Scott Sumner on Education

Reading time ~5 minutes

In a post titled Education and Marriage Sumner shall we say muses on education (and a bit on marriage, off topic for my purposes). His post underscores our inability to understand cause and effect in complex domains.

Marriage and education are highly correlated with positive life outcomes, in many dimensions. Many people believe the causation runs from marriage and education to better life outcomes. Some believe that policies that improve education and encourage higher marriage rates will lead to better societies. This is where I become a bit skeptical.

From this, two Outcomes for education that he ascribes to others:

Increase positive life outcomes. Lead to better societies

Sumner presents an Argument against Ideas that target the second preferred outcome:

The argument I’d like to make is that good educational systems don’t cause good societies; rather good societies cause good educational systems.

He offers some characteristics of a “good society:”

In other words societal quality is related to many variables beyond income, including social cohesion, low levels of violence, lack of tribalism, liberal attitudes in the broadest sense of the word (such as empowering women and allowing dissent), etc.

This provides a starting point for Criteria with which to evaluate policy impacts:

  • Higher social cohesion
  • Lower levels of violence
  • Absence of tribalism
  • Increased empowerment of women
  • Increased tolerance for dissent

He speculates why “intellectuals” might think that education can “help solve society’s problems?”

Partly because they are intellectuals.

Which seems like a weak ad hominem attack?

Partly because at the individual level better educated people do better on average, in many respects.

This seems an argument supporting society investing in education, if the preferred outcome is better individual outcomes in those respects, or if the preferred outcome is a better society and that depends on better individual outcomes.

Partly because the idea appeals to both liberals and conservatives. They might disagree about precisely how the educational system should be improved, but can at least agree that an improvement would significantly improve society.

I’m not sure who this leaves out. If no one, this can be restated as, “Everyone generally agrees that improving education is worthwhile, although they don’t agree on how to improve education.”

He remains skeptical of these arguments that education can contribute to bettering society:

I’ve never met anyone who’s life outcome seemed strongly linked to the quality of their schooling, particularly the parts of education that get all the attention (quality of structures, teachers, curriculum, etc.)

This assertion surprises me. I split my university time between the main campus and a branch campus of a state school. In general the instructor quality at the branch seemed lower than those on the main campus. In Computer Science the main campus had a functioning computer lab, and the branch didn’t, which impacted the quality of those classes there. Of course the school I attended had a much weaker staff than MIT, Stanford, and the UC schools that many of my professional peers attended. I’m not sure how strongly linked these differences were to my and their life outcome(s), but I think the differences mattered.

Nor have I ever participated in a discussion of another person not present (i.e. gossip) that attributed that person’s failings to the schools they went to.

My guess is people harmed by poor school systems don’t occupy much attention in some circles. Sumner seems to focus here on low quality schools, and anecdotal evidence that no one he knows or has heard discussed has achieved lower life outcomes as a consequence of the poor quality of the schools they attended. I don’t think it’s hard to find evidence that better quality education leads to better life outcomes on average:

  • Studies comparing for-profit universities with public and private universities.
  • The impact of the GI Bill on the post World War 2 economy (assuming no school is lower quality than school).

It’s always personal qualities, or a tough home environment.

Given our bias to explain contingent outcomes with stories that make those outcomes seem correct and inevitable, it’s not surprising that “gossip” about others blame their failures on their character flaws.

More importantly this signals the challenge in understanding cause and effect in complex self-adaptive systems (where participants are lightly constrained, have their own intentions and interpretations, and influence and respond to the actions of others). Education outcomes depend on more than school quality (not that we know how to measure that anyway). The strongest predictor of life outcomes isn’t academic test scores; it’s Adverse Childhood Experiences (ACE).

I can imagine that there might be cases where education was to blame, but typically those cases are associated with attending schools with a climate of fear, a lack of respect for learning, and a non-supportive home environment. I.e., things that are not easily reformed through public policy.

Here we move from the non-education influences on outcomes to an assertion that those fall beyond the practical reach of public policy. Unsupportive parents, unsafe schools, antiintellectual cultures. They exist, and they can’t be changed, so they must be accepted, as must the impacts they have on others we would like to support if only we could?

FWIW, one reason I support vouchers is that I believe that private schools would be less likely to be dominated by a climate of fear. So I’m not saying policy has no effect.

Great–an Idea:

Use vouchers to support private schools.

With the interesting supporting argument:

Private schools would be less likely to be dominated by a climate of fear.


comments powered by Disqus