^^You know that there is a place on the GC rec for that, right? This information is being disclosed already because it is so important.
Yes, @gardenstategal … And low income students also provide actual proof of economic adversity, a status that the College Board might guess “wrong” based on the school they attend or their street address…
But, @calimex, that “proof” is in a separate file that admissions may not see. I don’t believe that schools are using this to disprove a student’s economic need, but we differ on this. It seems you believe that colleges will use this information to say that poor kids are not poor? It’s interesting – most other people are worried that rich kids will use this to get the "advantage " of being poor. But everyone who dislikes the idea of putting a test score in the context of the environment, especially the school, of the test-taker seems to think someone else will be the winner here!
@gardenstategal I’m very wary of ANY decision being made based on an algorithm with base assumptions that might be downright wrong – especially when it would be EASY to add a box to the common app that says “Check here if your household income is below X and/or if neither of your parents attended college”.
Here’s how how an algorithm based on home or school zipcode can get it wrong:
- Low-income kids living in rent controlled apartments in what appear to be expensive neighborhoods will appear wealthy
- Low-income kids in tony private schools or BS may appear affluent
- Wealthy kid living in mixed neighborhood and/or attending HS with LMI population will appear to be “overcoming adversity”
And there is no way to find out how your family has been “coded” and whethe they got it right or wrong. It is one thing to now know how your essay was scored or rated. That is meant to be a subjective measure. But this adversity score is pretending to be both accurate and objective and there’s no way to check it or hold someone accountable.
If you had to check an optional box regarding income, first gen status, or other kinds of adversity, they could verify the data for those on the “admitted” pile. If you claimed to be low income and didn’t apply for FA, for example, that would certainly warrant a phone call before the acceptance letter is sent. There’s a penalty for lying. THere’s no penalty for the College Board if they get the “score” wrong. No appeal process. No way to even check the score’s accuracy!
A more accurate description of the Environmental Context Dashboard (aka “adversity score”) is provided here (including a screenshot of the dashboard):
https://professionals.collegeboard.org/environmental-context-dashboard/detailed-data-description
FYI, the data is not based on zip codes but on census tracts, which are much smaller areal units, roughly equivalent to neighborhoods (think of several blocks in each direction).
The dashboard provides information both about the school (based on the school’s address) and the student’s residential neighborhood (based on the student’s home address). Thus all students residing in the same census tract will have the same residential score, and all students attending the same school will have the same school score. The school score includes information about the school’s SAT/AP performance, etc., to put the student’s SAT score in a context of peers from the same school. The neighborhood score is composite score of disadvantage calculated from 15 American Community Survey variables.
FWIW, I routinely create and use such indices for health outcomes research, so here are a couple of thoughts that may alleviate concerns: Census tracts are designed to be socioeconomically homogenous - e.g., an apartment complex or Section 8 housing will not be in the same census tract as the McMansion next door, although both will be in the same zip code. There is a very high, though not perfect, correlation between area-level and individual-level SES, and area-level data is routinely used as a proxy of individual-level data in health care and other fields.
Also, the College Board is working toward making the dashboard available/visible to the student, per those links above. That said, I would be even more excited if they also included schooling/residential history, to reflect changes (upward, downward, residential instability, etc.) in the student’s environmental context over time.
Census tracts are exactly what the San Francisco Unified School District is currently using to provide top choices in the public school lottery to kindergarteners living in census tracks that usually have low test scores on standardized tests…
The result? Affluent families pay top dollar to move to those census tracks when their children are infants or preschoolers in order to lock-in a kindergarten spot in a top school. THeir next door neighbors might be living in overcrowded conditions in a rent-controlled apartment and their kids may qualify for subsidized meals at schools.
They will both get the same residential score?
People are blatant about marketing their apartment rentals/sales as being in a CTIP area that provides advantages in San Francisco’s public school lottery. I saw a listing last week for a $6,000+/month rental apartment that emphasized that fact!
https://www.zillow.com/homedetails/1109-Broderick-St-San-Francisco-CA-94115/15084328_zpid/
“For families looking to enroll children into top SF public schools, the condo is currently in CTIP”
I understand what you are concerned about. But a kid who needs FA and who is at a good BS HAS had the benefit of an excellent education and excellent prep for the SAT so may have that “context” score adjusted accordingly. Iow, it would be expected that that kid is more privileged (not more wealthy) than the kid left behind. I think we would both agree that’s accurate. (Or we wouldn’t have picked BS for our kids.) At some level, you can’t have it both ways. And the essays on the common app are yet another way to bring up the family situation if it’s important.
I believe the “decision” though, is how to view the SAT scores and what other clues might be in the app. Not whether to admit that kid! My guess is that this is just a little piece of data that goes into the 10 minute app scan.
As a quick aside, the frenzy that is college applications today is stressful. Really stressful.
And it seems like no matter who you are and where you stand in the process, it feels like it’s stacked against you. To kids who aren’t wealthy, the “rich” kids have tutors and expensive ECs so can put together impressive apps. To those “rich” kids, whatever they’ve accomplished is negated by the fact that it was all "expected " of them and “easy to achieve”. You probably won’t believe this until you come out the other end, but colleges see through a lot of this pretty well and most kids, particularly ones with parents like the ones here who can help guide them and who have spent time understanding the landscape, will emerge happy with their very good options.
@gardenstategal You think it is fair that wealthy, privileged families who move into a low-income census tract to get a top choice public school in SF should have the same residential “score” as their truly disadvantaged, low income neighbor attending the same school?
Census tracts can and do get redrawn each decennial Census to keep them socioeconomically homogenous. I don’t know anything about the SF Unified School District, so I cannot answer your questions, @CaliMex. However, if there is such a dynamics as you describe, it will undoubtedly result in a tract being split or redrawn accordingly.
The residential address that serves as the basis for calculating the neighborhood score is the address of the student’s primary residence (e.g., for BS kids that would be their home address in their home state). I don’t know what happens if a family is wealthy enough to have multiple residential addresses, some of which they’re using to work the system as you’re describing @CaliMex. That may be one more reason why multiple neighborhood scores should be included when applicable, to reflect both historical changes in residential context as well as identify those families you’re referring to. But no one asked my opinion… I do hope, however, that the CB has consulted with social scientists. One can do a lot of damage (or waste a lot of resources) with such tools if they don’t know what they’re doing.
@GoatMama I’m hoping SFUSD drops lottery preferences by census tract. The lines are not redrawn often enough to curb abuse and gaming the system. By the time the lines are redrawn, people have already benefited …
The colleges have the actual data… they shouldn’t need to pay a third party to make a best guess.
So let’s assume they both have the same score which suggests they are both disadvantaged. This triggers a harder look by the adcom. Will other things on the app point to the differences? My guess, but it is only that, is yes. But at least the kid who is truly disadvantaged has gotten a better look.
And again, I realize cities in particular may have complicated geographies in terms of SES. But a lot of the country is very segregated along those lines. Live in xyz county in flyover state? No options except your LPS.
Really, neither of us know exactly what goes into this “score” and neither of us know what an AO will do with the info. I am hopeful it’ll identify kids with potential who have had obstacles to achieving it so that they get a second look and want to be optimistic because colleges want that second look too. Programs like Questbridge do wonders for disadvantaged kids at the top of the talent pyramid but there’s much less opportunity for those a little further down. Maybe this will help a teeny bit?
What would trigger a harder look?
The AO doesn’t have any reason to think something is amiss.
And as you stated earlier, they don’t have time for a hard look or they would handle looked at actual income data instead of the College Board’s guess based on census tract 
They could ask directly… then verify and penalize liars later in the process…
The SFUSD lottery is hopelessly dysfunctional and should be scrapped imho, for reasons far beyond the census track.
Pulling back to the abstract concept: Is it possible to develop a “good enough” (not 100% accurate but does add useful information to an admissions file most of the time) adversity measurement? Probably. I have no idea if this is it, nor does anyone else.
The SAT is flawed. Grades are flawed. ECs are flawed. Letters of Rec are flawed. Every single measurement AOs have is flawed. Perfection is never the standard. Scammers will always exist.
To me the question is whether the intended goal is appropriate. And if it is, then I would want to look under the hood, to see if the goal is being met. The tool should improve over time as they develop it. But to toss it out because it is imperfect is premature.
And let’s remember that the thing is not an “adversity score,” despite people wrongly calling it that. It’s an “environmental context dashboard,” which does just that: puts the individual SAT score in the context of the school, and the student in the context of his/her residential neighborhood. You can see what it looks like in one of the links in post #744.
The SFUSD lottery is in the process of being revised. The challenge is that they want the benefits of racially and socioeconomically integrated schools, but can’t ask about race or income directly. Colleges actually have access to the income information, they just don’t have time to review it. To me, that dictates a different solution (a score based on actual data) than having a third party without access to the actual data SELL a flawed best guess.
Individual SES (household income, parental education, occupation) is certainly important, but so is the residential context. Why not take a look at both? We know that neighborhood context affects child health outcomes after controlling for individual SES. Why wouldn’t it be similarly important for educational outcomes? In fact, a quick search shows that it is.
Also, keep in mind that the neighborhood context score is not based on income alone but is calculated from 16 indicators:
- Median family income
- Percentage of all households in poverty (poverty rate)
- Percentage of families with children in poverty
- Percentage of households with food stamps
- Percentage of families that are single-parent families with children and in poverty
- Percentage of families that are single-parent families with children
- Percentage of housing units that are rental
- Percentage of housing units that are vacant
- Rent as a percentage of income
- Percentage of adults with less than a 4-year college degree
- Percentage of adults with less than a high school diploma
- Percentage of adults with agriculture jobs
- Percentage of adults with nonprofessional jobs
- Percentage unemployed
- College-going behavior
- Probability of being a victim of a crime
And by the way, all the data is free and publicly available. You guys give me your home address and I’ll tell you exactly what your neighborhood score is.
FWIW, we recently introduced a similar neighborhood context score to the electronic health record of all pediatric and adult patients in our health system, as it strongly predicts a variety of health outcomes and can be used as a risk prediction tool. It doesn’t replace the patient’s individual data, it just alerts you to potential risks that come from the environment rather than the individual (food access, air quality, walkability, rurality, housing quality, living conditions, etc.).
I’m sure the data is incredibly valuable in understanding a particular community. But you need to know whether the individual you are seeking to understand is typical of their community or an outlier to ensure you aren’t making gross generalizations about them. If you are making decisions about an individual, you need to understand that individual first, then understand them in their context. I is dangerous to start by understanding a context – which they may or may not identify with or feel is relevant – and then making generalizations based on that context.
I don’t think the CB, or anyone here, is advocating for not using individual data. The entire application is based on individual data. What’s wrong with adding some contextual data as well? There will always be outliers, but the majority of poor people in this country live in poor neighborhoods, and the majority of wealthy people live in wealthy neighborhoods. Should we not worry about the majority before we worry about outliers? And with that said, I realize we’ve digressed to the type of back-and-forth argument that CC doesn’t seem to condone. So we just have to agree to disagree. 
@GoatMama Are you a data scientist? There is a growing literature on algorithmic fairness and the trade offs between accuracy and fairness… and increasing concerns about the unintended negative and often discriminatory effects of decisions-by-algorithm.
At the end of the day, the College Board is looking for a revenue stream. THAT is the impetus for their latest product.
Changing topics… In looking at interesting summer programs, it seems as if some of the most interesting ones – highly competitive but also highly subsidized or FREE – are designed for the summer between junior and senior year. Do any of you have kids who will be participating in any of these programs?