A full 83 percent of students statewide in grades four through eight were correctly matched with their teachers as school officials calculated job ratings that were sent to local districts earlier in the month.
State school officials said those calculations excluded the other 17 percent of students -- more than 160,000 statewide -- whose test scores from the 2011-12 school year were not used in teacher ratings. However, those students' scores were used in rating school principals.
Students were excluded from teacher gradings either because school records indicated they were not taught by those teachers for the minimum number of days required by the state, or because records were not available. Student transfers out of classes accounted for some exclusions; teacher transfers accounted for others.
Percentages of students excluded varied by grade level and subject. The state Education Department reported, for example, that 20 percent of seventh-grade math scores, but only 13 percent of fourth-grade English scores, were omitted.
The department expects higher percentages of students to be included in future years, as the ratings system is refined.
The new figures were released after several Long Island school superintendents voiced concerns about possible inaccuracies in ratings. Some doubts persist, local administrators have said.
Concerns stemmed largely from an earlier report by the American Institutes for Research, the nonprofit agency that developed the rating system. AIR reported that more than 50 percent of fourth- and fifth-graders could not be matched to teachers in a trial round of ratings based on test scores from the 2010-11 school year.
AIR's technical report, released Aug. 16, did not address the percentage of students unmatched in the latest round of ratings, which was based on 2011-12 testing. The Education Department encouraged, but did not require, districts to link students and teachers in 2010-11, but now requires it.
Tom Dunn, an Education Department spokesman, noted that only 2011-12 test scores were used in the latest ratings. He said all districts had student rosters and could arrange for teachers to check rosters for accuracy. "Therefore, we are confident that the growth scores [ratings] reported for educators for 2011-12 link student performance with the proper teachers as accurately as possible," Dunn said in a prepared statement.
Under a new job-evaluation system, state ratings will count for 20 percent of teachers' marks. The remainder will be based on results of tests selected by local districts, as well as on classroom observations and other subjective measures.
Teachers rated as less than effective face possible job loss. The new system will not kick in for most teachers until next year.
Questions about the system's accuracy linger. Tom Rogers, superintendent of the regional Nassau BOCES, addressed the issue in an Aug. 22 memo to local district chiefs.
"The linkages [between students and teachers] are considerably better in the 2011-12 school year, although still fall well short of 100%," Rogers wrote. "So some will undoubtedly continue to raise concerns about the methodology. However, perfect linkage may not be feasible."