3 Comments
User's avatar
Sebastian Garren's avatar

Hmm... I'm not sure what you are claiming the greatest education intervention is. Surprisingly, removing absenteeism is not, in fact, what the literature ranks as the strongest education intervention. In the big meta-analyses of Rachel Glennerster[1] and the RCTs of Duflo[2], the best bang for buck in classroom learning is tracking and ability grouping, outflanking teacher absenteeism, class size, and teacher training by quite a margin.

But you know how it is. Government capacity to make or even allow good interventions to occur is often weak. But almost all developing countries have a private school system, and those systems do frequently and reliably can outperform the government system as you say[3]! But is ability grouping likely to fail? I don't think so. You only need the administrators and teachers at the school to get on board. That seems to be a job which NGOs could do successfully, without stepping on any government toes or regulations.

[1]https://www.aeaweb.org/articles?id=10.1257/aer.101.5.1739

[2]https://openknowledge.worldbank.org/bitstream/handle/10986/34658/How-to-Improve-Education-Outcomes-Most-Efficiently-A-Comparison-of-150-Interventions-Using-the-New-Learning-Adjusted-Years-of-Schooling-Metric.pdf

[3] https://econjwatch.org/articles/big-questions-and-poor-economics-banerjee-and-duflo-on-schooling-in-developing-countries

Expand full comment
Craig Palsson @ Market Power's avatar

Thanks for the articles. I'll clarify that I was speaking to this intervention in total as the greatest education intervention, and it included several treatments. I think the teacher training/monitoring and the language instruction were the most important features. But the sum total of the interventions led to the biggest gains I've ever seen.

On ability tracking, I agree that it is an incredibly underrated intervention. A recent article even evaluated how it performed in Texas and found "positive effects on high-achieving students with no negative effects on low-achieving students" (https://www.nber.org/papers/w30370). Strangely, the authors concluded that was a bad thing because it increased learning inequality. And I think that's why it's a politically tough sell for wider adoption: you're labelling who is learning quickly and who isn't. Silly that we let that stop our kids from learning the most possible! I think I'll write a follow-up post elaborating on these ideas.

Expand full comment
Sebastian Garren's avatar

Good article. Thanks for pointing me to it. I didn't take the authors' comments about increased inequities in learning outcome as negative, except insofar as the word "inequity" is a synonym for "bad."

They do point out some secondary effects which point towards helping the bottom 25% of the distribution and have a nice 'ring' to them: smaller class sizes, decreased sorting by race or SES.

Expand full comment