This weekend, triggered by a few readers who disagreed with my assertion that socioeconomic status is a huge driver of educational attainment and performance, I decided to respond the way any nerd would in my situation: I made a chart.

In a moment of Excel fervor, I took data from the College Board’s 2011 Total Group Profile Report of college-bound high school seniors and plotted the mean combined (Reading, Math, and Writing) SAT scores for various income cohorts.

Take a look at the chart below. On the horizontal axis is family income. On the vertical axis is the combined average SAT score for students from families in each group. The general story is pretty simple: The higher the parents’ income, the higher their kids’ SAT scores.

(And this, sadly, is but one example. For a fuller account of the link between income and education, see the work of Duke’s Helen Ladd.)

UPDATE (2/21, 5:15PM EST):  

Thanks for the mega-response to this post. A few things:

1. Let’s all say it together: “Correlation is not necessarily causation.” Just ask XKCD.

2. Several readers have suggested that the real driver here is the parents’ education level rather than their income. Maybe. But without acquiring the underlying data and subjecting it to a more sophisticated analysis, we can’t say for sure. One thing we do know, from Table 11 of the report, is that among households where at least one parent has a graduate degree, the average combined score is 1687 — which is higher than for most income levels but lower than the average score for the highest income households. What’s more, there are three times as many households with graduate degrees as there are households with incomes over $200K, so something else might be afoot.

3. My hypothesis about that something — a guess rather than an assertion — is that the households in the top tier often have two parents with graduate degrees. That is, they’re rich and they’re well-educated and that’s a hard combo to beat. If that turns out to be true, it suggests that one of the most influential, but least remarked upon, social forces in America is assortative mating by education level.

53 Responses to “How to predict a student’s SAT score: Look at the parents’ tax return”

  1. Dan, you make an excellent point and your graph is compelling. However, the untold component of this story is that highly effective instruction can actually disrupt the pattern you found. John Hattie’s recently revised research lists a host of factors that have a greater effect on student achievement than SES. If anything, I find your post a “call to action” to make sure that all students have access to the same “top quality” instruction. (Read more here: http://grantwiggins.wordpress.com/2012/01/07/what-works-in-education-hatties-list-of-the-greatest-effects-and-why-it-matters/) Thank you for giving me some motivation on this COLD Tuesday morning!

    • Nathan says:

      How is it worse to be born into poverty and get bad scores than it is to be born stupid. If you’re born smart you somehow deserve it more than being born rich? Doesn’t really make sense to me.

      Besides, human genetics should be consistent enough that you don’t see much variation in the intelligence of different individuals.

      I think there will always be external factors like nutrition, disease, environment (toxicity).

      If you level out everything, then I guess you’re just left with motivation, and when you level that out you have equally achieving students.

      My point is… the only reason you do better than someone else is because you have an advantage. Being better at something is the same thing as having an advantage… unless you don’t count motivation as an ability, but it will inevitably come to be seen as one when people are failing and they need to improve.

  2. Rohan says:

    Wow.

    That’s telling.

    Dan, over the past 4 months or so, I’ve been running a fledgling feature on my blog interviewing ‘Real Leaders’ inspired by a not-for-profit I am an active alumnus of, called RealAcad.

    I’m wondering if you might have 20 minutes of time for a short interview..?

    Do let me know. Happy to reach you via email if you feel that’s best.

    Rohan :)
    rohan[at]rohanrajiv[dot]com

  3. COD says:

    My score scored well above his family income range :)

  4. Julie says:

    COD, averages are averages for a reason. People score both above and below them, sometimes greatly above and below them. It’s great that your son scored so well, but that doesn’t change the fact that his results are not the norm.

  5. Beyond the obvious, disappointing point DP’s data makes, look at the jump in performance from incomes from $200,000 to the next cohort, over $200,000. Scores leap 85 points in that one step.

    85 points of increase takes you all the way back four income groups going in the other direction.

    My question is this:

    What is income a surrogate for here? Why would kids with parents making $250,000 a year do so much better than kids with parents making $140,000, when both families would be considered quite well off.

    • benmooremerq says:

      Charles,

      Family incomes over $200,000 are highly correlated with one or both parents having post-graduate degrees and living in neighborhoods where achieving high SAT scores is both encouraged and expected. In most of the USA the highest income neighborhoods are peopled by PhDs, MDs, DDS, MBA, and JDs as well as people scoring in the top 10% on the SAT. This is especially true in the DC metro area and Silicon Valley.

      I strongly encourage anyone who is interested in this to read Charles Murray’s seminal work “Coming Apart”. Murray uses Census.gov data as well as Alumni directories of the Ivies to show very profound clustering of highly educated people, as well as the income correlation. Look more into the “Super Zip” term that Murray coins and you can find it. There is a significant income difference between families of National Merit Scholars and those who simply do well on the test. It is just reality.

  6. Craig says:

    This is interesting data but it is important to point out that the correlation between income and scores does not necessarily mean that making more money leads to higher test scores. In fact, there are probably a number of reasons that this data correlates this way.

    It could be that there is a genetic component. There could also be access to educational opportunities. Income probably determines location of home and therefore quality of schools. There may be paid tutors at higher incomes, especially over $250k. etc, etc, etc.

    In any case it is interesting data.

  7. David Jacobs says:

    Big caveat! You cannot deduce causation from correlation. For example, Harvard grads make more money than non-Harvard grads, but further study discovered that Harvard wasn’t the cause, they were just good at identifying and accepting those who would likely be more successful.

  8. Jennifer Friedman, Ph.D. says:

    I am glad that someone exposed this trend, as I see it everyday in both my position as a school psychologist and as a private therapist for many affluent teenagers. My colleagues and I discuss this sutuation often and lament its impact in our children. I am saddened that these direct correlations persist and continue to negatively impact the future of our society. I frequently see affluent teenagers taking the SAT or ACT upwards of five times in order to elevate their scores. They also have private tutors and/or pay for test courses in order to maximize their performance. If every student had access to these supports, then the college entrance playing field might be leveled. However, this is not the case at all, and the result is a perpetuation of class stratification.

  9. James Evans says:

    “Highest level of parental education” (the next section in the report you cite) would show the same pattern, with children of parents with a graduate degree having the highest mean SAT scores and parents without a high school degree having the lowest.

    And since education and income are positively correlated, maybe it is parental experience with education themselves that explains the pattern in your group.

    In any case, you need a model with multiple predictors to really understand educational or SAT performance.

  10. Robert Holdar says:

    Charles, the reason for the large jump on the chart above $200K is that entry includes ALL incomes above $200K whereas the lower entries are in $20K increments. Continuation of the $20K steps would most likely show a continuation of the smooth curve.
    The causes are most likely a combination of native intelligence and a family cultural support of educational achievement.
    As for outliers, remember that income levels are fluid and families frequently move from one cohort to another as individual situations change.

  11. Alex Sirota says:

    This is a fascinating correlation, Dan. Mr. Fishman makes an interesting observation asking the why the jump at the 200K+ level?

    Here’s what I think income is a surrogate for in general for all income levels: the amount of time available for parents (and parent proxies) to spend with their children.

    At certain income levels you are just working to make ends meet. You may be working 2 jobs. You may be working night shifts. You simply don’t have time to engage in conversation much less any serious work at home for any extra-curricular or even intra-curricular work.

    At the 200K+ level you basically have met all your basic needs for the ability have a nice house, with maybe a separate room for homework, away from the TV. You also most likely have support at home in the form of a housekeeper/nanny, which gives you the most precious gift of all as a parent: time.

    If used appropriately that time can be spent taking kids to sporting activities (freeing up their brains away from school), to social activities and of course to the all important — helping your children with homework and even prepping for the SAT. Of course a few extra bucks can buy tutors and prep materials too, which others may find a luxury if they don’t have the basics taken care of.

    It takes 2-4 hours of engagement per week on a regular basis, I have learned, to increase my son’s grades from all Bs to a mix of As and Bs. This just happened to him for the first time in 6th grade. We started a regime of 2-3 hours of work using Khan Academy as a way to introduce homework into his schooling. His school doesn’t regularly provide homework outside of major projects, but I do believe engagement with my son around a set of activities he gets to choose (rather than ones chosen for him) bled into his schooling and ability to achieve better scores on tests. Surprisingly (or not surprisingly) this ability to provide time and effort, also allowed him to improve scores in subjects he and I did not explicitly work on, like drama.

    So higher income = ability for parents to provide time and/or materials for kids = higher test scores.

    That’s my theory, at least.

  12. Dan Pink Dan Pink says:

    James –

    Yes indeed. My guess — and it’s only a guess — is that if we held income constant, the parents’ level of educational attainment would be the very best predictor of all. (But again, education and income are so tightly connected these days that it might not show much of a difference.)

    Also, remember that these numbers include only those students who’ve taken the SAT and therefore are presumably thinking about going to college. If we widen that net to *all* students, the differences would be even more stark.

  13. Brad Pollen says:

    It saddens me to see children held down. As has been posted correlation is not causation. For example, look at the phenomenal scores of Asian students and the unbelievable outcome of KIPP Academies in the worst possible neighborhoods. Our educational spends more with less results and is fundamentally constrained by the teachers union.

  14. Wittybrit says:

    Dan, from just the first few comments your really seem to have hit some nerves here – great job at sparking the discussion. Whatever the deduction of causation from correlation, this process is based on looking at how we got to where we are today.

    More interesting, I think, are the questions it raises about the future:

    What it would take to create a world where there was NO correlation between parents’ income/education and the child’s acedemic performance (for whatever causes)?
    Is that a world we want to create?
    If not, what is a tolerable amount of inequality of opportunity?
    If so, are we individually and collectively willing to pay the price to create it?

    • Gael N says:

      We definitely do want to create a world where money has zero correlation to performance. If you read Outliers by Malcolm Gladwell he talks about a study in which they followed the careers of a group of kids in the top 0.1% of IQ. Before the study there was an reasonable assumption that high IQ was an indication of success. Instead what they found surprised them. Most were in average jobs, a small percentage were in careers considered “successful” and there was even a small percentage that were unemployed or in low paying jobs. This stunned the researchers so they processed all the information they had. The one thing they saw that consistently came up was that the ones from higher income families were usually more successful then those from lower incomes.

      IQ tends to be a bit more distributed than income. We need a system that allows kids with high potential to succeed and not by bound by the circumstances they were born in.

  15. Gadsden says:

    Nice little chart, but not very meaningful without controling for parents’ academic achievement.

  16. Dan Pink Dan Pink says:

    Gadsden –

    More from Table 11:

    Students from households with incomes above $200K have an average score of 1721. Students from household with a parent with a graduate degree have an average score of 1687.

    Of course, there are limits to what we can conclude from that difference. And there are obviously more sophisticated ways to analyze the totality of numbers in this report. Indeed, I encourage someone with access to the underlying data — and perhaps a graduate student or two — to do that.

    But this income-SAT correlation, which is based on the results of 1.6 million test-takers, is pretty stark.

  17. Patrick says:

    Ms. Friedman, as a PhD you should know that correlation does not imply causation. The chart shows that income and SAT scores are related, but it does not necessarily show that income caused those scores.

    I do not doubt that a high household income contributes to higher SAT scores, but there are other, confounding variables that are likely to at least some degree driving both SAT scores and income.

  18. Adam says:

    Higher income == ability to afford tutors and/or SAT prep classes. Wow, what major insight.

  19. Cheryl says:

    The book ‘Freakonomics’ takes this subject as one of the areas in which it makes it ‘s point that direct correlation does not say much about causation. However, they take on this topic in great detail, although the data is older of course. And they look at elementary education, regents, etc. as well as college. They look at and hold constant 12 different socioeconomic factors and parental activities. The outcomes are not what you would think, of course. The gist of the study is that ‘who’ the parents are (including income, education levels, genetic ability)is the prime determinant in childrens’ academic achievements. Surprisingly, what they ‘do’ for their children (e.g., move to a better neighborhood), comprising at least 6 of the factors they included, has little quantifiable effect. That’s a vast simplification but I recommend the chapter on this topic to anyone concerned about this issue.

  20. julia says:

    I would like to see the statistics related to the groups that fall OUTSIDE this Score/Income trend. What are the instances of students from LOW income achieving HIGH scores? and students from HIGH income achieving LOW scores.

    What, if any correlating patterns can be perceived by viewing and comparing these chart trends?

  21. Chris Santos says:

    Dan -

    A couple of points: First, blaming the teacher’s unions is not the solution – there are ways to rid the system of bad teachers, but administrators don’t/won’t take the time it takes to do so. The union provides a due process: one cannot fire a teacher just because…Second, the education/poverty correlation makes it a challenge to break out: we mirror what we are taught/experience. Now this isn’t to say that one cannot break out of this cycle, and there are cases where extraordinary people have broken out the educational/poverty cycle. Third, can we really expect much change when poverty levels have not changed since the war on poverty during the Johnson administration? (http://www.npr.org/templates/story/story.php?storyId=1589660)
    Fourth, we all would like simple answers to the current topic of choice: education. However, as many previous posters have indicated, whether it is causation or correlation, we are ultimately dealing with kids. In a child’s mind, there are factors that are beyond the control of many adults, no matter how influential, that affect how s/he sees the world and therefore interacts with it. As an educator having taught for nearly 17 years, I have seen good teaching and bad teaching. I have come across students who are remarkably resilient in the face of adversity and students who crumble at the sight of the simplest problem. Despite what we’d like to happen, the reality is that we are dealing with humans, and like I said before, children. This makes this even more difficult. As any parent knows, each child is different, try dealing with 32 in a class. There are no easy answers to this problem. Ultimately, we can correlate and find causation, but as you have indicated before in other posts and in your books, people are different, are motivated by different things (money, etc). We can never perfect the human condition, we are not robots and cannot input specific data to get specific results. As a child of poverty, I now sit here with a MA in Education and my Administration credentials, my SAT scores were in line with the chart (low). So, say what you want, the reality is that nothing is certain. Human intelligence is varied (see Ken Robinson’s The Element), much of which we truly do not understand (brain science is still developing). I don’t need a chart to know what is possible and what is difficult. Until you walk in someone else’s shoes, we can truly never understand the struggles and pain others go through.

  22. Nick says:

    While there is evidence that socioeconomic status is loosely correlated positively, the real answer to checking it isn’t too difficult. Extend your trend until you are mapping out the wealth of the top 500 wealthiest people in America. We know most of their degrees and their incomes, so you should be able to map that to data about the SAT scores. I’m pretty sure the results at that tail will turn out to be contrary to your analysis, but they should point to something truly accurate.

    We know the greater predictive correlations are that of quality of education and work done by the student. Higher income means it is easier to gain access to better education for their children. Hence the reason so many inner city parents place so much emphasis on the lottery for charter schools and why we see a loose correlation between income and SAT scores.

    Unlike Chris Santos mentioned, I disagree and look at evidence about teachers unions being problematic because they typically are the force that prevents certain changes that can make teaching students better easier. I’d say blocking the firing of teachers that are under-performing, but ignoring that, by helping to set curriculum and teaching methods that are archaic. Not always the case, but often. Unique education can be done in a class of 32 as it can be in a class of 50 or 10. Our system isn’t set up that way and there are other systems around the world that do manage to do it. Unfortunately, it means choice of education style and schools and uniform mandates don’t help that.

    The big one that I want to address in his statement though is that poverty hasn’t changed. The problem with statistics is that people believe them without applying context. If you assume income is the only measurement of a person’s status then you’d also need to include the fact that there are multiple measures of income that all say different things. Like some saying those in poverty have had an increase in wage growth since Johnson and so forth. Also you should include in benefits and other things that are now given by employers in addition to income. Those increase total compensation and it is total compensation & assets that is a more accurate measure of wealth and growth in poverty rates not income.

    The students you’d want to study though are the outliers that Julia mentions. There is something different there that might be useful in helping to develop education models that can be applied to students in a region.

    Honestly though, it’s too complex a problem for simple mandates and the only way to simplify it is to break down the numbers into smaller and smaller chunks. Let schools work the way they want, so long as they meet certain goals over time. Give students the option to go outside of zones so they have a chance in better schools for them. Etc.

  23. Mike Germaine says:

    So you have essentially shown that smart people tend to make a lot of money, and the kids of smart people tend to be smart themselves. Not to be cheeky, but what exactly is the revelation here?

    • Jackie says:

      Also, university students marry other university students nowadays, not the secretary at the office which might have smoothed the IQ spread. Genes are becomming more concentrated.
      Then there is marriage. It as a high correlation for kids’ educatioonal outcomes. So the reality is that if you want equality, stop the smart people marruing the smart people and staying married.
      The USA’s lowest people are still way higher than someone from Mali. Should that mean the USA shoukd transfer wealth to Mali to make it fairer?
      As for tutors seeing kids take the test 5 times to up their scores, good for them. That is tenacity.
      My one son got 800 out of 800 on his SAT. I am from Central Africa. It is time spent and interest in your kids’ education that made the gap close for us in one generation. Means less time for fun for myself but dedication to my children paid off.

  24. TR says:

    Another possible explanation is that people with more money live in better communities with better schools at younger ages. I work in the admissions office at an elite New England boarding school, and we have kids from every walk of life. We did a similar study on our own graduating students from the past 5 years, and found similar results. The students that were on financial aid, as a general rule, were slightly worse on the SAT than those that could pay the full tuition, even students that had been at our school for 4 years. The students on financial aid did out perform the full pay students in their graduating GPA by a wide margin.

    Our theory was that the students that could afford the full tuition had been in better school systems for a much longer period of time, and that was the root of their better preparation for the SAT. Students have access to the same SAT prep testing and are in the same classes in our boarding school, but the more affluent students can still out perform the students on financial aid in the SAT. The most interesting numbers for us were those students that lived in “good” zip codes, but could still not pay the full tuition. Those students, while on financial aid and by this graph would be on the lower end of the income spectrum, still scored very high on the SAT because they had access to the better public schools (that the families with high incomes created by paying heavy school taxes) from a very young age.

  25. Greg says:

    Regarding the xkcd comic, the best part for me is the “hover text” that shows when the mouse hovers over the cartoon.

    “Correlation doesn’t imply causation, but it does waggle its eyebrows suggestively and gesture furtively while mouthing ‘look over there’.”

  26. Morgan Rich says:

    As a life coach for teens and their parents, I get to hang out with all the kids who struggle in school and on the SAT tests (regardless of income). I help families move away from all the pressure and stress that comes with the drive to academic achievement.

    What we want is healthy, happy, and motivated human beings, not little grade and test score machines.

    My fear in these conversations is that we continue to bring relevance to SAT scores. By doing so, we are perpetuating a dangerous paradigm – that test scores & grades are more important than cultivating the caring & passion in our young people.

    What I see from this data is info about the question “who gets to be part of the club?” & more difficult is who doesn’t get to be part of the club (& what happens to them)?

    They end up thinking something is screwed up with them because their SAT score sucks. They doubt the qualities inside them that make them quality human beings, and no matter the income level, we need as many of these as we can get.

  27. Rob Savarese says:

    One thing that surprisingly no one has mentioned is how much hard work plays in this correlation. Is it possible that hard work leads to financial and academic success and those lessons are then passed down from parent to child?

    Character matters. Success principles typically are either learned through trial and error or passed down from parent to child. Granted opportunities are more abundant at higher income levels but it’s not the tutoring or extra help that is raising scores, IT’S THE ADDITIONAL WORK THE STUDENT IS PERFORMING! Whether this is done in a expensive private school or at the kitchen table is immaterial. It continues to surprise me how much income disparity statistics lead those in the education field to be blinded to the benefits of actually ‘doing the work’. Let’s stop thinking in terms of social justice and start actually teaching our children the benefits of hard work.

  28. Ben Knight says:

    I am having my son take the PPPPPPPPPSAT. Every 7 month old should! I will report back the findings.

  29. jeff says:

    Then parents and kids read things like this and believe their kids can’t do better. Then they work in mini-wage jobs and envy the rich. I have been a student and most of my buddies were a lot smarter than me academically( so many assumed) and when we started working my best bud was always messing up simple things and got fired a lot. My other Bud had piano, guitar and TONS of EXTRA classes has never worked a day in his life. That chart is only showing the students that have external situations dominating their lives. Not to mention SAT is just a joke all they really care about it our money.

  30. Morgan Rich says:

    That’s funny Ben. My kids didn’t do so well on their in utero UltraSAT, hopefully that PPPPPPPPSAT will elicit better results. I’m calling a tutor now.

  31. Jack says:

    If you plot parents SAT scores (i.e. while in high school earning minimum wage busing tables) with their kids SAT scores 25 years later, you’d get the same correlation. So much for the income equality argument.

  32. Bob Morrison says:

    Hi Dan… another way to look at this is using a graphic we created about SAT scores and per pupil expenditure when you factor in the economic category of the community. Here is the view for all schools in NJ:

    http://www.tableausoftware.com/public/gallery/new-jersey-test-score-analysis-visualization

    Again, as you noted, this is a correlation. Just adding another data point to support the case.

    Your friend,

    Bob

  33. Susana says:

    Without having read all the comments, I wholeheartedly agree with you, Dan, based on my personal experience in a variety of fields and settings as well as on what I’ve read. I find that we feel very uncomfortable talking about social class in the U.S., but this (and/or racism and sexism) is the heart of most of the inequities in our country. Also, I would guess most — if not all — of us on this list are rather privileged, which might account for some of the squirminess around this topic.

    Another factoid from my current field (healthcare): in the U.S., a person’s WEALTH is a better predictor of his/her health than diet, exercise, insurance status and whether or not s/he smokes. Source: Adler, N.E. et al. (2007). Reaching for a Healthier Life: Facts on Socioeconomic Status and Health in the U.S. Report
    from the Macarthur Foundation Research Network on Socioeconomic Status and Health.

    Kinda goes against “common sense”, eh!

    Keep on telling the truth and challenging us!

  34. John Thacker says:

    “it suggests that one of the most influential, but least remarked upon, social forces in America is assortative mating by education level.”

    What’s interesting is that one of the few who consistently remark upon it is Charles Murray, but he tends to be attacked quite a bit. Partly, I think, for his policy recommendations that follow his data about assortative mating, but partly because the phenomenon strikes at the new elite as well, a group that considers itself liberal and meritocratic.

    Also of some note is that according to the Ladd paper, the gap between rich and poor has grown, while the gap between black and white has shrunk (thankfully). As a result, the gap between rich and poor is now more than twice as large as the gap between black and white. That could perhaps give an incentive to alter the focus of affirmative action programs to some degree.

  35. Tom Cooper says:

    One factor which has not been mentioned has to do with mindset. People who make more money are achievement oriented.

    Countless studies have shown that achievers think differently from those who do not. Children of achievers are taught to think like their achiever parents.

    If we can teach people to think like winners – those who overcome – rather than as victims who are held down by “the man” – we will see an increase in those who achieve.

  36. Ale says:

    FALSE. I got an 1820, and my parents don’t even make more than $20 K.

  37. bj says:

    Another take on the issue.
    Social status makes it easier to use our intelligence in groups. Would be interesting to see if the “poor” did better in a room by them self or if we randomly didn’t tell them which in a series of practice tests would actually count.
    My daughter did well in a series of music performances where it was just her and a judge in the room. Then she had to do one where her peers got to sit in the room and watch her being judged. She choked. It’s not uncommon.

    Social pressure to appear brilliant is much stronger on those who weren’t born with a golden spoon in their mouth.
    Wealthy achievers do NOT teach their children how to overcome – they teach them that they deserve everything and can do anything to get it. Private schools for the affluent have much greater cheating problems than public schools.

    Also, there is just more pressure on the less affluent to perform well. If you’re rich, daddy will buy you into a good school regardless of your scores and he will make sure his friends offer you a cushy job you’re less qualified for than any of the average Joes around. If you’re at all nervous under pressure, the stakes for a poor kid are overwhelming. I’m amazed those in the under $101K families do as well as they do.
    See research below.

    http://psychcentral.com/news/2012/01/24/iq-drops-when-youre-in-a-group/34036.html

  38. James Uhl says:

    My SAT score said I should not even be breathing, but I maintained a 4.0 throughout my entire Bachelors and Masters experience, make about $125K/yr and do highly intrinsically rewarding work (public safety & teaching). The SAT is fraud!

    I do agree that your socioeconomic status will impact how, what, and whether or not you learn, but I was on the opposite end of this spectrum. My parents were upper middle class, I did good in school, but this limited, standardized, and highly algorithmic test essentially told me I was a waste of flesh. It is good thing I chose to ignore the results because SATs are not true measurements of your capacity for contribution and success in this world…the openness of your mind and heart are the truest of indicators.

  39. bob dobbs says:

    If you have all the data then why not do actual statistical inference first and give confidence intervals to support rejecting a hypothesis of equal means?

  40. EngrGrad says:

    I think the link between income and scores is correlated, though not directly. These SAT scores are a direct reflection of the primary education that students receive, which can be correlated to income through a few logical steps.

    In the US, we have a tendency to segregate neighborhoods based on socioeconomic status. People who have little live near other people who have little, and people who have a lot live near other people who have a lot. I think that is pretty well known (think of both gated communities as well as rent-controlled areas).

    In addition to this, we have a education system in this country where local communities largely fund their own schools, receiving minimal funding from State and Federal sources. If a community has very little overall income and property value, the community will receive very little funding for their schools.

    This provides a bit of a double-whammy for these students, as they live in low income houses and neighborhoods, likely influenced by people without significant education (just going on the trend that better education leads to better jobs) and additionally attend schools which are massively underfunded and likely incredibly ineffective at providing education and social services to these students.

    Compare that to students living in wealthy neighborhoods, where the schools have massive amounts of funding from the community. They are provided with not only the income advantage as well as the advantage of being in households with better educated parents, but also schools which can provide better teaching and better resources. I remember as a high school student travelling to high schools which had massive campuses, swimming pools, tennis courts, football fields with the same turf used in professional stadiums, highly advanced science labs, massive libraries, and no building was older that 15-20 years (seemed that there was always a new construction project going on).

    I know that throwing money at a system doesn’t make it better, but it is hard to believe that the most qualified educators would not like to gravitate to nicer neighborhoods, better paychecks, and better facilities to work in.

    As a college student, I had the opportunity to take part in NSF funded research which had a large outreach component. We took on high school students from local low-income areas to work in lab with us. These students were often selected by their teachers/counselors as being some of the best students. When they began working with us, it was obvious that at best they had 3rd-4th grade educations in math and science. These students were 16-18 years old and unable to multiply numbers together (even with a calculator). They were unable to solve BASIC algebra equations, and had never heard of most of the simple scientific concepts one would expect a high school student to know. This is because they were never exposed to it. How can someone expect to score highly on the SAT math section without knowing algebra? These school systems are clearly ineffective…

    We have created separate societies within this country, ones for the wealthy and ones for the poor. The education systems between the two are hardly comparable, and we are somehow surprised when the outcomes also are not comparable. Separate but equal should not be an appropriate educational policy in this country, not if we want to continue saying we are a land of opportunity.

  41. Very interesting correlation. I read through most of the comments too. At Bay State Tutoring, due to our location and the nature of out business we work with the high income levels. Unfortunately, at others have pointed out I think there are a few things at play here. Culture of success and means to afford SAT study programs seem to stand out the most for me.

  42. Late to this…but I wonder if there is any way to correlate scores to amount of SAT test prep purchased through tutors and or practice materials. Also would be interesting to see a cohort of kids from low income families assisted with access to free SAT tutorials and materials online and see the impact.

  43. Sentient says:

    I’ve attended several schools in several counties. From some of the worst crime ridden neighborhoods to some rather affluent neighborhoods. The schools in the worst nieghborhoods seemed to teach the basics while the schools in the middle of the really expensive neighborhoods seemed to teach subjects geared specifically towards SAT’s and IQ tests. Money really did not have any impact on the individual student, the curriculum seemed to make the difference. I’ve attended atleast nine different schools in about three different counties and what i’ve noticed is what you learn varies greatly on the average income of the population in each neghborhood even in the same city.

  44. michael roth says:

    Dear Mr. Pink:

    I worked in the field of education for the past forty-eight years. Approximately twelve years ago in an interview with a reporter for a local newspaper I stated my experience has shown a correlation between family income and standardized achievement scores of students. This did not sit well with many people. I am presently completing an article on this topic. I am emphasizing that it is not a level playing field. Children have advantages and disadvantages along the continuum and as a society we need to find the causalities and not just examine the symptoms.
    I ask your permission to use the graph you made of the SAT scores in my article. I will give appropriate credit. If you like I can send you a working copy of this work with the chart incorporated so you can make your determination.
    My best wishes for your continued success and efforts to bring data, information, concepts and thoughts to people for reflection.

    Michael Roth
    rothkokoro@aol.com

  45. David Roth says:

    Is this in any way surprising?

    Also, the strength of the relationship you see — a correlation of almost exactly 1 between income group and average SAT score — is not at all the same as a correlation of 1 by student observation. The correlation there could be quite low.

  46. Eric Love says:

    Your graph is not 100% accurate. I received a score of 1740 on my sats and my parents are not wealthy people. It is slightly offensive that you are asserting that “poor kids are dumb.”

    • Dan Pink Dan Pink says:

      Eric –

      I think you missed the point — several points, actually.

      For starters, the chart doesn’t mean that all people with high incomes score well and those with low incomes score poorly. It means there’s a correlation — a tight one — between family income and SAT scores. For instance, it’s true that men are generally taller than women. But one can’t conclude from that truth that all men are taller than all women.

      Second, the idea that I (or the chart or any of the commenters) asserts that poor kids are dumb is absurd. You also haven’t supported the claim with a quotation or a shred of evidence.

      Still, I appreciate your weighing in here.

      Best,
      Dan

  47. Gael N says:

    I would really like to see a chart on 1st generation Asian-Americans. They are known to be over-represented in top schools by being 40% of the enrollment despite the USA being 5% Asian. As a group Asians are not known as being in the highest income levels.

    My own theory is that even poor Asian families that immigrated were generally from the upper class back in their home country. The poor just didn’t have the resources to leave the country and the vast majority of them stayed. Even though they come with nothing, they still retain many traits from their affluence and education in their old country. So anyone trying to bring up Asians as a counter-example would actually be proving your chart correct.

  48. Fred Dobbs says:

    This is partially true but not entirely, many Chinese work in restaurants and their kids do well. They prove that you can do well in spite of poverty if you choose to and that a child’s decisions make a difference. If you study 16 hours a week and are poor and watch TV fewer than 7, you can get SAT prep books from the library and do well. Poor Asian kids with parents with no degree do better than white kids whose parents both have a degree. An obsessive focus on education is what the poor need but rarely have. You need to be really obsessed with grades. If you just do your homework 5 hours a week and a rich kid does, the rich kid will beat you, but if like Asians you become obsessed and work 15-20 hours a week, you’ll beat them. Also, the prior poster was right about the teacher’s unions. They hold kids back tremendously.

Leave a Reply