The number of student suspensions in Fairfax County schools fell by 24 percent last year, School Board members were told yesterday, but they disagreed over whether student behavior is improving or administrators are simply more reluctant to mete out punishment.

The drop far exceeded a goal the board had set to cut the number of suspensions by 10 percent. But some board members did not see that as good news, saying they suspected that administrators had suspended fewer students in order to meet the board's numerical target.

"Anecdotal evidence leads to grave suspicion as to how we got here. . . . It tends to show that there has been continued pressure on teachers not to suspend," said board member Gary A. Reese (Sully), who said teachers had told him of facing such pressure.

Reese and board member Mychele B. Brickner (At Large) said the board needs to reassess whether setting a numerical goal on suspensions is the right approach.

But other board members and school district staff disagreed, attributing the drop in suspensions to alternatives such as increased use of "timeout" rooms and a new program that keeps some disruptive students in school but removes them from a regular classroom. They also cited a school district survey last year in which teachers said that student behavior had improved and that they were spending less time on discipline.

"We ought to be making policy on the basis of data, not anecdotes," said board member Stuart D. Gibson (Hunter Mill).

The debate came as the board held its annual retreat and heard how the school system is progressing in meeting the 10 goals the board set two years ago. Based on the status reports, the board will decide whether to modify or eliminate those targets and whether to add new ones.

School officials said the district fully achieved three of the 10 targets--two goals related to technology and one related to allowing schools more flexibility in funding and staffing. Partial success was achieved on the other seven.

The goal involving student discipline was not fully met, despite the 24 percent drop in suspensions from 1997-98 to 1998-99. The board also had wanted to reduce by 10 percent the gap in the suspension rates for white and minority students because of concerns that minority students were being suspended in disproportionately high numbers. That reduction did not occur, as suspensions declined by at least 20 percent for every ethnic group.

Board member Ernestine C. Heastie (Providence) said she was concerned that eliminating the numerical goals would send the wrong message to minority parents.

"[The target] dealt with an issue that is very important to minority parents," Heastie said. "My concern is that minority parents will see this as the school system backing off" its commitment to reducing the racial disparity in suspensions, she said.

Preliminary data show that overall suspensions for the first quarter of the current school year were down about 9 percent over the first quarter of the 1998-99 school year.

School staff said that in addition to the new alternatives to suspension, there has been a rethinking of what kind of punishment is appropriate for certain offenses. For example, Fairfax schools used to regularly suspend students who cut school or skipped class, but now those students are assigned to in-school or after-school programs or to Saturday school, officials said.

Board members had wanted the overall number of suspensions cut because they thought the punishment was being used too often. They will decide at a future work session whether the numerical goals should be revised or abandoned.

The board will also consider adding new targets. Among the proposed numerical goals discussed yesterday were increasing the number of children who are reading at grade level by the end of first grade; placing more disabled children in regular classrooms at their neighborhood school; increasing student enrollment in career academies at high schools; and reducing the number of portable classrooms in the district.

CAPTION: Ernestine C. Heastie spoke against eliminating numerical goals.