Which Elementary Schools Would You Close?

And why the elementary decision is even more complex than the others

There are several considerations that make the elementary school process even more complex than that for middle and high schools.

1. Transitional kindergarten (TK) will become available to all four year olds starting in 2025-26. That should cause an increase of over 600¹ in TK enrollment. It might be even more. Today, only a fraction of eligible students sign up and the district has set up TK classrooms in schools that have empty classrooms which are, by definition, not the most popular schools. If the district were to offer TK in every school, and guarantee that every TK student could proceed to K-5 in the same school, this would greatly increase demand for TK slots and might have the effect of bringing all the angst and tension of the kindergarten enrollment lottery forward one year².

2. The district voted a few years ago to adopt a new zone-based kindergarten assignment process. In classic San Francisco fashion, the entire debate was conducted in the abstract on the basis of principles and values. No set of candidate zones was ever published, presumably because every possible zone configuration is going to annoy somebody. If some elementary schools close, will the district just modify attendance areas to account for the school closures or will it roll out the new elementary school assignment zones at the same time as all the closure announcements?

3. Offering language and general education programs in the same school leads to wasted space in grades 4 and 5 when class sizes increase. The district has signaled that it wants to consolidate language programs in fewer schools³. It’s unclear how aggressively they will be move towards complete separation of language and GE programs. A key difference between biliteracy and immersion programs is that biliteracy programs require the incoming kindergartner to be fluent in the target language whereas immersion programs are designed to have a mix of English only, target language only, and bilingual students. A Spanish immersion program will thus contain a mix of ethnicities whereas a Spanish bilingual program will be nearly entirely Latino. In San Francisco, there are three schools (i.e. BVHM, Marshall, and Huerta) that are 100% Spanish immersion and two that are 100% Cantonese immersion. These five schools have no general education programs. In contrast, there are zero schools that are 100% biliteracy program, presumably because this would lead to schools that are nearly 100% Latino or 100% Asian. The district has historically been very conscious of the racial composition of its schools (it was often compelled to be so by the courts). It has preferred to have the paper integration that comes from having biliteracy and general education programs under the same roof even if the children in the two programs never take class together. Is the district ready to accept the self-segregation that would happen with 100% biliteracy schools?

I don’t know where the district is going to come down on these issues but the answers will have a significant influence on which schools close.

Enrollment Today

Here’s a map that shows the location and enrollment of the schools that serve SFUSD’s elementary school students.

How Much Capacity Does SFUSD Have?

For middle and high schools, I estimated a school’s capacity to be the higher of:

• the highest number of students enrolled in any year since 2000

• a quick-and-dirty SFUSD estimate based on the number of classrooms in each school⁴.

Thanks to Philippe Marchand’s public records request, I now have a third capacity estimate for each school, which SFUSD calls a “program capacity”. SFUSD has defined five different “program capacities” for its elementary schools.

I’m guessing that 418 represents one TK class of 22, and then 66 students per grade from then on. Similarly, 572 comes from an additional 22 students in each grade and 726 from an additional 44 students in each grade.

A few things should be noted about these capacities.

• For many schools, these program capacities understate their true capacities. Monroe, Stevenson, West Portal, and Lakeshore exceeded their supposed capacities as recently as last year. In Monroe’s case, it had 522 students even though its capacity is supposedly only 418.

• It’s unclear why some schools were given the capacities they received. Feinstein and Sanchez were given capacities of 418 but, according to the facilities master plan, they both have more classrooms than Alamo, Longfellow, and West Portal which were given capacities of 572. Lakeshore had 587 students one year but its capacity was set at 418 instead of 572.

• On the other hand, the Chinese Immersion School at De Avila was given a program capacity of 572, implying four classes per grade level, even though it has enrolled only three kindergarten classes every year of its existence despite being a popular school among kindergarten applicants. Space has suddenly been found there where none existed before.

• The K-8 schools were given two separate program capacity estimates, one as elementary schools and one as K-8s. The K-8 capacities were set at 561 (SF Community), 616 (Lilienthal, Yu, BVHM, Lawton, and Rooftop), and 770 (Revere and Carmichael).

I estimated the actual capacity for an elementary school to be the highest of the three available sources (actual enrollment; classroom-based estimate; program capacity). Adding the actual capacity estimates for each school gives a total of 34,008. If we add up just the program capacities (and ignore the evidence in the enrollment record), we get a total of 31,368. If we think of the program capacity as the capacity that SFUSD would like to fill, the difference between the actual capacity and the program capacity (i.e. 2,640) is capacity that has disappeared from the system.

How much space does SFUSD need?

SFUSD has said its goal is that schools should be filled to between 90% and 95% of capacity. SFUSD had 23,258 TK-5 students in 2023-24. That implies they need capacity for between 24,500 and 25,842 students. Given the estimates of program capacity at each school, this means they have about 5,500 more seats than they need. (They’re going to need to accommodate 600+ extra Transitional Kindergarten students in a couple of years but they’re also going to be opening the new Mission Bay school with its 550 student capacity. Let’s call that a wash for now.)

Elementary schools currently enroll about 72% of their program capacities. The utilization varies widely across the district.

Schools in the Sunset are generally near full capacity. Ulloa, for example, is supposed to have 88 students per grade and it has an average of 87. Its program capacity is “only” 91% because the program capacity assumes it has 44 TK spaces even though it has zero TK students today.

It is interesting to compare this map with the one that shows where students actually live. To my eye, the under-enrolled schools fall into two groups:

• north and east from the Castro are schools in regions that just don’t produce a lot of SFUSD students. There is far more school capacity in this region than there are students. Reasonable people might differ on which individual schools should close but nearly all would agree that, if any schools have to close, at least some should be from this region.

• from Ingleside east to Bayview, there are under-enrolled schools in regions that produce tons of SFUSD students. In fact, there are far more students than there are school spaces and the schools are still half-empty.

What is the appropriate policy response? This is where genuine philosophical differences arise. The quickest and surest way to fill all the schools in the south of the city would be to switch to a neighborhood school system. If families had no choice but to send their children to the local school, all the schools in the south of the city would be full and schools elsewhere in the city would have to close due to low enrollment. This solution is not on the agenda because people really love the ability to choose which school to send their children to. Few of the people who oppose school closures in this part of the city would support switching to a residence-based enrollment system to guarantee their future.

The district is thus faced with an unhappy choice. Does it close some local under-enrolled schools and accept that families are going to continue to send their children outside those areas for school? Does it close schools elsewhere so that parents have no alternative but to send their children to the local under-enrolled schools? Does it design its elementary school enrollment zones so that local families are forced to fill the local schools? Does it reimplement some sort of diversity lottery so that the under-enrolled schools are filled with children who live in other parts of the city? Do you keep the schools open, even if they are under-enrolled, as a penance for past sins?⁵ My biggest complaint about the resource alignment process is that public input was never sought on these big-picture tradeoffs. Instead, nearly all the public involvement was focused on figuring out how to best weight a bunch of poorly-defined quantitative criteria.

Let’s look at what those criteria say.

Decision Criteria

The data SFUSD made available in response to Philippe Marchand’s public records request was sufficient for me to estimate the values for most of the criteria that comprise each school’s Composite Score. The most important criterion is School Access, which we looked at in the last post. The second most important criterion is Program Access, which is worth 18.3% of a school’s Composite Score. It is defined as the “percentage of students in each school identified as low socio-economic status or participating in Language programs, Special Education programs, AVID or Career Technical Education and Pathway programs”⁶. The more of these students a school has, the less likely it is to be closed. To calculate the Composite Score, this percentage is expressed as a Z Score which is the number of standard deviations above or below the mean.

Here’s a map showing the score of each school. At a high level, this probably looks like what the proponents of this criterion expected: schools on the east side score better than schools on the west side and the school with the lowest score is Grattan which has the fewest disadvantaged students. However, it’s worth noting a few oddities.

Notice that there are lots of schools in pale blue but none in deep blue. The program access definition includes so many different categories of student that the average across all schools is 75% with a standard deviation of 21%. This means the criterion cannot provide much of a boost to a school because it is impossible to be more than 1.2 standard deviations above the mean.

The second thing to notice is that the inclusion of language program participation in this criterion produces some results that are hard to defend. The schools with the highest scores are the five where 100% of students are in language programs. Alice Fong Yu has the highest possible Program Access score despite having the fewest special education students, and way below average numbers of disadvantaged students and English learners⁷. Meanwhile, schools like Alamo and Stevenson have more English learners, more special education students, and more disadvantaged students than Alice Fong Yu but, because those students are in general education programs, their schools are heavily penalized by the formula.

I attempted to replicate the entire Composite Score calculation but was stymied.

• SFUSD didn’t make available the data for the Historical Inequities criterion.

• The School Access criterion requires an undefined adjustment for the student density of the zip code.

• Several of the Excellence and Effective Use of Resources criteria are defined so loosely that reasonable people could come up with different numbers using the same data, depending on which years’ data they choose to include, how they adjust for missing datapoints etc.

I did calculate a partial composite score that included all bar the School Access and Historical Inequities criteria. Since the two missing categories account for nearly one-third of the overall composite score, I don’t think there’s much point in putting out a ranking of schools by this partial composite score. However, if two schools are in the same part of the city and draw similar students, they should get similar scores for the missing Historical Inequities criterion. Hence, I will sometimes assert that a school will get a higher or lower composite school than another school in the same part of the city. For example, I’ll compare schools in the northeast with other schools in the northeast and schools in Bayview with other schools in Bayview but I won’t compare schools in the northeast with schools in Bayview because of the possibility that the missing scores will change the order.

Which schools to close?

I’m going to come up with a list that achieves the needed capacity reduction based on publicly available data. My guiding principle will be that the number of seats that will be left after the closures in each part of the city will exceed the number of students who currently attend schools in that part of the city. In other words, no one will be forced to attend a school in a different part of the city than their current school. Since I can’t fully reproduce the composite score, my list should not be considered a prediction for what SFUSD will do. Nor is it even really a suggestion for what SFUSD should do. I don’t pretend to know enough about individual schools to make that call. My goal is to show that any conscientious effort to identify schools for closure is going to have to balance a lot of issues.

There are 71 schools with elementary school students. That’s too many for any reader to keep track of. To make it easier to follow along, I’ve therefore divided the city up into eight different regions and will consider each region’s schools in turn.

North-East

Northeast is the area east of VanNess and north of Market St. It contains eight schools: Lau, Chin, Garfield, Parker, Yick Wo, Spring Valley, Redding, and Tenderloin. Collectively, 891 (29%) of the 3,047 seats are empty.

Lau is one of the largest schools in the city; 83% of its students are disadvantaged, and 57% are English learners. It has also played a central part in the history of public education in the city. It obviously stays.

Tenderloin also stays. As we saw in the last post, its zip code has the highest student density in the city and it’s the only school in that zip code. Assuming it does stay open, the challenge of how to drive enrollment up will remain. No one who doesn’t live there is clamoring to send their children to Tenderloin.

Chin, Garfield, Parker, and Yick Wo are all small capacity (i.e. max 297) schools that are close together. Parker has the lowest enrollment. Chin, Garfield, and Yick Wo have all been recently renovated and have FCIs of 0.05 or less. Parker is the obvious choice to close. The only question is whether it’s the only one because Garfield and Yick Wo both have fewer than 200 students too. The other alternative is Spring Valley, which fills only 214 of its 418 seats. I’d opt for Spring Valley because it’s closer to Tenderloin so I think there’s more chance of its students transferring there.

The closures of Parker and Spring Valley would bring spare capacity in this area down to 8% from 29%.

North Central

North Central contains six schools. From North to South, they are Lilienthal, Sherman, SF Public Montessori, Cobb, Rosa Parks, Muir. As a group, 1,276 (45%) of their 2,849 seats are empty, and it’s only that high because Lilienthal is at 113% of its supposed capacity. This is also not an area where a lot of SFUSD students live so it’s vulnerable to closures.

Lilienthal stays. It’s one of the most popular schools in the district.

SF Public Montessori goes. This school was started in 2010 but never really achieved critical mass. It never exceeded 176 students.

Cobb is a tiny school with only 117 K-5 students but the district is going to be loathe to close it because it has the highest percentage of Black students of any of these schools. It’s also in the best physical condition of any school in this area with an FCI of 0.16 (which is still worse than three of the schools in the northeast).

Sherman is the only school in this area other than Lilienthal to be at more than 50% of capacity (it’s at 67%). Among elementary schools, it has the second-highest percentage of students with disabilities but still has a low score on the Program Access criterion because it a school that just offers General Education and it’s hurt by the definition of the Program Access criterion, which penalizes schools that just offer General Education.

Muir is a surprisingly big school with a capacity of 572 but it’s never exceeded 318 this century and currently has 226. You’d imagine that a school with a student body that is 85% Latino or Black would score really well on the Program Access criterion but, because the average was around 75%, it doesn’t score that highly relative to other schools.

Rosa Parks is massive. It has a reasonably healthy enrollment of 344 but that’s still less than half of its capacity of 726. Only two other schools have a capacity this high. Parks is home to one of the two Japanese programs in the city but its program is not as popular as Clarendon’s. If the district is looking to consolidate language programs, the one at Parks might shift to Clarendon, making Parks itself vulnerable.

I think Lilienthal is safe and SF Public Montessori is an obvious choice to close. That’s not enough. A case could be made for any of the others but I’m going to go for Parks, simply because it has the most spare capacity. Closing Parks and SF Public Montessori brings the spare capacity down from 45% to 8%.

Richmond

Richmond has six schools: Alamo, Argonne, Lafayette, McCoppin, Peabody, and Sutro. Between them, 458 (19%) of their 2,453 seats are empty.

Alamo, Argonne, and Lafayette have capacities of 418 or 572, making them too big to close without sending students outside the area. The second three each have capacities of 297, just the right size. To complicate matters, they are also the three in the best physical condition, each with FCIs between 0.09 and 0.11. However, Sutro is closest to Alamo, which has the most spare capacity, making it the one to close.

Closing Sutro brings spare capacity in the area down to 7% from 19%.

Mission/Potrero/Bernal

This area, bounded by Market, Dolores, and Cesar Chavez streets, contains 12 schools (Bryant, Buena Vista Horace Mann, Carmichael, Cesar Chavez, Flynn, Marshall, Moscone, Revere, Sanchez, Serra, Starr King, Webster). Not including the new Mission Bay elementary, which will add 550 new seats when it opens, 1,347 (26%) of its 5,269 seats are empty.

Unlike some other regions, no school in this area is doing terribly. Seven of the schools are at 80% or more of capacity. None has an enrollment below 240. All of them have language programs. It’s possible to make a decent case for and against most of them. That’s going to make eliminating the spare capacity even more controversial because every decision will seem capricious.

Until Mission Bay opens, there’s no local substitute for Carmichael. It’s also one of only two schools with a Filipino program. On the other hand, it may end up with the lowest Composite Score in this area. The district has talked about making its school closure decisions using a mix of art and science. If the composite score is the science, Carmichael is going to need the art to survive. For my purposes, it stays.

Starr King is an interesting case. Two thirds of its students are in the Mandarin immersion program. The remaining 111 are not enough to sustain a school by themselves so the fate of the school depends on whether the district decides to keep the immersion program there or combine it with Ortega’s. If Starr King closes, the GE students could be transferred to Bryant or Webster, provided GE programs remain in those schools.

If the resource alignment process had happened five years ago, Buena Vista Horace Mann would have vulnerable because it was in such a bad state and so many other schools are so close. With the remodel underway, it’s golden. That means the axe might fall on one of its near neighbors: Chavez and Moscone. They’re very similar schools serving the same neighborhood and they’ll end up with almost identical Composite Scores. My pick is Moscone because it is surrounded by other schools: there are five within 0.75 miles of Moscone. Bryant or Cesar Chavez or Flynn could easily be chosen instead of Moscone.

There are 12 schools in the district with Spanish biliteracy programs. Five of them (Bryant, Chavez, Moscone, Sanchez, and Serra) are in this region and two others (Muir and Glen Park) are just outside. If consolidation is to happen, Muir and Glen Park might lose their biliteracy programs because the general education programs at both are way more popular than the biliteracy programs. GE is also more popular than Spanish at Sanchez and Serra but it’s the other way around at Bryant, Chavez, and Moscone. Chavez and Moscone are right next to the each other so it wouldn’t make sense to have them both be 100% biliteracy. If Moscone is to close, maybe Chavez becomes 100% biliteracy and/or incorporates the newcomer students who I earlier displaced from the Mission Education Center.

The second school on my list from this region is Revere. It’s only at 52% of capacity, the lowest in this area. Displaced students can be accommodated at Flynn and Hillcrest, which are quite close.

Closing Moscone and Revere brings the spare capacity in this region down from 26% to 8%.

Central

This area stretches from Golden Gate Park down to Ocean Ave and encompasses 13 schools: New Traditions, Chinese Immersion, Grattan, McKinley, Milk, Alvarado, Rooftop, Clarendon, Miraloma, Sunnyside, Glen Park, Mission Education Center, Huerta. Between them, 1,014 (19%) of the 5,386 seats are empty.

Chinese Immersion and Huerta are both 100% immersion schools, one for Cantonese, one for Spanish. As such, they have golden tickets.

New Traditions, Grattan, Rooftop, McKinley, Milk, Miraloma, and Sunnyside are all pure general education schools with below average numbers of disadvantaged students. As such, they score very badly on the Program Access criterion but compensate with good to very good scores on the Excellence, Family Choice, and Student Enrollment criteria. All seven are at least 37% White.

In the last round of closures eighteen years ago, New Traditions was going to be merged with Grattan before getting a reprieve. It’s not possible to merge the two any more because both have been close to capacity for the last fifteen years. What changed? My guess is that the end of the diversity lottery allowed the schools to be filled with local parents. Now they are the only two majority White schools in the district. Both are in relatively poor condition with FCI scores of 0.38. New Traditions has only two classrooms of more than 600 sq. ft (the district’s benchmark) so it’s the more vulnerable of the two.

McKinley and Milk are even weaker. Both are at less than 60% of capacity. Normally, I’d suggest closing Milk, which has only 154 students, because it’s the smaller school but it somehow got prioritized for refurbishment so its FCI is an excellent 0.02. This is yet another example of the district prioritizing work on small under-enrolled schools instead of larger schools filled with students. McKinley then is my choice to close.

You might think that a newcomer⁸ school for Spanish-speaking students would be located in the Mission but the Mission Education Center, despite its name, is located on the Glen Park/Noe Valley border. Its numbers fluctuate with immigration but there were 65 last year in K-5. There was a comparable number in the new transitional kindergarten program but this was not restricted to newcomers. The other newcomer school, Lee, was recently folded in to Lau, a regular elementary school. One obvious change would be to close this school too and move the newcomer program to one of the schools in the Mission with biliteracy programs.

Closing McKinley and Mission EC would reduce the spare capacity from 19% to 9%.

Sunset

This area has 11 schools (Key, Sunset, Lawton, Stevenson, Jefferson, Yu, Ulloa, Lakeshore, Feinstein, West Portal, Sloat). Collectively, only 218 (4%) of their 5,214 seats are unfilled.

The district’s target is to operate buildings at 90%-95% capacity so with only 4% of seats being empty, no changes are needed. If a school has to be sacrificed to spread the pain around, Feinstein or Stevenson might be the vulnerable ones.

Excelsior

This area, from Ingleside east to Portola and Visitacion Valley, has 11 schools (Ortega, Sheridan, Longfellow, Guadalupe, SF Community, Monroe, Cleveland, Hillcrest, Taylor, El Dorado, Visitacion Valley). Taken together, 1,364 (27%) of their 4,972 seats are empty.

In the Mission, Central, and Sunset areas, the schools within each area had a lot in common. The schools in this area are very diverse. Six schools have Spanish programs (five biliteracy; one immersion); four have Chinese programs (three Cantonese biliteracy; one Mandarin immersion); one has a Filipino program. Three have multiple language programs. SF Community is the only one to have below average numbers of disadvantaged students. Ortega and Monroe are the only two to be above average in kindergarten applications.

Taylor and Monroe both have over 500 students and they are the only schools to offer Cantonese, Spanish, and General Education programs. I can’t imagine either school closing but I’ll also be surprised if all three programs remain at both schools.

Ortega and Longfellow have unique features. Ortega is one of two schools in the city with a Mandarin program and Longfellow is one of two with a Filipino program. I am going to assume those features keep them safe.

Let’s try to find some good candidates for closure. The most under-enrolled schools in the area are El Dorado (41% of capacity) and Sheridan (45% of capacity). El Dorado’s 112 K-5 students could easily fit into the spare capacity at its closest neighbor, Visitacion Valley, and there’s no prospect of Visitacion Valley getting students from anywhere else. However, El Dorado is in the best physical condition of any school in the district (FCI is 0.01). Yet another example of the district wasting its facilities dollars on small, under-enrolled schools where they get the least value for money. I suspect that whoever did the prioritization calculated that modernizing the buildings would increase the enrollment without realizing that there would still be massive over-capacity even if every school were pristine. There just aren’t enough kids to go around.

Meanwhile, Sheridan (169 K-5 students) has also had a lot of work done on it (FCI is only 0.11). Sheridan also does extremely well on the School Culture and Climate surveys and the Social-Emotional Development surveys and has a low teacher turnover rate. Its fate may be bound to that of Ortega, its nearest neighbor. If Ortega become a pure Mandarin-immersion school, then the General Education students could be shifted to nearby Sheridan.

Cleveland and SF Community are very close to Monroe. SF Community is by far the smallest K-8 program and has only 181 K-5 students. It’s a clear candidate for closure.

Cleveland has a lot in common with Guadalupe to its south and Hillcrest to the north-east. They all have a mix of GE and Spanish biliteracy programs and, consequently, high percentages of disadvantaged students and English Learners. Hillcrest has the most students of the three but it also has the most spare capacity because it is the biggest. Revere was on my list of schools to close in the Mission region so I have to keep Hillcrest open because they are neighbors. Similarly, if SF Community is on the list to close then Cleveland, its near neighbor, should stay open. That means the axe falls on Guadalupe. One challenge for Guadalupe is its location on the city’s southern border. Its nearest neighbor is a Daly City school, not an SFUSD school. (Longfellow is also closer to a different Daly City school than it is any SFUSD school.) Hence, Guadalupe can only attract students from the north and those students have to pass Monroe and Cleveland to get to it. How many will do that? However, my calculations show Guadalupe having the highest composite score of the three, so who knows what the district will decide.

Closing SF Community, Guadalupe, and El Dorado would still leave this region with 9% spare capacity.

Bayview

This region has 4 schools: Carver, Drew, Malcolm X, and Harte. They are at 23%, 33%, 25%, and 43% of capacity respectively. Combined, they have 636 students but space for 1,980, meaning that 68% of the capacity is unused.

It didn’t use to be like this. Back in 2000-01, Carver and Malcolm X both had 370+ students. Today they have barely 100 each. Drew and Harte have seen enrollment decline too, but not by as much.

Harte is probably safe. It is actually at its highest enrollment since 2009. If Candlestick Point ever gets redeveloped, Harte, as the closest school, should see a bump in enrollment from that. If either Visitacion Valley or El Dorado closes, it might get a bump from that too.

Today, Carver, Drew, and Malcolm X have a combined 389 students. Drew alone has a capacity for 572 and it was recently modernized, leaving it with an FCI of just 0.02. There is a natural temptation to just consolidate the three schools into Drew.

Carver is a plausible alternative. It will, I think, end up with a better composite score than Drew, has the further advantage of being almost equidistant between the other two, and has just enough (418 students) capacity for all the students. But it would be a tight fit.

Even if the numbers are unassailable, the optics of closing two schools in this neighborhood might dissuade the district from doing so. If only one is to close, it will be Malcolm X. The cost of the required maintenance there works out at $165,000 per student.

Even after closing Carver and Malcolm X, 44% of the capacity in this area would be empty.

Summary

My list had 14 schools on it from all over the city, except Sunset. The closures would reduce capacity in the elementary grades by 5,562, almost exactly hitting the target of 5,524. Given the uncertainty about future transitional kindergarten enrollment, I doubt that the district will move so dramatically.

With my list, 12% of students would be affected by the closures, including 8% of Asian and White students, 17% of Latino students, and 23% of Black students. Would that disparity be sufficient to sink the plan? Such disparities are inevitable if minimizing the number of affected students is a goal because Latino and Black students are over-represented in the most under-enrolled schools. The schools that are at less than 60% of capacity contain 5% of Asian students, 9% of White students, 19% of Latino students, and 43% of Black students. The schools that are at more than 90% of capacity contain 46% of Asian students, 44% of White students, 25% of Latino students, and 19% of Black students. It’s just hard to find under-enrolled schools with lots of Asian and White students. Of the 27 schools with the most Asian and White students, only one (Parker) is less than 65% full.

Next

The district has said that they will release the list of school closures on Wednesday, September 18. I’ll be out of town next week so my analysis will have to wait until the week after.

1. In 2023-24, TK was available to students who turn 5 between September 2nd and April 2nd and 890 students enrolled. When it’s available for all four-year-olds, the total number should be 890 * 12/7 = 1,525, implying an increase of 635.

2. It might also reduce the availability of early child care in the city. The argument is that preschoolers require less intensive care than infants. The older preschoolers effectively subsidize the cost of taking care of the younger kids. If the older preschoolers shift to TK classrooms, this will drive up the cost of preschool tuition.

3. Consolidating programs does not mean reducing the number of seats. To be clear, there is no need to reduce the number of seats at the kindergarten level because the district has already done that in the last few years. In 2023-24, the kindergarten spots for Spanish Immersion and Biliteracy programs were at 97% capacity.

4. SFUSD’s quick-and-dirty estimate was to count the number of classrooms over 600 sq. ft., assume that 85% of them are utilized and that they contain an average of 25 students per class (25 being roughly the weighted average 22 per class in K-3 and 33 per class in 4-5). I had ignored the significance of the “over 600 sq. ft” clause until I noticed that New Traditions had only two classrooms of that size which meant its capacity was officially 2 * 25 * 85% = 43. It must have at least ten classrooms under 600 sq. ft. because it has had as many as 260 students. I don’t know if other school buildings also have smaller classrooms or if New Traditions is an architectural anomaly.

5. Of course, the ideal outcome would be for parents to want to send their children to the local schools and for the parents who have to send their children to schools elsewhere in the city (because the local schools are full) to be clamoring for new local schools to be built. If the district knew how to accomplish that, it would have done so a long time ago.

6. Interestingly, the district’s data shows that English Learners were also counted as part of this definition even though they are not explicitly mentioned. I don’t have a problem with this. Given that this is supposed to be an equity-focused criterion, it makes more sense to include English learners than language program participants.

7. Counting the percentage of English learners makes K-8 schools look bad because many English learners, particularly Chinese-speaking learners, are reclassified as fluent by the end of 5th grade. A K-5 school that has the same percentage of English learners in kindergarten will show a higher overall percentage of English learners because the students will have moved on to middle school after they become fluent in English.

8. I must confess that I don’t understand what value a newcomer program adds on top of a biliteracy program.