, but they said that until individual students can be tracked more accurately, it is difficult to know whether students drop out, move out of state or transfer to private schools.

"We don't have the ability to look at the data and say with any certainty what happened to those students," said Charles Pyle, spokesman for the state Department of Education. "To make that comparison and suggest that the SOLs are responsible is irresponsible."

Virginia students take SOL exams in English, history, math and science in third, fifth and eighth grades and in high school. This year, for the first time, seniors have had to pass both the high school SOL reading and writing exams, plus four other SOL exams or tests from a list of accepted alternatives, before they can graduate.

When the testing regime began in 1998, passing rates were low, and educators feared that thousands of seniors would be at risk of not graduating this year, when the results would officially begin to count toward their diplomas.

Based on the survey, they now believe that the percentage of seniors who fail to graduate with their class will be similar to that of past years. But educators increasingly calculate graduation rates beginning with the freshman year, as required by the federal No Child Left Behind law.

Pyle noted that the survey, taken in May, was preliminary, and that several school districts have reported more students graduating than had been projected. City officials in Hampton, for instance, had predicted that 122 students would not graduate but have reported that only 27 did not get diplomas, he said. Final numbers will not be available until after school districts have finished holding graduations, many of them scheduled this week.

Pyle said state officials have asked the Commonwealth Educational Policy Institute to analyze graduation rates. The group, based at Virginia Commonwealth University, is headed by former Virginia schools superintendent William C. Bosher Jr.

A new information management system is expected to enable educators to track individual students beginning in summer 2006, according to Pyle. In the meantime, he said, school districts have made tremendous efforts to identify students who are struggling with the tests and provide local- and state-designed tutoring to help them pass.

Lisa Abrams, a Boston College research associate who has examined similar data for all 50 states, said calculating graduation rates starting with freshman year is more comprehensive. Otherwise, she said, "you're not factoring in students who have left school before 12th grade. It doesn't give you as much information about the capacity of schools in the state to graduate students in four years."

By that measure, Abrams's study found that Virginia ranked better than 36 states in recent years. She said the study shows that high-stakes testing has hurt graduation rates elsewhere, but she cautioned against making judgments in Virginia based on this year alone.

"It is only one year," she said. "You'd want to see what happened next year and the next."