Survivor bias and the 92% of Southern Black men who support spanking

In today’s New York Times both Michael Eric Dyson and Charles Blow write about spanking. Blow doesn’t mention race and the South, but that’s in the background when we writes:

I understand the reasoning that undergirds much of this thinking about spanking: Better to feel the pain of being punished by someone in the home who loves you than by someone outside the home who doesn’t.

Dyson goes further, and ties the practice back to slave plantations:

Black parents beat their children to keep them from misbehaving in the eyes of whites who had the power to send black youth to their deaths for the slightest offense. Today, many black parents fear that a loose tongue or flash of temper could get their child killed by a trigger-happy cop. They would rather beat their offspring than bury them.

Here are a couple of logical points, and then some data.

First, please note that the rationale some Black parents use doesn’t need to explain all of the practice of beating children, just the difference between Blacks and Whites. Blacks are more likely to support spanking than Whites, but a strong majority of both groups in this country agree spanking is “sometimes necessary.” So not every case of Black parents beating their children is attributable to slavery and racism. Some may be, and the rationale no doubt is in many cases, but that’s not the whole story.

Second, it’s common for people who suffer some disadvantage and survive to attribute their survival to the hardship they suffered. NFL player Adrian Peterson, who beat his 4-year-old son with a stick, said, “I have always believed that the way my parents disciplined me has a great deal to do with the success I have enjoyed as a man.”

If there were 15 boys on a lifeboat, and one survived, he would probably say, “I have always believed that my lifeboat experience has a great deal to do with the success I have enjoyed as a man.” If he were successful in his post-lifeboat life, people might agree with his explanation. In fact, statistics might even show that lifeboat survivors are more successful. Statistically, that’s a survivor bias – the people who should be dragging down the average aren’t alive to weigh in.

A more subtle effect is not just statistical bias but real survival selection — the one lifeboat guy who lives was probably the strongest. So his story seems credible, even though lifeboat populations have very high mortality. In fact, the “Black-White mortality crossover” is a classic puzzle upon which many demographers (including me, though I sadly didn’t get it published) have cut their teeth: At old ages, Blacks have lower mortality rates than Whites (here’s a recent update). That’s partly because to live to old age in Black America you have to be tough (and partly because some old Blacks exaggerate their age, intentionally or unintentionally, which is a cultural expression of the same thing).


Anyway, kudos Harry Enten at 538 for turning to the General Social Survey to show trends in spanking attitudes. He shows that born-again Christians, Blacks, Southerners, and Republicans are all more likely to support spanking. And he did a regression showing those variables all predict spanking agreement when entered together. However, what he doesn’t show is the the interaction most important for today’s news: The support for spanking among Black men raised in the South. (Enten uses the GSS code for where people currently live, when for a question like this I think it’s more appropriate to use the code for where people lived when they were age 16.)

To get a decent sample size (this is down to 211 Southern Black men), I pooled three administrations of the GSS (2008, 2010, 2012), to get this:

spanking race and region.xlsx

Notice the huge gender gaps, which Enten for some reason didn’t consider.  And see that the Southern-at-age-16 people have higher rates of supporting spanking than the currently-Southern. If spanking were a reasonable adaptation to hardship, necessary for children to toughen up and learn to follow orders so they don’t get killed by Whites, why would Black men support it more than Black women?

So 92% of Southern Black men support a “good, hard spanking,” and Charles Barkley was probably right, empirically, when he said spanking was ubiquitous in the South in his childhood. But 75% of non-Southern White men support it, too. So it’s variations on a nearly-universal theme.

And the people who think it helps children because it helped them are not alone among the survivors of difficult childhoods. But that doesn’t mean they’re right.

Clarification: Don’t take the term “survivor” too literally. The lifeboat analogy is just an extreme version of, “15 people experienced harsh beating as a child, and one ended up a successful football player.” People who suffer and succeed often incorrectly attribute their success to their suffering.

26 thoughts on “Survivor bias and the 92% of Southern Black men who support spanking

  1. The concentration of sectarian Christians and fundamentalists in the South explains most of the regional gap, and much of the racial gap in support for corporal punishment, as Ellison and I showed a long time ago (1993, ASR).


  2. SURVIVOR bias!

    Please! We need some perspective. Live for 3 months in Africa or India, and understand what survival mans. Exaggeration and hyperbolism seems to bring in page views, but what you are saying is not true.


    1. The point is that people who feel they have succeeded in life look back on adversity and determine that it helped them, without considering all the people who experienced the same adversity and may have been harmed by it.

      The “survival” need not be literal; the post is importing a technical term from statistics to relate its argument about what constitutes good evidence about the effects of spanking to a larger point about evidence.

      Liked by 1 person

        1. Come on; some 90% of black men, in south or otherwise, think spanking is OK, and this result dos not depend on whether they had a “difficult childhood” or “success in life” or anything. Where is the evidence that “they survived a difficult childhood’? Is there a GSS variable that is missing here? Where is the evidence that those who had a difficult life and succeeded later in life? Your data point is Charles Barkley?

          “The point is that people who feel they have succeeded in life look back on adversity and determine that it helped them, without considering all the people who experienced the same adversity and may have been harmed by it.” Where is the evidence? 90% suggests that almost every black man, who has succeeded in life or not, or whether they faced adversity of not, or they have harmed by it or not, voted in favor of spanking.

          My point is black male support for spanking is not in anyway correlated to adversity in life, or success in life, or anything. I will not explain the reasons; Dr. Darren Sherkat should; he has published some 10 papers.


        2. “That’s partly because to live to old age in Black America you have to be tough (and partly because some old Blacks exaggerate their age, intentionally or unintentionally, which is a cultural expression of the same thing)”

          Where are you getting these things?


  3. I also was about to refer to the two Professor Sherkat articles, namely,:

    1. Ellison CG, Sherkat DE. 1993a J. Sci. Stud. Relig. 32, 313–29
    2. Ellison CG, Sherkat DE. 1993b Am. Sociol. Rev. 58: 131–44.

    and who decides to show up? Welcome Dr. D.S; I wish you explained more.


  4. Can I use this space to ask a few questions to the authority, DR. Sherkat?

    1. Why some sectarian Christians and fundamentalists only in the south support physical punishment? Why not even more sectarian Christians e.g., Mormons, Amish, Quakers, etc.

    2. Why the racial gap? We do not see a Hispanic preference for caning or physical punishment? Is there a “Macho” component?

    Is there any truth to survivor stories? I have lived in west Virginia which is as fundamental as poor as you can get, but caning or physical punishment is looked down there.


    1. You’re wondering why the Quakers don’t tend to support corporal punishment? Nonviolence is one of the central tenets of their faith, as is equality, lack of hierarchy, and consensus-based decisionmaking. They have been involved in corrections reform longer than anyone else, and they got massive influxes of hippie converts dodging the draft a generation ago. The Amish are also a “peace church.”

      Now, if the Mormons have lower numbers of support for corporal punishment, that might actually be surprising.


      1. yeah, quakers is a bad example. but nonviolence towards others, is not a deterrent to beating the child, as most Hindu or Buddhist children might know.


  5. I agree that the idea that childhood adversity leads to better adult outcomes is a silly ex post rationalization. Adversity is common, so it’s a common rationalization. Common rationalizations seem correct because they’re common.

    The idea that spanking is good for kids (i.e. makes them nicer, more respectful, etc.) reminds me of the 18th century argument that keeping the poor starving made them work harder.


  6. i want to know the definition of spanking. stern slaps on the butt? and “good, hard spanking” — is that gss terminology? what petersen did was a proper, maybe even extreme, whupping. does gss ask about degrees of spanking, or is it yes/no. do they define it? does anyone? enquiring minds want to know without having to go look it up themselves!


    1. That’s the whole question text in the caption to the figure. It’s strongly agree to strongly disagree that “it is sometimes necessary to discipline a child with a good, hard spanking.” One wonders how different it would be if they didn’t include “good” — which could be a term of moral acceptability or could be a modifier of “hard,” meaning, “really hard.”


  7. Male-female differences are much larger outside the South, both @16 and now, and for both Whites and Blacks. What’s up with that? Are gender-role differences generally smaller in the South?? How does this greater M/F similarity in the South fit with that old time religion that Sherkat mentions?


  8. As for the “I was whupped as a child, and I turned out great” idea, why do we pay any attention to it at all? Would any successful adult say, “I was whupped, but I’d be even more successful today if I’d been whupped more”? It’s also likely that successful people who were not beaten as children say things about their parents’ non-violent techniques that are similarly warm and approving. Does that prove anything?

    Liked by 1 person

  9. Given the known association between authoritarian parenting and low SES, I would really like to see how these spanking approval numbers break down by class, and see the regressions on that. Obviously class isn’t independent of race or being from the South. But we shouldn’t be leaving out such a major variable here.

    “My parents beat me, and I turned out just fine.” seems like an oxymoron. You didn’t turn out fine if you support violence against children, when all the evidence indicates it is ineffective for its stated goal of controlling behavior. Since it is ineffective, the only reason to do it is for the feeling of power it imparts. It’s the classic cycle of abuse– someone made to feel powerless as a child often feels the urge to take power back by abusing others. If someone says they can no longer beat children, that person is trying to take the abusers’ power away, and they react defensively with weak justifications like this one. Sure, there are worse forms of abuse, but that doesn’t make it OK.

    Liked by 1 person

      1. Vijay, yes. Annette Lareau’s book Unequal Childhoods talks about the parenting style of poor or working class parents being more authoritarian than that of middle or upper-middle class parents. But you could quickly go to the GSS and run a simple crosstabs for spanking or other measures of authoritarian parenting and, say, subjective social class or something like educational attainment. Also, the 2007 Baylor Religion Survey has a number of questions about parenting style that you could also run across levels of education. I’ve already checked and both these sources show a pretty clear relationship between more authoritarian parenting and educational attainment (proxy for SES). I didn’t look for income.


    1. Great idea, and it may even sound probable in 20th century, when people still were fighting against findings of high heritability of all psychological traits. Of course today we know that aggressiveness is highly heritable.
      People, who spank their children, have genes making them more prone to aggression. Their children receive those genes and that’s why they also spank. That’s why there is so little data on any long-lasting negative effects of spanking after taking heritability into account (if you don’t take
      heritability into account, your research on spanking effects will sound as 20th-century science, in other words is crap).

      [Rest of very long comment deleted. Feel free to submit a link to your writing for those who want to read more.]


      1. While kindergarten analogy was just teasing Alex, I admit, I thought the fact that spanking effects differ for minorities may be relevant for your entry. To repeat: for minorities even WITHOUT corrections for heritability, there is either no negative effect at all, or even _positive_ effect of spanking (I have provided a reference to scientific paper) – i.e. maybe there is no survivor bias, but impressions of black males correctly reflect the reality.


Comments welcome (may be moderated)

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