indicates that system capacity lags far behind existing enrollment. BOE measures building utilization annually in its enrollment capacity utilization (ECU) report. The most recent ECU report in evidence was for the 1998-1999 school year. The report describes the formulas by which BOE determines the capacity of each school. Because buildings are used differently based on level and type of schooling, BOE uses different formulas to determine the capacities of elementary schools, middle schools, high schools and City-wide special education. The utilization numbers reflect widespread overcrowding. Almost 60% of all elementary schools, serving 63% of the City's [187 [187 Misc.2d Misc.2d 50] 50] elementary school students, and 67% of high schools, serving 70% of the City's high school students, are overcrowded under the formula used by the ECU report. Overall, the numbers indicate that approximately 59% of New York City public school students attend overutilized schools. Overcrowding tends to plague certain school districts more than others. For example, all of the Queens community school districts are overutilized, and five of seven are operating in excess of 110% capacity. Overcrowding is even worse than indicated above because the ECU formulas actually overstate schools' capacity. This inflation occurs because the formulas adjust for overcrowding by adding to schools' capacity nonclassroom space if such space is in fact used for classrooms. For example if a crowded school is forced to convert its gymnasium or auditorium into classroom space, the capacity formula indicates increased capacity. Second, there was credible testimony that the amount of space allocated to each pupil is too small, failing to allow sufficient room for a teacher's desk, for a set-back for student desks sufficient for all students to view the blackboard, or for space for educational tools such as computers. Additionally, the formula assumes class rosters larger than what is optimal from a pedagogical standpoint as measured by State and Federal goals. 27 Overcrowding has numerous negative consequences for students. Specialized spaces in school, such as gymnasiums, science labs, libraries, and art rooms, must be taken over for full-time classroom space, thereby depriving students of the programs intended for those rooms. Marginal spaces in a school building, including undersized offices, hallways, and storage space, have been converted to classrooms. Many school administrators have had to resort to the use of trailers and other temporary structures. Such structures outside the main school building isolate teachers and students from the rest of the school and often occupy much-needed playground space. The students taught in the temporary structures still must use the main school building's cafeteria and gymnasium. Temporary structures are often difficult to heat and they frequently lack sufficient power for computers. [187 [187 Misc.2d Misc.2d 51] 51] Some schools, particularly in Queens, have been forced to employ extended day and multishift schedules to relieve overcrowding. SED has granted waivers to seven Queens schools, with a total enrollment of approximately 25,000 students, to operate two shifts of six periods a day, instead of the State-mandated seven. This response reduces students' school day, obviously a less than optimal solution. At-risk students in particular often need a longer than usual school day—more "time on task" in current educational jargon. School buildings become extremely crowded during the overlap of the two shifts, and the educational program is distorted, resulting in anomalies such as 150-student gym classes. Overcrowding often necessitates the creation of multiple lunch shifts. There was credible testimony that such staggered lunch shifts over the day can have a negative effect on students who must take lunch early in the morning or late in the afternoon. It would not be news to generations of school teachers that hungry students find it difficult to focus in the classroom. For these reasons increased utilization of existing school facilities is not a satisfactory solution to overcrowding. 1. Class Size The most significant negative impact of overcrowding is its effect on class size. In schools that are bursting at the seams there is little or no room to allocate the necessary space to reduce the number of students per class. The evidence demonstrates that class size has an effect on student outcomes, and that smaller class size can boost student achievement, particularly among at-risk children. The advantages of small classes are clear. A teacher in a small class has more time to spend with each student. Fewer students mean fewer administrative tasks for each teacher. Student discipline and student engagement in the learning process improve in smaller classes. The Federal Government, SED, and BOE all agree that smaller class sizes help students to learn, and all have initiated programs to reduce class size. Although the Federal and State programs differ slightly in the specifics and methods of implementation (the Federal program seeks to lower class size in grades 1 through 3 to an average of 18 students, while the State program seeks to lower class size in grades K through 3 to an average of 20 students) their goals are similar: to reduce class size in the lower grades to 20 students or fewer. The [187 [187 Misc.2d Misc.2d 52] 52] court also finds credible the testimony of the numerous New York City district superintendents called by plaintiffs who were unanimous in their assessment of the advantages of smaller classes. The positive effect of smaller classes on student outcomes is also documented by statistical evidence. Plaintiffs' expert Jeremy Finn, a professor of statistics at SUNY Buffalo, testified at length about the Tennessee student teacher achievement ratio (STAR) project, a landmark study of the effect of class size on student achievement. The STAR project was conducted in Tennessee with cohorts of students in 79 schools. The students were randomly divided into three types of classes: (1) small classes of between 12 and 17 students with one teacher, (2) regular-sized classes of between 22 and 26 students with one teacher, and (3) regular-sized classes with one teacher and one teacher's aide. Teachers were randomly assigned to the classes. The controlled part of the study followed the students that started kindergarten in 1985 through their graduation from third grade. Students enrolled in the STAR program were tested at grades K through 3. The students were tracked after they left third grade to see how they performed as they moved through elementary and secondary school. Nearly 50% were receiving free lunch, an indicator of low economic status, and approximately one third were minority, almost entirely African-American.