Charter Schools in San Francisco
Some have closed the achievement gap and are outperforming SFUSD.
Basic Facts and Figures
Enrollment Growth
Charter school enrollment has grown steadily since they were first allowed in the 1990s.
Five Keys is a network of three charter schools which primarily serves adults, many of whom are in jail, many of whom are in Los Angeles. The CDE includes them in the SFUSD enrollment numbers because they receive their charters from SFUSD but I will be ignoring them in the rest of this report.
Even excluding Five Keys, the other charter schools have been increasing their share of the public education market. In 2020-21, they served 8.1% of all public school students in the city, up from 4.9% in 2010-11.
Charter Schools by Grade
Charter schools are most popular at the middle and high-school grades.
Where They Are
All the charter schools are located in the center and east of the city. This has a predictable effect on who attends them.
Who Attends Them
In SFUSD schools, Asian students form the largest group at 35% of all students. But the aggregate population of charter schools is 55% Latino, 15% Black and only 5% Asian.
It follows that the fraction of public school students attending a charter school in preference to a SFUSD school also varies by race.
Charter schools serve 8% of all public school students in San Francisco but 16% of all African American students and only 1% of all Asian students. This is not spread evenly across grades. In the middle school grades, 25% of African American students and 22% of Latino students attend charter schools.
Having said that, it doesn’t really make sense to speak of charter schools in the aggregate, not only because they have no common governance but also because individual schools serve very different populations. Mission Prep, Edison, and Leadership are all > 80% Latino. KIPP’s high school is over 90% Black and Latino. The three other KIPP schools (two middle schools and an elementary school) and Life Learning Academy are all > 80% Black and Latino while City Arts and Tech is at 78%. Meanwhile, the combined White and Asian population is below 11% at all those schools. The two Gateway schools are more balanced. Meanwhile New School is > 40% White and Creative Arts is > 50% White.
We can compare the charters to SFUSD’s own schools. The horizontal axis shows the percentage of kids who are eligible for free or reduced-price meals (“FRPM eligible”). The average in San Francisco public schools is close to 50%. The vertical axis measures how close the school’s racial/ethnic composition is to the public school average using a measure called a divergence index that I’ve previously used here and here. The smaller the divergence index the closer a school’s racial/ethnic composition is to the public school average.
Here’s the picture for elementary, middle, and K-8 schools:
Gateway Middle and New School come closest to the public school racial mix but New School has considerably fewer FRPM-eligible students.
The three KIPP schools (one elementary and two middle) serve particularly poor populations. They all serve a much higher percentage of FRPM-eligible students than the three majority-Black SFUSD elementary schools.
Here’s the picture for high schools. Gateway High sits right at the city average in terms of both racial/ethnic composition and FRPM-eligibility. Of all city high schools, KIPP serves the highest percentage of FRPM-eligible kids and Leadership is not far behind.
Are The Charter Schools Doing a Good Job?
Enrollment
One quick and dirty way to assess how well a charter school is doing is to look at its enrollment. Schools that are doing well will show more stable enrollment because they retain more students and can fill any vacant spots from a waitlist. Fluctuating or declining enrollment is an indication of problems.
The KIPP schools are thriving. Each new school seems to be able to ramp up enrollment as it adds grades and enrollment then reaches a steady state.
Gateway’s two schools are also thriving.
Among the other K-8 schools, Creative Arts, Edison, and Mission Prep all seem to be thriving and New School is growing fast as it adds grades. Note that some charters have failed, most recently OnePurpose which had its application for charter renewal rejected by SFUSD.
The other high school charters are a mixed bag. Leadership went through a rocky phase in the late 2000s but seems to have recovered. City Arts and Tech was solid for a decade but appears to be struggling now. Whether it meets the fate of its sister school, Metropolitan Arts and Tech, remains to be seen.
Class Sizes
Why do parents choose to send their kids to charter schools? It’s not because they have more money. Just like SFUSD, charter schools get most of their money from the State of California using a formula that’s based on average daily attendance, grade, and the Unduplicated Pupil Percentage (UPP). The UPP is the percentage of students who are English Learners or FRPM-eligible. Districts (and charter schools) with more of these high-needs students get more money. Money buys staff so we would expect the class sizes to be smaller at schools with more high-needs students.
The most recent class size data that is publicly available dates from 2018-19. Since the schools have a mixture of grade spans (K-8, 6-8, 5-8 etc.), I picked sixth grade as the point of comparison and calculated the average class size for 6th grade English and 6th grade Math at each school.
As expected, there is a bit of a downward trend: the schools with the fewest high-needs students have the largest class sizes. But this explains only a tiny part of the observed class size variability.
It is noticeable that the four charter schools with large concentrations of Black and/or Latino students (Edison, Mission Prep, KIPP Bayview Academy, and KIPP San Francisco Bay Academy) have 6th grade class sizes in the 26-30 range while the SFUSD middle schools with similar demographics (Everett, Lick, Visitacion Valley, Willie Brown) have class sizes in the 14-22 range. A key question for everyone associated with SFUSD should be: Why are Bayview parents choosing to send their kids to KIPP Bayview Academy with its average class size of 29 when they could send them to nearby Willie Brown Middle School with its average class size of 15?
Elementary and Middle School Achievement
The benchmark for how to measure charter school performance is set by Stanford University’s Center for Research on Educational Outcomes. Their last national report dates from 2013 but they have published more recent reports on particular states and cities, but not San Francisco. The report that attracted most attention was the one on New York City, partly because that’s the biggest school district, and partly because the results were so striking. The report showed that Black and Hispanic students gained a significant academic advantage from charter school enrollment with Hispanic charter school students performing at the same level as their white district school peers. Within the universe of charter schools, there was a lot of variability but “charter schools affiliated with a Charter Management Organization [such as KIPP] have better learning gains than district school peers in both reading and math. The positive impact is equivalent to about 57 days of learning in reading and 103 more days in math”1.
I don’t have that data to produce a similar analysis for San Francisco. All I have are the most recent standardized test scores, which are from 2018-19, broken down by student subgroup. Let’s focus on the largest subgroup in charter schools, viz. “socioeconomically disadvantaged Hispanic or Latino” students.
There are 2.5 times more (2400 vs 900+) economically disadvantaged Latino students in SFUSD than in charter schools. But the number who meet or exceed the standard in English is barely higher in the district schools (just over 500 vs just under 500). In Math, more socioeconomically disadvantaged Latino students at charter schools met or exceeded the standard in Math than at all SFUSD K-8 and middle schools combined.
How is that possible? Well, here’s the percentage who met or exceeded the standard in Math at each school:
The two KIPP schools and Edison all have lots of socioeconomically disadvantaged (SED) Latino students and all do much better than comparable SFUSD schools such as Everett and Lick. But Mission Prep is the stand-out with an astonishing 74% of its students meeting or exceeding the standard. For ELA, the figure is even higher: 81%. That’s not just better than Latino students do at any SFUSD school, it’s better than socioeconomically disadvantaged Asian (67% in Math; 66% in ELA) and White (43% in Math; 53% in ELA) students do in SFUSD. It’s even comparable to the numbers for non-disadvantaged Asian (78% in Math; 80% in ELA) and White (74% in Math; 82% in ELA) students. In short, Mission Prep’s socioeconomically disadvantaged Latino students meet or exceed standards at about the same rate as non-disadvantaged Asian and White students in SFUSD in both Math and ELA.
Here’s the picture for socioeconomically disadvantaged Black students but showing English Language Arts instead of Math. Mission Prep doesn’t have enough of those students to register a score (remember it’s >80% Latino) but, of those schools that do, the two KIPP schools do much better than any of the SFUSD schools with similar numbers of those students.
These results may explain why parents have been forgoing the small class sizes at Brown, Lick, and Everett.
High School Achievement
It’s practically impossible to evaluate individual high schools without being able to adjust for the standardized test scores of the incoming students. So I’m just going to mention a couple of eye-popping statistics relating to AP tests.
AP Calculus
I came across this statistic serendipitously when I was doing research for the Families for San Francisco report on SFUSD’s Math Pathway. It didn’t make it into that report but here it is:
In 2018-19 (which is the last year for which we have course enrollment data), more Black students took AP Calculus at one charter school, KIPP, than in all of SFUSD.
Having introduced this statistic, I am honor-bound to admit that I suspect this was a one-year anomaly that we won’t see repeated:
This was the first year KIPP offered AP Calculus so there may have been an artificial first-year bump.
Four was an all-time low for SFUSD. It was 20 the previous year2 .
It’s possible that the new SFUSD Math sequence, which came into effect for the Class of 2018-19 caused some of the SFUSD decline because it forced most Black students who wanted to take Calculus to have to take a one-year compression course combining elements of Algebra II and Precalculus instead of taking those courses over two full years. But this explanation only goes so far - the KIPP students took a similar compression course.
We don’t have information on the number who actually passed the exam, just those who were registered for the class which is what the graph shows. It would not surprise me if few or none of the KIPP students actually passed the AP test:
Across all subjects, KIPP students take a lot of AP tests (more per student than at any SFUSD school except Lowell) but pass them at a low rate (lower than at any SFUSD school except Marshall).
Compared to AP students in the other subjects, the AP Calculus students had poor preparation because they had to take the aforementioned compression class the previous year instead of Precalculus.
AP Course Enrollments
So much for Calculus. What about all the other AP classes? Do students in charter schools take more or fewer AP classes than their SFUSD counterparts?
I calculated the average number of AP classes by race/ethnicity in SFUSD and in San Francisco charter high schools. For each group, I divided the total number of AP course enrollments by the total number of 12th graders in that group. To limit the comparison to regular schools, I excluded the continuation high schools, SF International, and Independence High from the SFUSD count and excluded Five Keys from the charter count. I used 2018-19 enrollment numbers because those are the most recent available.
In the charter high schools, there were 434 enrollments in AP courses by Latino students and only 172 Latino 12th graders giving an average of 2.52 enrollments per 12th grader. Of course many AP courses are taken by 10th and 11th graders so this average does not mean “Latino students enroll in an average of 2.52 AP classes in 12th grade.” It’s something closer to “Latino students enroll in an average of 2.52 AP classes in high school”. By comparison, in the regular SFUSD high schools, there were 873 Latino 12th graders and 1419 AP course enrollments by Latino students, giving an average of 1.23 AP classes per 12th grader.
In SFUSD high schools, Black and Latino students take far fewer AP courses than Asian and White students (0.74 and 1.23 compared to 3.19 and 3.73 respectively).
But charter schools close that gap considerably. Latino students in charter schools take an average of 2.52 AP classes, more than twice the 1.23 taken by their counterparts in SFUSD. In fact, since 2.52 is closer to 3.19 than it is to 1.23, we can say that, in this one respect, the performance of Latino students in charter schools is closer to that of Asian students in SFUSD than it is to that of Latino students in SFUSD.
Black students still lag the other groups but their 1.86 average at charter schools is more than 2.5 times their 0.74 average at SFUSD.3
Of course, this does not prove that students learn more at the charter schools. Other explanations are possible. It may be that charter schools attract disproportionate numbers of ambitious, hard working students who would do just as well in SFUSD. Or it may be that the charter schools are pushing students to enroll in AP classes that they won’t be able to pass. But it could just as easily be that high expectations at charter schools push the students to achieve more than they would in SFUSD.
Conclusion
Charter schools are attracting increasing numbers of Black and Latino students. Evidence from standardized test scores and AP participation rates, while not conclusive, indicates that Black and Latino students do much better in some charter schools than in SFUSD.
SFUSD’s reaction has been extremely defensive. They tried to force KIPP Bayview Elementary (which serves the highest proportion of FRPM-eligible students of any school in the city) to move to Treasure Island. Famously, a school reopening consultant wasn’t hired because she’d worked for a charter school. It seems they are prioritizing the interests of the school district as a legal entity over those of the students.
People sometimes have a visceral reaction against KIPP because it has a very particular approach and because it’s not based here. But Mission Prep is based here and has managed to achieve SFUSD’s holy grail of disadvantaged Latino students to excel at the same rate as non-disadvantaged Asian and White students. Instead of dismissing all good charter school results as “cherry picking”, SFUSD should learn from them. Perhaps the district should instruct its teachers to follow Mission Prep’s curriculum.
I’m using the course enrollment data from CDE because that enables comparison with the charter schools. SFUSD has itself published slightly different numbers for the 2018-19 and 2017-18 (3 and 17 instead of 4 and 20). If you’re curious, see the Families for San Francisco report for a more detailed discussion of these numerical discrepancies.
No charter school average is shown for Asian and White students because only 1% of Asian 12th graders and 4% of White 12th graders attend charter schools compared to 29% of Latino 12th graders and 38% of Black 12th graders.
I wonder if the racial disparity would be different if the were a charter school that catered to an Asian population, for example a Chinese immersion charter.