Reader mail: questions on student assignment

Updated to reflect comments from Vicki of PPS-SF (see comments section below). Also I mistakenly put the low test score area (CTIP 1) preference ahead of the SFUSD preschool preference so I have fixed that.

I received these questions in an email tonight and thought I might as well post them for everyone. I hope the answers are helpful.

Can you confirm that we still need to put our neighborhood school as our first choice if we want it?    NO. You do not have to put your attendance area school first if you want it. The district’s algorithm will work best for families if they rank their choices in order of their true preference. Let’s say you live in the attendance area for School C, but you would prefer School A and then School B ahead of School C. If you rank your choices truthfully — A, then B, then C — the algorithm will attempt to place you in Schools A and B before School C.  Here are the order of preferences:

  1. Siblings of current School C students;
  2. Any students currently enrolled in an SFUSD preschool located in School C’s attendance area who list School C as their first choice;
  3. Any CTIP 1 residents (people living in census tracts with the lowest test scores) requesting School C as their first choice;
  4. Residents of the School C attendance area;
  5. All other applicants listing School C as space permits.

If we put our neighborhood school (let’s say West Portal) as our second choice, and a non-West Portal neighborhood family puts West Portal as their first choice, they would get in before us.  Is this correct?  i.e., the process would run through everyone’s first choice and try to give as many people their first choice before moving on to considering everyone’s second choice?  NO.  If you live in the school’s attendance area, you would have preference above any other family who does not also possess a higher preference (e.g., sibling, CTIP, preschool) even if that family listed the school higher than you did.  
How is priority in the waitlist process determined – e.g., if you get 0 of out of 7, you are at the top of the waitlist; if you get your 7th choice, you are next on the waitlist; if you get your 6th choice, you are next, so on and so forth?  Is there any write-up explaining the waitlist process?  In the new system, being on the waitlist doesn’t give you any priority — families will no longer be “rewarded” for listing seven choices; nor will families who list fewer than seven be penalized if they don’t receive one of their choices. Families will be allowed to list an unlimited number of schools if they wish. I don’t think it’s been determined how many additional “runs” of the wait lists will be necessary, but I’ve been told that there will be no restrictions on what families request on their wait lists. Let’s say you list Schools A, B, and C and don’t receive any of them — you can continue in subsequent wait pool runs to list your same choices and the assignment algorithm is supposed to work the same way each time, awarding seats based on the order of preferences listed above. 
 Do neighborhood families get a preference for their neighborhood school in the waitlist process?  Assuming again that West Portal is our neighborhood school, would a non-West Portal neighborhood family who got their 6th choice be higher in the waitlist for West Portal than us if we got our 2nd choice but not our first choice of West Portal.  Or would we get preference in the waitlist because we are a West Portal family (even though we got our 2nd choice)?  My understanding is no. See above — subsequent wait list runs should consistently apply the order of preferences listed above each time. 
Related to the above, if you did not get your neighborhood school, does it give you higher priority in the waitlist — and does it matter if you put your neighborhood school as one of your choices or not?   The only things that determine preferences are the factors listed above. Based on my understanding, each of those preferences should be applied consistently in each round of the assignment process.  If you do not list your attendance school at all, you could still be offered an assignment to that school IF each of the following conditions are met:

  • You were not assigned to any of the choices you listed; and
  • Everyone who listed your attendance area school has been assigned either to that school or a higher choice; and
  • Spaces remain in your attendance area school.

Does it make a difference if your waitlisted school is one of your initial 7 choices vs. being a totally different school?  Again, each run of wait lists will not be dependent upon what you listed in previous runs, and you will not be limited to a certain number of choices (nor will you be rewarded for listing more, rather than fewer, choices). There will be no more “wait pool” or “amended choices.” If you don’t want what you got in the first run, you can continue to request different choices, in any order and in any amount, until you get something you like better.  Except in cases of oversubscribed attendance areas (which the staff believes will not be all that common), your “worst case scenario” should be your attendance area school — unless you don’t list it. Even then, you may be offered your attendance area school if you can’t be assigned to any of the schools on your list and your attendance area school is not full.

Advertisements

36 responses to “Reader mail: questions on student assignment

  1. I’ve read several times that children who attend a SFUSD preschool in the attendance area get very high preference but I can’t seem to find a list of those preschools nor the process to apply, eligibility requirements, etc. (The SFUSD website needs some serious help, by the way!) That would seem an easy route to getting our son into a school near us but I’m fairly sure there’s something I’m missing. Can you enlighten me?

  2. @Nervous Nelly, the scenario you describe is possible, but only if every single family in your attendance area requests your attendance area school and is not assigned to a higher choice.

  3. Hi Rachel,
    I just thought of an application scenario that I haven’t heard addressed – and was hoping you could shed some light..
    Is it possible for me to *not* be given a seat at my neighborhood school (let’s say I want one – and I put it down first) – and to also *not* have the dense population tie-breaker advantage?? Example: say my neighborhood school has 40 seats, and there are 30 neighbors applying for kindergarten (my neighborhood is not too dense, right?) But once you place siblings (who may not be in the “neighborhood”), and CTIP1 (who are not in the neighborhood)- you only have 20 seats left to work with. Now say I am not among the random 20 neighbors who gets a seat. Am I just out of luck in musical chairs? Or does the assignment system account for these earlier tie-breakers when determining if I’m in a dense population area?? Thanks.

  4. @Bryan, I’m sorry, your question is making my head hurt. Maybe it’s the heat.:-) In any event, the way I read the district’s FAQ, the densely populated tiebreaker would be issued to residents of attendance areas where the total number of applicants, to ANY list of schools (not just the attendance area school), exceeds the number of seats available in the attendance area school. Let’s say you live in the attendance area for school A, which has a capacity of 66 incoming kindergarteners. If 67 people who live in that attendance area submit applications for kindergarten, any kindergarten, that attendance area would then be classified as “densely populated.” In other words, the densely populated tiebreaker is calculated by the number of applications originating from the attendance area and has less to do with the number of people who want to attend the attendance area school.

  5. Hi Rachel. My questions about the dense population tie breaker was never answered. In reading this blog and the SF K Files, it looks like there is a lot of confusion about it so I’m posting it again:

    I’m still confused about the Dense Population Tiebreakder, especially after your 9:11pm e-mail: “If the number of students who submit applications and live in a particular school’s attendance area is greater than the number of seats available in that attendance area, then all students who live in that attendance area will get the dense-population tie-breaker. ”
    What does it mean by “number of seats available.” Does it mean number of seats available in total, or number of seats available AFTER sibling preference and CTIP1? If it means number of seats available in total, then there is a signficiant possibility that a neighborhood school will not be considered oversubscribed before the preferences, but many neighborhood applicants could be shut out of their school because of siblings and CTIP1 and still not get the Dense Population Tiebreaker. Am I understanding this correctly?

  6. Rachel,

    As you suggested, I looked at the demographic analysis of school age children in SF that was carried out by LGDRI. One piece of data that looked weird to me was the number of K-5 children in Pacific Heights (43), Presidio Heights (75) and Marina/Cow Hollow (96). Those numbers seem way too low. I have to imagine those don’t include K-5 students who went to private schools?

  7. @Leslie – it’s impossible to know exactly how many children are in any particular neighborhood, but we relied on our demographers’ — Lapkoff and Gobalet — projections based on census data. These projections are considered to be fairly accurate and are used by a number of different agencies for planning purposes. Over time we might well see more families with young children *moving* to a particular attendance area and so our projections don’t take into account future changes in behavior, but for year one there’s no reason to assume they’re not reasonably accurate.

  8. @Wayne – the system is “locked” for this year (meaning, no further changes) but if there are unexpected patterns like the one you suggest then I think we would certainly consider changes for subsequent years.

  9. Rachel said:
    “Let’s say that CTIP 1 requests take another 5 percent of seats (an assumption I find questionable based on the statistics noted above, but for the sake of argument, let’s make it)”

    Hi Rachel,

    Thanks for the additional info. My concern is the assumption that the ctip1 enrollment will match the current ctip1 enrollment level rather than the Ctip1 requests. If I am understanding the new student assingment system correctly: If 30 Ctip1 students in the whole city pick Clarendon as their first choice then it’s possible that not a single non-sibling neighborhood student would get in. This is a possibility as Clarendon’s is one of the most requested schools city-wide.

    Doesn’t this seem to be a flaw in the algorithm which could be corrected? If the school board is so confident that the enrollment will have a reasonable portion of neighborhood students, why not go ahead and reserve some portion of the seats for neighborhood students.

    Thanks
    Wayne

  10. Rachel,

    Thanks for graciously answering my question even though I ranted and expressed my frustration. The concern I have now is whether SFUSD factored in the possibility that for the attendance areas with good schools, the parents who live in those areas and never considered public schools under the old system may now be willing to send their kids to their area public school. I guess we will have to see what happens.

  11. An additional answer for @Wayne, who is concerned about his child’s chances of getting into Clarendon, his attendance area school — I wrote this as a response to a private email I received from a parent enrolling for the 2011-12 school year and wanted to reprint it here for others who might be interested in the same question:
    Honestly, if I were you I wouldn’t be too concerned about getting into Clarendon this time around if you are sure you live in the attendance area (I mean, you’ve checked the maps that were approved on Sept. 28 — Clarendon’s attendance area was changed with a sliver closer to Grattan reassigned to Grattan instead of Clarendon). An updated map hasn’t been drawn but as long as you are west of Woodland Ave. you should be still be in Clarendon’s attendance area.

    I can’t absolutely guarantee you that you will get a Clarendon spot under the new system but your chances are way, way better than under the old system. Why do I think that? There is no SFUSD pre-K in Clarendon’s attendance area. Second, there will be very little busing from CTIP 1 areas to Clarendon and even under the old system very few families from those census tracts requested Clarendon — only five percent of the the total enrollment of the school lives in a CTIP 1 census tract. It would be a huge shift — and one that would immediately draw interest and scrutiny from the district and the Board — for this percentage to change in one year.

    Secondly, all of our data indicates that the number of families in the Clarendon attendance area is well below the capacity of the school. There are 44 seats in Clarendon’s general ed strand (the one for which attendance area residents have priority). Last year there were 17 siblings who requested placement in that track, leaving 27 openings for new applicants. Let’s say that CTIP 1 requests take another 5 percent of seats (an assumption I find questionable based on the statistics noted above, but for the sake of argument, let’s make it) — that would represent perhaps 2 additional seats, leaving 25 for attendance area residents and everyone else. Are there 25 children between the ages of 4 and 5 in your area? I would say — based on my knowledge of the rough area designated as Clarendon’s attendance area — probably not.

  12. One more thing for @Wai — check out the district’s FAQ on the new assignment system, which I have relied upon for the answers to many of the above questions.

  13. @Wai – I know you’ve reviewed the testimony from Muriel Niederle and Clayton Featherstone where they described how the algorithm would work. Basically, if we use “rankings” to weight towards outcomes, we encourage people to strategize on how to get the outcome they want. Al Roth talked a little bit about this on his blog post announcing that SFUSD would be developing its algorithm internally. We achieve the promised result — getting everyone his or her highest choice possible — by the “transfers” round after people are assigned.

    As far as making the algorithm open source, I am talking to people internally about doing this and in the absence of any compelling arguments AGAINST making the algorithm public I think we should go ahead and do so. Stay tuned about that. I regret that I can’t say more about the district’s decision to go “in-house” on the development of the algorithm but I am prohibited under the Brown Act from disclosing the Board’s discussion with staff on this decision. I can only say that this discussion was substantive and the implications of the district’s decision were extensively considered.

  14. @Lindsey —
    1)yes, you have to live in the attendance area containing the preschool AND the school you are requesting. So, for example: the Argonne pre-K is in the same attendance area for Argonne Elementary. If you live in the Argonne attendance area AND attend Argonne Pre-K, then you receive the pre-K preference for Argonne only.
    2)I would have to say probably not. The district’s preschool program (called the Child Development Program) runs at a significant deficit, which has been up until now acceptable to the Board because we believe a dollar invested in preschool represents two dollars saved in remediation later on. But maintaining current offerings is very different than expanding offerings in a time of serious budget pressures.
    3)I don’t know the answer to that, but you could find out by contacting our Child Development Program.
    I know it’s extremely difficult for middle-income families to find affordable preschool options. I remember the incredulity I felt when I was told I’d better get my 3-month-old daughter on a waiting list for preschool!

  15. @Leslie – I think that assuming oversubscribed=densely populated is a big mistake. From every piece of data the Board considered (much of which is located here for your perusal) we got the strong impression that the only schools that don’t have capacity to absorb the students who live in the new attendance area for those schools are located in the southeast corner of the city. But you’re right that past performance doesn’t always equal future results, so this is something we’ll be monitoring closely in the next year.

    Anyway, it’s impossible for me to KNOW how it will all work out for people in your situation, but based on what I have been told by staff, I don’t think you should presume that you have “last pick” for wherever it is you want to go. I think the system is far more fluid and flexible than that, but as you say, I guess we’ll see in this coming year.

  16. Q: If we put our neighborhood school (let’s say West Portal) as our second choice, and a non-West Portal neighborhood family puts West Portal as their first choice, they would get in before us. Is this correct? i.e., the process would run through everyone’s first choice and try to give as many people their first choice before moving on to considering everyone’s second choice?

    A: NO. If you live in the school’s attendance area, you would have preference above any other family who does not also possess a higher preference (e.g., sibling, CTIP, preschool) even if that family listed the school higher than you did.

    How is this accomplished? The only rules I can see from the official material is the 5 rules you have cited. It says nothing about how the each applicant’s ranked preference list is going to be used in the enrollment process. Is there other unpublished information that lead you to this conclusion?

    In general I like the direction the new enrollment system is going. However it concerns me that there is a big void in the understanding of what this new system really is. Imagine the election office announce in the coming election we are going to adopt a great new system called “instant runoff”, which ensures we will elected someone with majority votes. Sounds great! However I need to know how? What steps are involved so that we will always elect some one with majority vote? We can’t just leave it to the election office to do it right. We need to know how. And for SFUSD’s enrollment, I don’t know how they do it.

    It concerns me more when I’ve learned that SFUSD is going to implement the new software using their in house IT. With all due respect, I have concern that SFUSD IT will do it correctly. Not unless they publish the algorithm or open source their implementation so that we can see what’s really going on. One small mistake can lead to incorrect result. Some angry parent will come back and ask why they don’t get the choice they entitle to even if they apply it in the right way Rachel Norton have said here.

  17. Hi Rachel-
    Question/ clarification about the sfusd pre-school priority factor.

    1)Do you also have to live in the pre-school’s attendance zone AND attend the Pre-school to get the sfusd Pre-school preference factor?

    2)Does sfusd plan on expanding Pre-school capacity in the next few years to accommodate likely increased interest in these facilities?

    3) How many tuition-based (non-subsidized) spots are currently available at sfusd Pre-schools per year?

    It’s tough for even middle class folks to find affordable full-day Pre-school options in the city without signing kids up utero 😉

  18. Rachel,

    I find the latest student assignment policy to be utterly unfair to middle class families who live in CTIP2 attendance areas with undersubscribed or underperforming schools (whatever you want to call it) who want out of their attendance area schools.

    It’s frustrating that we are the lowest in priority after CTIP1 families (lowest socioeconomic status) and then overflows from high demand attendance areas (highest socioeconomic status) for out of attendance area schools. As if I wasn’t annoyed already, I just found out from somebody else on another site that we would be lowest priority in city-wides too. Is that right? If that’s the case, the policy is even more unfair – the poorest and the richest get their picks and then the CTIP2s who want out of their undersubscribed schools are left with the scraps. Granted, I’m equating oversubscribed with densely populated but I have no choice considering the lack of history.

  19. I’m still really curious to know what happens if you move
    (within the city) after your child has already been assigned to a “neighborhood” school, assuming the move was for a legit reason, of course. We are likely moving to a bigger space next summer after our child will have an assignment and I’m wondering if we’re really now tied into only moving within our school boundaries. Most suburban school districts easily allow inter district transfers once the child is already at a school as long as the child lives within the district boundaries and I’m wondering what SFUSD’s plan is for this.

  20. Hi Rachel, I read your previous response but I still think that models should be run before the school enrollment process, particularly for high demand. I’m not personally affected this year (or for several more years) but I’d hate to see none of my neighbors get a local school assignment.

    Basically, I think the whole system should be simulated based on last years preferences and we should look for unintended consequences.

    The current Ctip1 enrollment of Clarendon is not a good metric because that was under the previous assignment system. I’m assuming 2 things about the new assignment system:

    1) Ctip1 applicants who choose Clarendon first will get in before neighborhood students.

    2) Most Neighborhood student will now pick Clarendon first ( even if they didn’t last year, as they had much longer perceived odds.

    Wayne

  21. I’m still confused about the Dense Population Tiebreakder, especially after your 9:11pm e-mail: “If the number of students who submit applications and live in a particular school’s attendance area is greater than the number of seats available in that attendance area, then all students who live in that attendance area will get the dense-population tie-breaker. ”
    What does it mean by “number of seats available.” Does it mean number of seats available in total, or number of seats available AFTER sibling preference and CTIP1? If it means number of seats available in total, then there is a signficiant possibility that a neighborhood school will not be considered oversubscribed before the preferences, but many neighborhood applicants could be shut out of their school because of siblings and CTIP1 and still not get the Dense Population Tiebreaker. Am I understanding this correctly?

  22. @Wayne – I addressed your question as best I could here. I don’t think staff has broken down first choice requests by CTIP 1 census tracts.

  23. Rachel,

    Can we revisit my previous question? How many students living in a ctip1 area requested Clarendon as their first choice last year?

    Thanks
    Wayne

  24. More on densely-populated attendance areas, taken from the district’s FAQ. It appears I was wrong on how they are determined — it’s based purely on the number of applications, regardless of whether these applications list the attendance area school or not, originating from addresses within that attendance area:
    How will the tie-breaker for dense-population areas work?
    If the number of students who submit applications and live in a particular school’s attendance area is greater than the number of seats available in that attendance area, then all students who live in that attendance area will get the dense-population tie-breaker.

    Do I have to list my attendance area school for the dense-population tie-breaker to apply?
    No, you will not have to list your attendance area school as a choice for the dense-population tie-breaker to apply. We do not want to create any rules that would require families to list a school on their application form that they do not otherwise want to list. We want a simple system that allows families to list their true preferences when they submit their application form.

    Which attendance areas have more applicants than seats available?
    This is not something that can be known before all applications are submitted; it depends on the number of students who submit applications and the number of seats available in attendance area schools. We will know and share this information as we generate assignment offers, and over a period of years we will be able to assess the impact of this tie-breaker.

  25. @fj_murphy I think you’re unclear because I don’t think the district has been very clear about what happens after the first application period. We are trying to back away from “rounds” (as in Round I or Round II) and also the idea of wait pools — I hesitate to say more because I don’t have a clear understanding yet of how it all works. But I agree with you that having a consistent standard for when we stop placing students is very important.
    I’m told that many of the remaining questions will be answered by the Nov. 13 Enrollment Fair, when the district will officially kick off the new student assignment process. Stay tuned.

  26. Hi Rachel,

    Thanks for posting this, and especially posting the corrections. It’s hard to follow and people often have ideas held over from the previous system (e.g. listing a school first to get the tiebreaker).

    One thing that I think people are still unclear about is the subsequent rounds of assignment. You mention “wait lists” and “wait pools.” Presumably, those would all be “wait pools,” right? I’d imagine that the old “cohort” system would still have appeals and siblings, but instead of 0/7, 0/6, 1/7 and Late, we’d see preschool, CTIP1, attendance area, oversubbed area and “other.” I haven’t seen anything that is explicit about that.

    I’m also curious as to how the 10-day count will be managed. It seems that this year, it was managed strangely, with the “hard” 3-day count turning into a much softer 15-day count or something. I know it’s not most important part of the SAS, but having a consistent standard for dissolving the waitpools should be important.

  27. @KWillets – Those figures refer to past demand patterns under the old system. Even if such a breakdown were to be done, it wouldn’t be particularly helpful because we expect demand patterns to change drastically with the new attendance area preferences and the removal of the “0/7” preference in subsequent rounds.

  28. It’s on page 27 of this presentation: https://rpnorton.files.wordpress.com/2009/10/ad-hoc-committee-meeting_october-19-2009_stanford.pdf , “Proportion of students who receive a school they ranked”. The original data just has to be broken down by census tract or school attendance area.

  29. @KWillets I don’t think that projection has been done.

  30. The original studies had a calculation of the rate at which people receive their first and second choices, and the rate at which people receive none of their choices, etc. Can we get the same results broken down by school attendance area?

  31. Vicki of PPS-SF wrote me after seeing an excerpt of this post – her comments appear below and I have corrected the post above to reflect them (see the text that is struck through):

    A CTIP 1 resident would not need to list School C as their first choice to receive that priority. CTIP 1 residents would get priority in all of the schools they list regardless of where it was placed on their list. Now chances are they will get their first choice school but not necessarily especially if its a high demand school highly requested by CTIP 1 folks – E.R. Taylor comes to mind as a possibility. The same is true with SFUSD pre-k attendance area.

  32. Do CTIP1ers have preference in city-wide school applications, after siblings?

  33. @SunsetMom, the answer I’ve been given on that question is that the district should know, based on the number of applications that come in from a particular attendance area, which attendance areas are “oversubscribed” and which are not. Lets say attendance area school A has 40 seats after siblings and CTIP, but there are 43 applications originating from that attendance area (regardless of whether those applicants rank their attendance area school or not). That “overage” would designate that attendance area as “densely-populated” and the tie-breaker would apply to all applicants from that attendance area.
    Because this tie-breaker comes AFTER attendance area preference, you will never be displaced from YOUR attendance area school by someone whose application is eligible for the densely-populated tie-breaker (unless that person also has a higher tie-breaker, like CTIP 1). However, the “densely-populated” tie-breaker would give an applicant an advantage in applying to citywide programs, since no attendance area preference exists for those programs.
    The staff does not expect most of the high-demand schools on the West side will fill up with children from the attendance area — they actually expect that there will be seats left over. And we already know there is not actually enough space in South and East side schools to accommodate the children we know are there, so those applicants likely will get a densely-populated tiebreaker.

  34. Another assignment question along these lines — I understand that those in over-subscribed neighborhood schools receive a higher priority than the general applicant pool (so below sibs, CDCs, CTIP1 and neighbors — but above the general pool). Is that for starting round 2?? Because until a full run has been done, how can you know which schools are over-subscribed? It doesn’t seem like you could change people’s status half-way through the first run..

  35. I received a question offline about whether ranking one choice higher makes you more likely to get that choice over someone else in a situation where there are more applicants than seats. My understanding is NO. In order for the assignment system to remain strategically simple (meaning that families don’t have to use a strategy about how they rank their choices and instead are rewarded for simply listing schools in order of their true preference), how you rank a school cannot give you an “edge” over someone else. The way we maximize everyone getting their highest preference is the “transfers” round after everyone has been provisionally assigned. Links to video and powerpoints from Muriel Niederle’s presentation on this last year are here.