Why some black men succeed in college

Black males who do well in college have parents — and at least one K-12 teacher — with high expectations, concludes the National Black Male College Achievement Study.

Black male achievers typically come from working-class families, concludes Shaun Harper, an associate professor higher education at Penn who founded the Center for the Study of Race and Equity in Education. Nearly half have parents with no college degree. “As a group they shun the idea that they are cognitively smarter than their less-successful friends or cousins or other peers (and their high-school academic records largely back that up),” notes Inside Higher Ed.

In addition to parents who considered college a “non-negotiable” goal, and a teacher who took a special interest, achievers had adequate financial support to pay for college and support from black juniors and seniors when they started college.

Sixty percent grew up in homes with two parents. “Census data show that 35 percent of black children grow up in two-parent homes,” reports Inside Higher Ed.

Harper asked each of the 219 black men to talk not only about themselves but about the experiences of their three best black male childhood friends — and these differences virtually jump off the report’s pages.

“When asked what differentiated their own paths from those of their peers who were not enrolled in college, the participants almost unanimously cited parenting practices,” the study states. “Their friends’ parents, the achievers believed, did not consistently maintain high expectations and were not as involved in their sons’ schooling. By contrast, most of the achievers’ parents and family members more aggressively sought out educational resources to ensure their success — tutoring and academic support programs, college preparatory initiatives, and summer academies and camps, to name a few.”

Like the well-to-do parents in the preceding story, the black male achievers’ parents invested in their children’s success.

About Joanne


  1. That’s kind of a no-brainer for those of use that do have high expectations for our kids. I can’t wrap my head around the parenting of those that don’t.

  2. @Andrea: Agreed.

  3. Is this surprising? Doesn’t this address the success of charters that are chosen by motivated parents. Is there anything that society and schools can do to promote or improve on these facts?

    If not, how much credit/blame do schools and teachers deserve?

  4. Yeah, motivated by how poorly their kids are doing in the lousy district schools.

    Schools and teachers are the symptoms, not the disease. The disease is the structure of the American public education system which takes no notice of whether there’s much educating going on.

    The underlying assumption’s that the school board members who get elected will be elected because the ensure/promise to ensure, that schools and teachers will stay focused on education. But it’s a silly assumption from the get-go.

    By overseeing public education via the democratic process politics becomes, inescapably, the most important consideration and education the least. That situation will only change when parents decide whether schools prosper or die on the basis of their patronage.

  5. A study of the obvious.

  6. Richard Aubrey says:

    It may be obvious, but it’s also subversive.
    It suggests that certain lifestyles have different results than other lifestyles, which is being judgmental. I have a relation who says, wrt one thing or another, “It’s their culture,” clearly thinking that means the results we see don’t really happen, or the results we see are really okay.
    If we want to ithncrease the number of folks from this SES/ethnic group, something will have to be done about the culture.
    You know those stats telling us that black kids whose parents are in the top income quintile do worse on ACT or SAT than white kids whose parents are in the lowest quintile? Yeah. Those. It’s not likely a heritability thing, since the black parents in question are in the top quintile and you don’t get there by being dumb.
    It’s something else. Ogbu, looking a Shaker Heights, found black kids there watched a huge amount more television than white kids, in order to keep up with the pop culture, the discussion of which was apparently a social issue. Got to keep up or you’re a loser.
    Said it before: I have black clients who are so straightarrow they squeak when they walk. Although Pentacostals, they sent their kids to a Catholic school up through the eighth grade before exposing them to pub ed, which in their district included some low SES including but not limited to black families. I figure, these folks know what they’re doing. Got to be a reason.

    • Mark Roulo says:

      “You know those stats telling us that black kids whose parents are in the top income quintile do worse on ACT or SAT than white kids whose parents are in the lowest quintile? Yeah. Those. It’s not likely a heritability thing, since the black parents in question are in the top quintile and you don’t get there by being dumb.”

      If you believe (a) in IQ tests [which consistently show that American blacks, on average, score 1 standard deviation below American whites], and (b) that IQ has a non-trivial genetic component, and (c) in regression towards the mean [which, if you believe in (a) and (b) you *should*], then, sadly, this *will* have a large heritability component. I don’t think that heritability explains all of it, but heritability explains a lot of it.

      Using very crude numbers, the top quintile will be roughly 1 std-dev above average. But this is 2 std-dev above the black average. Regression to the mean would mean that black parents in the top quintile (and they will mostly be close to the edge of that break) will have kids with IQs around average (still 1 std-dev above the black average).

      White parents in the bottom quntile will be about 1 std-dev below average, and their kids will regress to about 1/2 std-dev below average.

      The 2 std-dev gap between the parents turns into (on average) a 1/2 std-dev gap for the kids. This doesn’t quite work out to the white kids from the bottom quintile outperforming the black kids from the top quintile, but genetics have gotten us pretty close.

      The good news is that if you don’t believe (a) that IQ tests measure anything important, or (b) that IQ is genetic (or has a large genetic component), then none of this matters. And you only need to disbelieve either of these, not both!

      The bad news is that if you believe (a) and (b), then a huge chuck of this *IS* explained by genetics.

      The worse news is that even if you believe that this is totally environmental/cultural , we don’t have any good ideas (as a society) for how to “fix” a sub-culture.

  7. Deirdre Mundy says:

    Question– I thought “regression to the mean” was about POPULATION LEVEL Genetics, not individual mating pairs.

    So, imagine a group where the IQs are 1,1,9,9,5,5

    The average for the group is 5.

    Some possible matings – 1+9 = average kid
    1+1 = below average
    9+9 = above average
    5+1 = slightly blow
    9+5 = slightly above

    If 9 marries 9, and each 1 marries a 5, we get next gen IQs( over simplifying )

    of 9 9 3 3 3 3 So the average is still 5., even though the kids of the above average parents are still above average.

    In the next generation, the population is too small for the 9’s to selectively mate, so:

    9+3 = 6, 6
    9+3= 6, 6,
    3+ 3 = 3, 3

    Average is still 5….

    (Yes, this is oversimplified, but…….)

    I’m pretty sure regression to the mean does NOT work for individual pairs. How many brilliant couples do you know with average kids?

    How many couples with a tall mom and dad have short offspring?

    • Deirdre Mundy says:

      After some fun with google, I see that regression to the mean only applies to RANDOM occurances… so, again, I’m not sure if you can make the case for individual pairs— and it’s for extreme outliers. If two smart parents have a smart kid, is he really an outlier?

      • Mark Roulo says:

        The kid would be an outlier, just not a very large one. Two 2 std-dev parents would have a 2 std-dev kid 1/6 of the time. This is slightly unlikely, but not worth getting super excited about.

        This assumes that the unexpressed genes are random for the population. If they aren’t (and Charles Murray would argue that they are becoming less random due to associative mating), then the odds are even better (and, of course, even *worse* for the not-smart counterparts).

    • Mark Roulo says:

      The way this is supposed to work (and I am not a biologist of any sort) is that people have lots of genes that influence various traits. But only 1/2 of those genes get “expressed”. So … imagine that we have a bright person (say, IQ 140 or so). The genes getting expressed for this individual work out to an IQ of 140. But this person has just as many genes *not* getting expressed. What do these genes work out to? If we know nothing else, the best guess is that those unexpressed genes work out to the average IQ for whatever population the person drew those genes from. For random white people, that population has an IQ of about 100.

      When having kids, the kids get a random mix of genes from both the expressed genes (IQ 140) and the unexpressed genes (IQ 100). The expectation, then, is that the kids will run about 120 for IQ. Will there be lots of exceptions? You bet.

      But this is still how regression to the mean works for genes.

      It works the same way in reverse for not bright people.

      And if we know lots about the person’s family (e.g. siblings, biological aunts and uncles) we can form a better guess as to the “population” of the unexpressed genes.

    • Mark Roulo says:

      Last bit … regression *TO* the mean should more accurately be labelled regression *TOWARDS* the mean. Bright parents are unlikely to have “average” kids, but the brighter the parents, the less likely that the kids are going to be just as bright as they are.

      One example doesn’t prove the point, but Marie and Pierre Curie had kids. One of them, Irene, won a nobel prize too, but the other was much less distinguished (as an author). I’d be reluctant to call the second kid a failure, but not in mom and dad’s league.

      Irene’s children are scientists (both have PhDs, I think), but not in Mom and Dad’s (or Grandma and Grandpa’s) class. I expect *their* kids to be even more “normal”.

  8. Richard Aubrey says:

    I get all that. Problem is, regression to the mean doesn’t explain such a large drop from the folks who got to the top quintile. It would be closer to regression to an outlying bottom.
    Thus, culture is the issue and, no, there’s no easy to fix a culture, especially one protected by PC.
    I heard that regression to the mean explains why a student pilot who makes a lousy landing will do better next time. Same reason a student pilot who makes a fantastic landing will do less well next time. It was in a manual for instructors. Encouraging the guy who makes the lousy landing, or screaming at him, both work. Either way, the instructor thinks he did good.

    • Mark Roulo says:

      “Problem is, regression to the mean doesn’t explain such a large drop from the folks who got to the top quintile. It would be closer to regression to an outlying bottom.”

      The reason that this explanation “works” is that the two populations are regressing to different means.

      The black parents in the top quintile are 2 std-dev above their population norm, so their kids move 1/2 way back to this norm, which is 1 full std-dev.

      The white parents in the bottom quintile are only 1 std-dev below their population norm, so their kids move 1/2 way back to this norm, which is 1/2 of a std-dev.

      We see this in lots of populations. Super-smart people tend not to have super-smart kids … although they are still more likely to have super-smart kids than random people from the population. Super-star athletes tend not to have super-star athletes … but are still more likely to have super-star athletes than random people from the general population.

  9. Actually, regression to the mean coupled with the fact that we’re not talking about salaries over $150K but rather $75K means that we’re talking about government jobs for African Americfans, which reward pretty well without requiring extremely high levels of g.

    So black couples might have an IQ of 95-11, making $75-125K year. Quite possible in government and union jobs of the past. Their kids could easily have IQs of 85-95, which is still average to slightly above average for African American IQs.

  10. Richard Aubrey says:

    Cal. That it takes $75k to make it into the top quintile doesn’t mean that’s what the blacks in the top quintile make. Given various aspects of history–which I have termed “Historical Accumulation”–it’s possible they are underpaid by a few bucks if their basic IQ is considered in isolation.
    For example, I dealt with a social worker who worked at a “changing” high school. No matter what the admin did, eventually all non-sport extra curricular organizations ended up run by white kids. Year after year. His view was that the black kids just flat didn’t know how to work in an organized group. I was introduced to Roberts’ Rules in, I think, the fourth grade and we even had a little student council, of which I was some kind of functionary, iirc.
    Later on, at a black college, I saw how they did it. The loudest, those who interrupted best and who resisted interruptions best, were left with the microphone and an empty room. They, quite literally, had never heard of Roberts. Or any form of order. This is an example of how HA can handicap someone who has even above-average native intelligence.
    I have a relation who works with a woman of average or slightly below average intelligence. She came from a disfunctional family where actually using your mind to solve problems was a waste of time because the family was so unpredictable and counterproductive that nothing you bothered thinking about ever showed up as you thought it might. And the parents and older sibs might eff you up just for spite.
    Thus, it never occurred to this woman to look at prices and stop doing her grocery shopping at a convenience store/gas station. Once my relation went over the math with her, a couple of times, she caught on. If the rules of your particular universe are not consistent, there’s no percentage in trying to think about it. Or even wondering if you should. Point is, culture matters, and not just in the big stuff.

    Thomas Sowell, writing on culture had a number of examples, of which I remember two: Brazilians of Japanese ancestry whose forebears came from the northern islands still tend to marry those whose ancestors came from the northern islands, and not even the hotties whose ancestors came from the southern islands. A mirror image for the folks whose ancestors came from the southern islands. And nobody can put a finger on the reason.
    In the eighteenth century, the Czar of All The Russias invited German farmers to immigrate, prizing their advanced methods and work ethic. By the end of the niineteenth century, European Russians included two percent of German ancestry, while the German-Russians were forty percent of the officer corps. I may be off a few points, it’s been a while, but the point would stand anyway.
    However, I’d be interested in any empirical evidence that blacks in the top quintile are largely in the bottom decile of the top quintile.

    • Mark Roulo says:

      “However, I’d be interested in any empirical evidence that blacks in the top quintile are largely in the bottom decile of the top quintile.”

      It isn’t empirical evidence, it is math. But it also isn’t quite this extreme.

      *IF* you assume that IQ tests are “real”, then we have that:
      (a) scores for various populations tend to be gaussian (more or less) around the mean for that population, and
      (b) the mean for the American white population is about 100, and
      (c) the mean for the American black population is about 85.

      Remember, we can duck all of this if we just decide that IQ tests don’t measure anything real.

      So … simplifying a bit, the top quintile is the same as 1 std-dev from the mean (ie, about 1 std-dev over 100). About 16-20% of the white population falls into this category and about 2% of the black population falls into this category. But because we are dealing with the tails of a gaussian, most of each population is nestled close to the mean. By the time you get out to 2 std-dev from 100, we have about 2% of the white population, and 0.15% of the black population. note that 2/20 is larger than 0.15/2 … this helps to illustrate how the shape of the tails of a gaussian fall.

      In table form, the white population in the top quintile breaks down as 16% above 1 std-dev, and about 8% also being above 1.4 std-dev. The black population beaks down as 2.5% above 1 std-dev and about 0.8% above 1.4 std-dev.

      So about 1/2 of the white population is in the bottom half of the top quintile and about 2/3 of the black population is.

      This is all based on IQ and heritability assumptions. If those are not valid, then none of this math matters.

  11. Richard Aubrey says:


    I was referring to income.

    • Mark Roulo says:



      I was still on the demographic could it be or not thread.

      I have no data for black income distribution … The census might cover this?