Improving SFUSD's School Funding Formula

SFUSD's current system pours money into the smallest schools. It could be spent more efficiently.

There is a lot of public information about school district finance. If you want to know how much money a school district spent on textbooks, you can look it up. If you want to know how much money was allocated to each of the schools in a district, you can sort-of do that too (more about this in a future post!). What you can’t easily do is compare the process that one district uses with the process that another district uses. I’ve yet to find a district with a clear publicly accessible description of its process. In SFUSD’s case, there are some well-written documents but no visitor to sfusd.edu is ever going to find them. The relevant documents circulate among school site councils, PTSAs, and others with a stake in the process.

Today’s post is going to look at SFUSD’s funding formula, note some peculiar features of it, and identify directions in which it could be changed.

Multi-Tiered System of Supports

Some students need more help than others. Those needy students are not distributed evenly across the schools in the district. From those two relatively uncontroversial statements arises the challenge of how to allocate money fairly to schools to reflect the varying levels of need at the schools. This is not a question with a single right answer. Reasonable people will disagree on the desired allocations because they will disagree on what is fair. But any debate about the funding process needs at least to be based on an accurate understanding of the process’s strengths and weaknesses.

SFUSD calls its system a Multi-Tiered System of Supports (MTSS). The district provides a number of centrally-provided services which don’t have to be paid for out of the individual school budgets. All schools then receive a site budget to pay for all locally-provided services. The foundation for the school budget is a flat dollar amount per student. Schools with a greater need then receive more resources from the district.

“School sites are ranked relative to one another on their inputs, and the resulting tiers are used to differentiate resource allocation and additional site supports…Schools with more need receive additional resources and site support, thereby providing each and every child access to the same quality of education irrespective of which school they attend.”

The supports vary based on the type of school. At the elementary school level, here’s what a school gets at each tier.

• Tier One: 0.5 FTE (0.5 social worker)

• Tier Two: 2.5 FTE (1.0 social worker, 0.5 nurse, 1.0 Instructional Reform Facilitator or Literacy Coach)

• Tier Three: 4.5 FTE (1.0 social worker, 0.5 nurse, 1.0 ARTIF (Academic Response to Intervention Facilitator), 2.0 instructional supports (Assistant Principal, Instructional Reform Facilitator or Literacy Coach).

At the high school level, Tier Two schools get only 1 additional FTE and there’s no real difference between tiers two and three.

• Tier One: 2.0 FTE (1 Counselor, 1 Nurse)

• Tier Two: 3.0 FTE (1 Counselor, 1 Nurse, 1 Instructional Reform Facilitator or Literacy Coach)

• Tier Three: 3.0 FTE (1 Counselor, 1 Nurse, 1 instructional support)

Here’s a job description for an Academic Response to Intervention Facilitator, one for an Instructional Reform Facilitator, and another one for a Literacy Coach. My sense from these job descriptions is that the roles are more about improving the functioning of the school and the competence of the teachers and less about directly instructing students. The words “facilitator” and “coach” by themselves demonstrate that these jobs are not directly responsible for delivering results. They are not assistants working for the teacher but more experienced people working for the school district and principal to make the teachers more effective. Increased student learning may be the ultimate goal but it’s a second order effect. I have no idea whether these roles represent the best way to improve a school.

A note on the scale of these supports. Schools in richer neighborhoods with richer parents usually have well-functioning PTSAs that raise money for the school. A good PTSA in San Francisco might raise $400 per pupil. In a school of 300 kids, that’s $120,000 which is not much more than the cost of one FTE, including both salary and benefits (based on SFUSD’s average salary of around $85,000).

The inputs used to rank the schools are as follows:

Figure 1: Inputs used to rank schools in SFUSD’s Multiple Tiered System of Supports

My first reaction on seeing this list was surprise that the first two factors were the number and percentage of African American, Latino, and Samoan students. I thought that including race so explicitly in a funding decision was illegal but I’m not a lawyer so what do I know?

My second reaction was to note that only one of the fourteen factors is based on a count. Everything else is a percentage. I would have thought that the chances that a nurse would be needed is directly proportional to the number of kids in a school (perhaps with some adjustment for age). An elementary school with 600 kids should be three times as likely to need a nurse at any one time as an elementary school of 200 kids. I’d be alarmed if the need for a nurse is dependent in any significant way on the percentage of English learners or the teacher turnover rate. And yet, with the way that SFUSD’s funding formula works, all the elementary schools with 200 kids have access to a nurse but the largest elementary school in the city (Lau, whose final enrollment in 2021-22 was 738) does not.

Suppose that these supports work marvelously. How do you move up from tier three to tier two to tier one? Having 100% of the students meet the standard on the SBAC tests wouldn’t change anything: you’d still be in tier three. Reduced teacher turnover would help because four of the fourteen factors relate to that. But the only way to really move up the tiers is to change your student body because the remaining ten factors relate to that. Unless your school attracts a bunch of English-speaking White and Asian students, you’ll be in tier 3 forever. That’s a peculiar definition of success: “we won’t know if the school is improving unless a bunch of rich White folks deign to send their kids there”. I had initially thought of MTSS as extra funding for school improvement but that’s wrong. MTSS is not about improving educational outcomes because it doesn’t look at outcomes. It’s just a way for the district to demonstrate the sincerity of its commitment to those schools.

Measuring MTSS Effectiveness

Even though outcomes are not part of the MTSS funding decisions, it’s entirely possible that the increased funding enables improved outcomes, either increased teacher retention or improved student learning. If it works, tier 3 schools should close the gap on tier 2 schools which should close the gap on tier 1 schools. The MTSS system has been in place for a decade. That’s long enough to know whether it’s working but it would take a sophisticated analysis, requiring knowledge of which supports have been in place in which schools in which years and how the various learning outcome measures have changed in response after controlling for the demographic composition of the schools receiving support and those not receiving support. Only the district’s RPA (Research, Planning, and Assessment) group has the data and expertise to conduct such an analysis. Is the superintendent or the board willing to give them that task and would they be responsive to the results?

I did conduct one crude analysis, which was to see if MTSS support led to increased enrollment. It can be hard to tell from the outside if a school is working well but parents can feel it. If things are improving, they tell their friends and enrollment increases. I compared the enrollment figures for 2021-22 with the enrollment figures for 2011-12, the last year before MTSS was introduced. I found that enrollment in the elementary schools that are currently in Tier 3 is down 10%, compared to 7% for the district as a whole. Of the three tier 3 middle schools, one (Brown MS) wasn’t open in 2011-12 and the other two (Everett MS and Lick MS) had a 14% increase in enrollment. Meanwhile, the four high schools (O’Connell, Independence, Downtown, and Wells) showed a 24% decline. Let's call that inconclusive.¹

Alternatives to MTSS

Here are a couple of ways that MTSS could be restructured that might be better than the current system.

Alternative #1: Focus on the schools with the most struggling students

We could keep the supports the same but place schools in tiers according to how many actual students in need they have not according to predictions of need based on demographic inputs. For example, we could estimate the number of students in need at any school by counting all those who scored Standard Not Met² on the SBAC tests. Schools with the most Standard Not Met students would get the most help. The ten schools with the most struggling students will go into tier three. The schools ranked 11-36 will go into tier two and the remaining schools will be in tier one. This matches today’s tier distribution and keeps the cost of the program unchanged.

Figure 2 below shows how the schools would be distributed to tiers under the alternative system. The x-axis shows the total enrollment of the school and the y-axis shows the number of struggling students at the school (based on the percentage who scored Standard Not Met on the SBAC ELA or Math tests). Schools that are currently in tier 3 are shown in red; those that are in tier 2 are shown in orange; those that are in tier 1 are shown in blue. The current tier system is not very efficient at targeting aid to struggling students: some schools that have a lot of struggling students are in a lower tier than schools that have fewer struggling students. For example, Longfellow ES has around 227 struggling students, 6th most among all elementary schools, which would put it in tier 3 in the alternative system instead of tier 1 where it is now. On the flip side, Carver ES has about 68 struggling students, 44th among all schools, which would put it in tier 1 instead of tier 3 where it is now.

Figure 2: Using the number of struggling students at a school as the basis for assigning tiers produces different tier assignments than MTSS

Bessie Carmichael is the school with the highest number of students who score Standard Not Met, but it’s only in tier 2 today. It would move from tier 2 to tier 3, as would Flynn and Huerta. Hillcrest, Bryant and Muir would move from tier 3 to tier 2 while Malcolm X would also move to tier 1. Lakeshore, Rooftop, Lau, Glen Park, and Alvarado would move from tier 1 to tier 2, while Parker, Garfield, SF Montessori, and Cobb would move from tier 2 to tier 1. The schools that move down tiers often have higher rates of struggling students but they are so small that they don’t have enough such students to earn the support under the alternative algorithm.

The pairing of Carmichael and Longfellow is not a fluke. Those are the two schools with Filipino language programs and their students are both about 25% Filipino, a far higher percentage than any other schools in the city. The schools lose out under the current system, even though they have many struggling students, partly because they are large (575 and 475 students respectively) and partly because their students are the wrong race.

The advantage of this alternative is that it reaches more students in need. The current tier 3 schools contain about 1786 struggling students between them whereas the alternative tier 3 list contains 2260 struggling students. That’s a 26% increase in students helped at no increase in program costs. The disadvantage of this alternative is that it reduces support for students who are struggling in schools with low enrollment.

Alternative #2: Vary the dollars by child according to need

SFUSD’s funding formula gives a flat amount per student and then uses MTSS to give extra staff to certain schools. California’s Local Control Funding Formula takes a different approach. It gives 20% extra funding for every child who is socioeconomically disadvantaged i.e. eligible for free or reduced price meals, or an English learner, or in foster care. SFUSD could follow this approach. Schools that have more disadvantaged kids would get more money and it would be up to the school site councils to decide how to spend it.

This approach would be transparent, easy to administer, and more predictable for schools because they wouldn’t suddenly move from one tier to another. The biggest gainers would be schools like Gordon Lau in Chinatown which has a high percentage of disadvantaged students (94%) and is large. The biggest losers would be the super-small tier 3 schools, for whom the value of the additional 4 FTE is far greater than what they would receive under this model.

1. But last year was very weird because the schools were reopening after the pandemic. At many of the tier 3 schools, enrollment increased significantly between census day in early October (the official date for enrollment figures statewide) and the end of the year. For example, Cleveland ES went from 318 on census day to 351 at year end. I’ve no idea whether this growth is typical or just a consequence of people enrolling late after the pandemic reopening. If this is primarily due to the pandemic, then it would be better to use the final enrollment figures in any comparison, in which case the results of my “analysis” would change significantly. Tier 3 elementary schools would then be down 7%, exactly matching the district average. Everett and Lick would then be up 23% instead of 14% (I wonder if the increased enrollment played a role in the much publicized discipline and violence problems that plagued Everett last year) and the high schools would be up 18% instead of down 24%. That’s not a misprint: Independence HS had 265 students in 2011-12 and only 137 students on census day but its final enrollment count was 360. Downtown and Wells, which are both continuation high schools, also saw their enrollments double after census day.

2. Standard Not Met is the lowest of the four categories, the others being Standard Nearly Met, Standard Met, and Standard Exceeded. In the 2022 SBACs, one quarter of SFUSD’s students were Standard Not Met in ELA and one third in Math. Of the Standard Not Met students, just over half were Latino, about 15% Asian, 11.5% Black, and 11.5% Two or more. Geographically, they are concentrated in the schools in the south and east of the city. The SBAC tests are not taken by all grades. To estimate the number of struggling students in a school, I did the obvious thing and multiplied the total enrollment by the percentage of those from the school who scored Standard Not Met.