The GiveWell Blog

Understanding the achievement gap

From 2004-2006, I gave all my donations to organizations focused on helping inner-city youth (particularly academically). Equality of opportunity was my favorite cause, and I assumed (without having time to really look into it) that inequality stemmed from the gap in quality between different grade schools. I now believe that this assumption was badly wrong, and that as a result, my donations were mistargeted.

The most surprising thing I’ve learned about the achievement gap between black and white / low-income and high-income students is how early in life the gap is present. Every source I’ve looked at is consistent in this regard: children from different socioeconomic and ethnic backgrounds have large, systematic differences in academic performance in kindergarten, and these differences grow only slightly as children get older. In other words, most of the systematic academic inequality we observe is present by age 5. There is a good deal of literature on this subject, but the paper I’d recommend for starting to take a look is “Understanding Trends in the Black-White Achievement Gaps during the First Years of School” by Murnane, Willet, Bub, and McCartney (Brookings-Wharton Papers on Urban Affairs, 2006).

This is a relatively simple observation, but it completely changes the way I think about promoting equality of opportunity in the U.S.

  • It makes me much more interested in interventions that focus on early childhood, the period during which most of the “achievement gap” appears.
  • It makes me much more skeptical of the idea that equalizing children’s schools will equalize their educations, something that once seemed obvious to me. It makes me much less optimistic about what one can accomplish with “small schools” (a Gates-backed initiative that has produced disappointing results), private-school scholarships, etc. (The Murnane paper also references research on disappointing results from programs in this category.)
  • It makes me especially skeptical of low-intensity grade-school or high-school interventions such as after-school tutoring. I find it very possible that a few hours of extra help a week just aren’t enough to make a dent in deep-rooted disadvantages.

I think it’s interesting that this extremely basic, fundamental, and important fact about the achievement gap – how much of the gap is present by age 5 – has not come up in any of our conversations with (or applications from) charities themselves. That includes both education charities and child-care charities. It seems to me that most development and fundraising professionals are focused on reinforcing and serving donors’ existing assumptions; if you want to challenge your assumptions to get the best understanding possible, you have to look elsewhere.


  • eduwonkette on June 13, 2008 at 8:45 pm said:

    Interesting insight on the charities. If others are interested in the size of these gaps at age 5, the average black student enters kindergarten testing at about the 25 percentile of the white distribution in math (a gap of .663 standard deviations), and the 35th percentile of the white distribution in reading (a gap of .4 standard deviations).

  • JillB on June 13, 2008 at 9:24 pm said:

    I have supported inner city programs that take children out into nature and give them experiences that are not related to education and for self-exploration.

    I think (but cannot back this up with data) that too much attention is paid to the academic gap and the idea that early formal education is the answer to that (the preschool for all approach). Is this just another assumption? Maybe we should look at poverty, prenatal care, and support of families as the keys to whether kids are ready for Kindergarten.

    I remember a study that showed that the key to how well kids did in school was parental involvement. What if we approached the achievement gap as a family issue and opened up learning centers for kids and moms (or dads) and engaged the family in the learning process. It would give more confidence to the parents and rather than taking the kids away during the day to provide something outside of the family, would simply raise the whole bar for the community.

    Just as you say that late intervention is not a real solution, perhaps more school is not the solution either.

  • Carl Shulman on June 14, 2008 at 10:11 am said:

    Devlin, in his refutation of ‘the Bell Curve’, did a meta-analysis of twin studies involving 50,000 people, and attributed 20% of the variance in IQ to shared environment in the uterus at the same time (and only 48% to genetics). This suggests that something that can affect a mother’s uterus at one time and not another, like an infection/injury /toxin exposure, significantly affects cognitive development.

    Research to pinpoint the differences between uterine environments looks promising relative to typical education charities. Prevention of exposure of pregnant women to alcohol, lead paint, STIs, and other environmental insults might be both cheaper to do and more effective. Certainly, iodizing salt and providing sufficient micronutrients to pregnant women offers high returns relative to education subsidies.

  • Carl Shulman on June 14, 2008 at 10:35 am said:

    Although, for correcting inequality in the U.S., where salt and flour are already fortified, micronutrients are less important.

  • Andrea on June 14, 2008 at 3:40 pm said:

    Carl Shulman: You are assuming that an IQ gap is the cause of differentials in school and life success. I don’t think that’s the problem. Researchers believe persistent poverty is a result of “cascading” risk factors, so that a person with a high IQ who is born in a poor inner city surrounded by violence and crime will likely do poorly in school, whereas a person with a lower IQ but born into a middle class community with little violence and well funded schools will do well in school. It’s not about IQ, or IQ tests. That’s really the lesson of the bell-curve dust up.

  • Carl Shulman on June 15, 2008 at 10:13 am said:

    Certainly IQ is not ‘the cause’ of differentials in school and life success, but it is quite powerful: high school graduation gaps almost disappear when controlling for IQ. If we know that there is a major source of individual differences that is not genetic but affects kids before they are born, it should be targeted.

    We know that low birth weight affects scholastic performance and IQ scores, and African-American teen mothers are much more likely to have low birth weight babies. Identifying and remedying the causes of that difference looks to be quite valuable.

  • Andrea on June 16, 2008 at 1:20 pm said:

    “African-American teen mothers are much more likely to have low birth weight babies..” Certainly, but it’s hard to control for all the other factors that affect many African American teenagers who give birth to low birth weight babies, factors that are also the cause of low birth weight. It’s really a cascading effect, not a single cause-effect relationship. I see people here sometimes wanting to zero in on that one cause that would make a difference, but I don’t think it works that way.

  • Holden on June 17, 2008 at 10:55 am said:

    We agree that early-childhood and pre-natal care holds promise, and that’s part of why we support the Nurse-Family Partnership. At the same time, just because one factor is important doesn’t mean the others aren’t; we believe in keeping our minds open to any program that has demonstrated success. We just think it’s appropriate to approach equality-of-opportunity programs with skepticism (i.e., look for empirical evidence and not just a plausible story) given the complexity and depth of the problems they’re addressing.

  • Andrea on June 17, 2008 at 11:07 am said:

    “At the same time, just because one factor is important doesn’t mean the others aren’t….”

    Yes, right, and the factors interact. That’s teh point. That’s why amelioration of human problems is so immensely difficult.

  • JillB on June 17, 2008 at 4:08 pm said:

    This conversation moves the point of intervention to prenatal – which is certainly a valid point. A comprehensive approach is needed for sure. And while we are discussing this, intervention for teenage girls to avoid pregnancy is another key opportunity. But then, did we just go full circle back to equal opportunity programs?

    The key, in my humble opinion is to address the problem holistically, not with any one point of intervention. I commented earlier about establishing strong learning centers in the communities most needing them – and then addressing everything from teenage pregnancy, parenting, education of parents, family, etc.

    A comprehensive approach to one community, rather than a generic approach to touch a number of communities.

  • JillB on June 17, 2008 at 4:10 pm said:

    Actually, instilling hope is maybe the most useful path. However we can encourage hope instead of hopelessness and cynicism would be a good path.

  • Carl Shulman on June 18, 2008 at 12:07 pm said:

    “Certainly, but it’s hard to control for all the other factors that affect many African American teenagers who give birth to low birth weight babies, factors that are also the cause of low birth weight. It’s really a cascading effect, not a single cause-effect relationship.”
    Negative effects of low birth weight are seen across all ethnic and SES groups.

  • Jason R on June 26, 2008 at 9:23 am said:

    Holden, the reason that this isn’t discussed more is that a lot of this has to do with the parents. With parents who themselves suffered this syndrome and do not read well or are not available to their kids, well, the kids are not going to have a head start. I’m seeing this in my own life. My 2.5 year old is almost reading. Why? She’s been read to by my wife and me every day since she could sit up. She’ll probably be too advanced for kindergarten. I don’t know if she’s unusually smart, maybe, but I credit it to the fact that she gets the attention she gets and that we’ve focused so much on reading to her. I don’t know any charity in the world that can address that gap.

  • Andrea on June 26, 2008 at 9:06 pm said:

    Carl, It’s a cascading effect, not a single cause-effect relationship. And as Jason points out, not all the elements can even be remotely reached by any intervention.

  • JillB on June 28, 2008 at 12:19 am said:

    So, are there any programs that support a whole community, bringing both the parents and the children out of poverty and which give long-term hope?

    Are there and innovative actions that have been successful in changing these problems, even if they are one time programs?

  • Andrea on June 28, 2008 at 2:05 pm said:

    I have read about some programs where families are moved out of the poor neighborhoods they live in and into middle class neighborhoods, but they only work when the families are carefully selected for high motiviation. There are also some programs that move families into mixed neighborhoods, but without careful selection the problems come along with the families, including crime and violence, and critical supports that the neighborhoods provided are no longer available. Read “American Murder Mystery” by Hanna Rosin in this month’s Atlantic Monthly.

Comments are closed.