The Lottery: A case study

I thought I had published this note a while ago, but am not finding evidence I did. I did something similar, but it only covers the summary worksheet. The following is using an example from the 2010 Kindergarten Lottery, and helps to show the inner workings of the Point System for the Lottery.

So for starters, imagine you have a spreadsheet with 688 rows by 64 columns. Each of the columns have strange headings like proxst, or ch1prises, or asgsch. I have figured out what most of those things are – probably most anyone could decypher this mess if looked at long enough. Now that I think about this, I should make this case study a multi-part series… but I wanted to get some information out quick for certain folks, so here is what I have now.

When you submit your application and the information eventually makes its way to Dr. Alves in Massachusetts, he fires up his program. The first thing it does is assign a random number to each applicant. This random number is identified as ccrnd, or “Controlled Choice Random”. Please keep in mind I am totally guessing at the human-readable text – it is what makes sense to me. If everything else is exactly the same (for instance, if you just happened to submit an application twice), this number would be the ultimate tie-breaker.

When you filled out your application, each of your inputs were codified. For example, you checked a box that marked which level of income you made. This automatically becomes a number, for instance, “Income Level Code 5”. The column for this data is incomecode. You get more points for lower code values; or in other words, the less you make, the more you are favored. In most cases, the point system uses a logarithmic scale, meaning that the high end is really high end. Yes, yes yes, I realize you could totally game the system and LIE. That is not the point of this post – a pox on you if you abuse your knowledge. So keeping that in mind, here are some more:

mealcode: in 2010, you got the same points no matter if you received free/reduced lunches or not

childno: more points for more children in the household

adultcode: generally more points for fewer adults – the data is really weird for this one

pedpts: a summation of p1educode (first parent’s education) and p2educode (second parent’s education). More points for less educational achievement

precode: a non-intuitive assignment of points. Lots of points if you went to Head Start, least points if you went to private school. Not going to preschool is just a notch above private school. Couple other values in there as well.

Your “SES” score (ch2sesprt – choice 2 SES part) is the sum of the above (incomepts, mealpts, childpts, adultpts, pedpts, prepts). In general, they are all weighted the same except childpts, which has about 3/5 the weight of the others. The max possible score is 2,300,000 for SES.

The next group of points is for “Proximity” and siblings. The points are all in millions and trump SES:

Proximity B (ch1prxb): 3mil

Proximity A (ch1prxa): 4mil

Sibling (ch1sib): 5mil

Proximity (doesn’t matter which one) + Sibling: 6mil

Proximity and Siblings only counts for Choice 1 – no other choices.

Your total score for Choice 1 would be:

Total SES (ch2sesprt) + Prox/sib points (ch1pripts) + the random number (ccrnd) = ch1aprnd

Your total score for Choice 2 (and all others) would be:

Total SES (ch2sesprt) + the random number (ccrnd) = ch2aprnd

You get your first choice if your total points is in the Top N scores, where N is the number of available seats. There is some special calculation whereby some Proximity B seats are reserved, thus a Proximity A person may be bumped by a Proximity B person if Proximity A seats fill up. Weird science, eh?

For the Second choice, this becomes the “Second round” if you will. After all the First choices are processed, a new number of available seats is calculated and becomes the new N. If your score is in the Top N for 2nd Choice, you are golden. So on and so forth.

Finally, there is one more strange quirk to this whole thing. Some folks are labelled as “Low SES”, while others are “Non Low SES”. There is absolutely no documentation about what distinguishes one from the other. Unit 4 will tell you it totally depends on Free and Reduced Lunch. But Dr. Alves twists this; the lowest Non Low SES total points was 205,000, but the highest Low SES was 2,000,000. A magnitude of difference. Go figure.


2 Responses to “The Lottery: A case study”

  1. charlesdschultz Says:

    Here are my annotations for all the columns in the 2011 Lottery raw data file:

    Column HeaderWhat I think it means
    stidnStudent ID
    stlnaStudent Last Name
    stfnaStudent First Name
    gradeStudent Grade level
    sibgroupSibling group – not exactly what this pertains to
    sesstatusLow or Non-Low Socio-economic status – not sure how it is derived
    preasgPreassigned school; Mostly from Magnet school selections in Feb; some due to being held back
    preprgPreassigned program
    asgschAssigned school after going through Choice Program (represented as a number)
    asgprg”Assigned Program (ESL, Regular, SPED, etc)”
    CH ASG”Which choice this applicant received, if any”
    ccrndControlled Choice Random number
    ch1aprndTotal of 1st choice points (Random + ch1prises)
    ch1prises”Sum of 1st choice Sibling, Proximity and SES points”
    ch1schWhich school was the first choice (id number)
    ch1schnaname of the 1st choice school
    ch1prg1st choice program
    ch1pripts1st choice proximity + sibling points
    ch1sib1st choice sibling
    ch1prxa1st choice proximity A
    ch1prxb1st choice proximity B
    ch2aprndTotal of 2nd choice points (Random + ch1prises)
    ch2schWhich school was the 2nd choice (id number)
    ch2prg2nd choice program
    ch3aprndTotal of 3rd choice points (Random + ch1prises)
    ch3schWhich school was the 3rd choice (id number)
    ch3prg3rd choice program
    ch4schTotal of 4th choice points (Random + ch1prises)
    ch4prgWhich school was the 4th choice (id number)
    ch4aprnd4th choice program
    ch5schTotal of 5th choice points (Random + ch1prises)
    ch5prgWhich school was the 5th choice (id number)
    ch5aprnd5th choice program
    aprndRandom points for Pre-assigned students

    statstatus: whether this child was assigend or not
    sibstatsibling status: whether the sibling is assigned or not
    asgtag”Assignment tag; why this student was assigned the school they received (because of pre-assignment, sibling, proximity or choice)”
    pxastat”If they have Prox a, this column either says they were assigned or the reason why they were not”
    pxbstatSame for Pox b
    ch1statThe reason why the 1st choice could not be granted (usually because the school ran out of seats; sometimes because the school ran out of reserved seats)
    ch2statThe reason why the 2nd choice could not be granted (usually because the school ran out of seats; sometimes because the school ran out of reserved seats)
    ch3statThe reason why the 3rd choice could not be granted (usually because the school ran out of seats; sometimes because the school ran out of reserved seats)
    ch1wno”These are all 0, so it is hard to tell”

    spedSpecial Education
    eslEnglish as a 2nd language
    homelangMother tongue
    speddesc”Spec Ed Description (ie, what type of Spec Ed)”
    mealcode”In the past, this was one of 4 values (R, F, P and S I think)”
    mealptsNot sure why Alves fills these in with all the same value (except for those that did not mark anything)
    incomeRange of household income
    incomecodeA corresponding code for household income
    incomeptsHow many points they get for their income.
    hhsizeHousehold size
    childnoNumber of children in the house
    childptsHow many points they get for number of children in house
    adultnoNumber of adults in the house
    adultcodeA code for number of adults in house
    adultptsPoints received for number of adults
    liveswithWho does the child live with if not both parents?
    p1genderGender of 1st parent
    p1eduEducation level of 1st parent
    p1educodeCode for Educational level of 1st parent
    p2genderGender of 2nd parent
    p2eduEducation level of 2nd parent
    p2educodeCode for Educational level of 2nd parent
    pedptsPoints for total educational level
    preschWhich pre-school did this child go to
    precodeThe code for each preschool
    preptsPoints for going to certain preschools
    whiteIs the child white?
    nativenative american?
    multimany ethnicities?
    genderGender of child
    aptapartment number
    zipzip code
    phonephone number
    proxbaIs the address within Proximity A of Barkstall?
    proxcbCarrie Busey?
    proxdhDoc Howard?
    proxssSouth Side?
    proxwvWest View?
    proxghGarden Hills?
    proxbsch”If not Proximity A, which school is Prox B?”
    appstartDatetime this record was processed by Alves program
    appsubDatetime this record was completed by Alves program
    cityCity of family residence
    chkaprndnot quite sure 🙂

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: