Middle School Class Sizes

San Francisco's are smaller than average but far from being the smallest

We saw last time that elementary school class sizes in San Francisco are among the lowest in the state. Today, we’re going to look at how class size varies from middle school to middle school, and at how the average class size here compares with the average in other districts.

The source of data is again the Student Accountability Report Cards (SARCs). SARCs are published annually but the data is delayed so the most recent data we have is for the 2021-22 school year. To give a sense of what data is available, here’s the relevant excerpt from the SARC for AP Giannini middle school.

Figure 1: Except from SARC for AP Giannini Middle School

I’m not going to do subject-by-subject comparisons. Instead, I calculated a weighted average class size (i.e. the average class size for each subject weighted by the number of classrooms in that subject) for each school. In the case of AP Giannini, that average is 25.6.

Figure 2 shows the average class size in each of the San Francisco middle schools, some of which are pure middle schools serving grades 6-8 and some of which are K-8 schools. The y-axis is the average class size and the x-axis is the unduplicated pupil percentage (UPP) i.e. the percentage of students who are English learners, eligible for free or reduced-price meals, or foster youth, which is a rough proxy for need. The dot size is the enrollment per grade in each school.

Figure 2: Average Class Size and Unduplicated Pupil Percentage for all SFUSD Schools That Serve Middle School Students

We can see that there is a huge range, from 28.3 (Alice Fong Yu and Bessie Carmichael) to 13.5 (Willie Brown). If we ignore the K-8 schools for a minute and focus exclusively on the pure middle schools, it is noticeable that the larger the school the larger the average class size. Giannini is the largest school and averaged 25.6 per class. The next six largest middle schools (Presidio, Hoover, Aptos, Denman, Marina, Roosevelt) all have average class sizes between 22 and 23.7 while the six smallest middle schools (Everett, Francisco, Lick, King, Visitacion Valley, and Brown) all have much smaller average class sizes, from 20.2 down to 13.5.

A few other observations:

• It’s hard to justify many of the observed class size differences by reference to need. Why is Bessie Carmichael’s class size (28.3) nearly twice as high as Visitacion Valley’s (15.0)? Of course, even if one school does have demonstrably more need, it’s not obvious that having different class sizes is the best way to address that need. Would it be better to have some tutors who can give focused attention to the students who most need help?

• Although six of the eight K-8 schools have more than 500 students, the fact that the schools serve nine grades means that there are relatively few students per grade. Some of the schools also mix language and general education programs. This leads to a lot of variability in class size.

• SF Community School is one of the schools with particularly small classes (average 15.5) but I heard a parent complaining during public comment at a Board of Education meeting that the school still didn’t have an assigned middle school math teacher. A larger class with a teacher is surely better than a smaller class without one.

• Some schools contain more special education students who require smaller classes, thereby pushing those schools’ averages down.

• The district has attempted to create pathways where each elementary school feeds in to a middle school. But the district still allows open enrollment and some middle schools are more popular than others. While there has been an overall drop in enrollment, it has not been evenly spread out. Enrollment has declined significantly at some schools but not others. Staffing has not been adjusted at the schools with reduced enrollment.

• Even before the recent enrollment decline, it appears to have been district policy to have much smaller classes at Willie Brown.

Staffing

The SARCs also include information on staffing at schools. Staffing includes not just teachers but support staff such as counselors, librarians, social workers, and nurses. Here is the total staffing at each school compared to its enrollment.

Figure 3: Staffing at each SFUSD school serving middle school students

Smaller schools have more staff per 100 students. That’s what we would expect given that we saw in Figure 2 that larger middle schools have larger class sizes which means fewer teachers. There are so few support staff that their inclusion doesn’t change the order that much. But the distribution of support staff is still head-scratching. Everett had the most support staff (1.6 per 100 students), but its demographic and geographic cousin, Lick, had one of the lowest (0.5 per 100 students). The schools with the lowest support staff ratios were Lawton (0.3 per 100) and Hoover (0.4). The schools with the highest support ratios after Everett were King (1.4) and Revere (1.3).

These support numbers should, however, be taken with a very large pinch of salt. As we’ve discussed in the past, because of budgetary shenanigans, some of the people who work at school sites are listed as working at the central office. As the SARC puts it:

For various reasons, including funding sources and employees serving multiple locations, many of the staff in these positions are listed as working at central administrative offices, and so, will not show up under a particular school's staffing count. This includes the following FTE: Counselors - 11.8, Librarians - 31.6, Nurses - 55.3, Psychologists/Social Workers - 111.6 and Resource Specialists - 155.

In other words, the SARC data significantly underestimates the total number of support staff but whether those central office support staff are allocated in an equitable way (however you define equitable) is impossible to assess because no data is ever provided.

Comparison With Other Districts

We can compare the average class size in San Francisco’s middle schools with that of other districts. In figure 4, the y-axis is the average class size weighted by the enrollment in each school and the x-axis is the unduplicated pupil percentage i.e. the percentage of students who are English learners, eligible for free or reduced-price meals, or foster youth. Districts with a higher UPP¹ get more money from the state and one way they could choose to spend that extra money is on smaller class sizes. I didn’t include K-8 schools in the average for technical reasons².

San Francisco’s average middle school class size is 22. Recall that AP Giannini had the highest average class size in San Francisco (25.6). That’s still below the average class size in large Bay Area districts such as San Ramon Valley, Fremont, and Mt. Diablo. Long Beach (27.2) is higher again but the highest average class size is to be found in financially stable Clovis (31.9).

Nevertheless, there are some large districts (e.g. Los Angeles, Capistrano) where the average is even lower than in San Francisco. Famously troubled Oakland has one of the lowest average class sizes (17.5) while Santa Ana and San Bernardino are also in the 18-19 range. These districts also tend to have a lot of variability from school to school. In Oakland, West Oakland Middle had an average class size of 10.7 while Edna Brewer Middle had an average of 23.3. In Santa Ana, the average ranged from 11.8 to 21.0.

Conclusion

Middle school class sizes are considerably smaller than the state average but a long way from the the smallest. There is huge variability from school to school that in part reflects policy and in part reflects the district’s delay in adjusting staffing to match enrollment.

1. The UPP that I’m using is the weighted average of the UPPs of the individual middle schools rather than the district-wide UPP. In practice, there won’t be much difference between the district-wide average across all schools and the average just for middle schools.

2. The technical reason is that I’m too lazy. In order to calculate the weighted average class size, I’d need the enrollment in the middle school grades (as opposed to the elementary school grades when kids are in self-contained classes, not classes by subject). I didn’t have the data for enrollment by grade and I was too lazy to add it.