79 Comments

Commenting has been turned off for this post
Slaw's avatar

A little off topic but I want to rebut the idea that the finish line for educational research and reform should be getting everyone into college and an "elite" career. That is economically illiterate. A healthy economy has a wide diversity of work and not all of those jobs are going to be in the field of computer programming. Blue collar labor is just as necessary to a well-functioning society as white collar labor. It is no failure if somebody doesn't want to go to college--that is a simple and basic diversity that society as a whole should understand and embrace.

What is remarkable to me is the belief that only white collar work should pay well and the blue collar work should naturally pay poverty wages. Why does this paradigm go largely unquestioned/unchallenged? It's factually wrong for the skilled trades and the worship of concepts like "creativity" and "innovation" has become practically fetishized.

Expand full comment
Carina's avatar

Really great post. “Hard science” folks like your friend will always be skeptical of social science, and a lot of that skepticism is deserved – but these are important questions, and there’s no good alternative, especially because we’re largely talking about quantitative outcomes (GPA, test scores, class rank, college admissions and performance, job salary, etc.) Ethnography can suggest hypotheses, but it can’t tell us what works at scale.

I encountered a lot of the same prejudices against statistics in grad school, mostly from people who specialized in inequality (race, gender, and LGBTQ studies). They would convince themselves it was a hate crime to even categorize people for quantitative analysis. For example, a checkbox for race/ethnicity = oppression because it reduces complex personal identities to a few categories. I would often point out that we wouldn’t even be able to talk about inequality without categories and numbers.

For me, one of the most feasible reforms is increased & improved use of randomization. We need greater tolerance for the inherent unfairness of randomizing interventions in the short term so that we can help everyone in the long term.

Another feasible one is publishing null findings--because this can be accomplished without billionaires. The discipline could come together to support a real journal for null results, or (even better) prestigious journals could dedicate a certain % of each issue to null results. A small group of big names could make this happen if they really wanted to. But they’re probably afraid of younger scholars publishing papers that disprove their pet theories.

Finally, more funding for schools to conduct their own quantitative research (in collaboration with professional researchers). Schools have access to their own student data, the ability to conduct new surveys / assessments, and the desire to find out what actually works in reality. I haven’t looked at the research closely enough to speak to the quality, but the concept behind the U Chicago Education Lab (partnering researchers with school districts) seems promising – at least, it’s better than school districts implementing expensive interventions without a plan to study the results that accounts for the many threats to internal validity.

Expand full comment
77 more comments...

No posts