YOU SHOULD BE ASKING QUESTIONS

Baltimore has had success in lowering crime by providing youth opportunities instead of relying on punitive measures. Why aren’t DC leaders interested in this? We should be learning from Baltimore and trying new approaches.

Your gentle, periodic reminder to not platform people who should not be platformed.

How Teachers Can Judge the Credibility of Research (Opinion) As a teacher, your time is limited. Don't waste it on programs that only serve the interest of the companies selling them.

Look, causal inference studies can be good or bad regardless of where the person who produced it works, what field their degree is in, and whether they have a PhD. You have to actually look at what they did. There are no shortcuts.

www.edweek.org/teaching-lea...

There is huge underemployment. Most former feds I knew quickly moved to part-time or consulting work, which you wouldn’t see in unemployment, but also doesn’t pay the bills.

And, last but not least, this paper uses @jepusto.bsky.social's new metaselection R package, which allowed us to examine publication bias for main v subgroup effects in the same model.

Subgroup Effects, Compositional Effects, and Treatment Effect Heterogeneity in the What Works Clearinghouse Study Data Education researchers are increasingly interested in understanding “what works” for whom and under what conditions, often through analyses of subgroup findings. Yet such findings are frequently rep...

Only for those who can't get behind the paywall, you can access the paper here: www.tandfonline.com/eprint/IFK2Q...

a man in a white shirt and tie is making a funny face . ALT: a man in a white shirt and tie is making a funny face .

In sum: Be curious and cautious! Avoid overgeneralizing from average or subgroup effects. At minimum, studies should report confidence intervals and treatment effect heterogeneity, and examine subgroup results with both empirical evidence and a clear theoretical rationale.

Evidence suggests that educational programs *can* help students facing systemic constraints—especially when programs are designed to expand their opportunity to learn. And yet, we did not find effects for other subgroups and cannot rule out all measurement artifacts.

a man wearing a red shirt with the words it 's just on it ALT: a man wearing a red shirt with the words it 's just on it

Finally, we find substantial treatment effect heterogeneity—even within the same study and outcome domain—which begs the question: when is the average effect not sufficient for decision-making?

We do not find evidence of compositional effects across studies once we account for intervention type and delivery method. We also do not find systematic variation in effect sizes by grade level or clear patterns by years of implementation (though our measure there is limited).

The subgroup effects are not explained by outcome differences or publication bias, and are unlikely to reflect ceiling or floor effects. We cannot rule out whether subgroups have less variation in outcome scores, which would make effect sizes appear larger without reflecting larger true effects.

We find larger positive effects for economically disadvantaged (0.06 SD) and female students (0.05 SD), relative to main effects. With an average main effect size of about 0.15 SD, that’s roughly a 40% and 35% increase in magnitude, respectively.

a man wearing glasses and a tie is sitting in front of a computer and says what is going on ? ALT: a man wearing glasses and a tie is sitting in front of a computer and says what is going on ?

Too often in ed research, we report subgroup findings without examining why they appear. Are they real or spurious? We set out to explore this. Using WWC data, we examine subgroup effects relative to main effects and find larger positive effects for economically disadvantaged and female students.

Subgroup Effects, Compositional Effects, and Treatment Effect Heterogeneity in the What Works Clearinghouse Study Data Education researchers are increasingly interested in understanding “what works” for whom and under what conditions, often through analyses of subgroup findings. Yet such findings are frequently rep...

New paper from me and Chris Klager. While at the What Works Clearinghouse (WWC), one question came up repeatedly: are educational programs differentially effective for different subgroups? The short answer: some subgroup effects may be real, but caution is needed. www.tandfonline.com/doi/full/10....

Colorized illustration of Black Washingtonians celebrating DC Emancipation Day in 1866 from Harper's Weekly.

Happy DC Emancipation Day!
 
Today, we honor the anniversary of the abolition of slavery in the District and the liberation of more than 3,000 enslaved people.
 
We also reflect on the ongoing work to build a DC where freedom is fully realized—through equity, opportunity, and dignity for all.

Presidential Essay: AEFP’s Role in an Evidence-Informed Education Ecosystem Historically, the U.S. federal government has invested in education research as a public good, with the implicit assumption that such research can support more efficient and effective educational oppo...

Presidential Essay: @aefpweb.bsky.social's Role in an Evidence-Informed Education Ecosystem via @efpjournal.bsky.social
direct.mit.edu/edfp/article...

Bar graph showing the reenrollment rates for program and control group students--21% and 11%, respectively--a difference of almost 10 percentage points

New RCT findings from MDRC on a program to reenroll stopped-out community college students within 15 credits of a degree

The program, REACT, includes outreach, advising, and tuition waivers

Early findings show it nearly doubled reenrollment, from 11% to 21%

www.mdrc.org/sites/defaul...

New rules mean many DC restaurants are closing streateries The District made streateries permanent. But restaurant owners say the rules are too costly and complicated to keep them up.

DC implemented a new policy for its streeteries that cost restaurants $20,000 to implement.

The result: virtually every restaurant eliminated their outdoor dining, many have reduced their hiring, & now a few cars can park more easily.

Congrats on the successful policy change, DC government!

If you've spent time teaching, you know that class size matters. Larger classes make it harder to teach and manage the classroom. That's a different question from what is the average effect of class sizes on test scores.

I don't remember seeing an image of Earth before that captures the (thin, fragile, tiny layer of) atmosphere so clearly. I mean, look at that.

Interesting that the ad doesn’t cite any specific accomplishments 🤔

Deliver YouTube video by Opportunity DC IEC

The pro-business group Opportunity D.C. (which has gotten funding from developers, philanthropist Katherine Bradley, and others) says it will spend $1 million on mailers and TV ads backing @kenyanmcduffie.bsky.social's mayoral campaign. The first ad dropped today: www.youtube.com/watch?v=_wmA...

A New Approach to Algebra in 8th Grade Seems to Produce Big Benefits Middle schoolers who took grade-level math and Algebra 1 together benefited, a study finds.

"The students who took two math periods saw big gains on 8th grade state tests over similar peers who were in 8th grade math alone, equivalent to almost a full year of additional learning" www.edweek.org/teaching-lea...

Delighted to share our revised paper (w/ @juliaturner.bsky.social & @jacobbastian.bsky.social ) showing that Universal Pre-K has broad effects on local labor markets — effects well beyond only mothers. Larger effects for full-day programs. nber.org/papers/w33767

#EconSky #EarlyChildhood #ECE

I agree with this. One of the best things I did last year was talk to journalists. We should do it more often! But it’s also incredibly insulting to place blame on the field for the damage caused by this Administration. People don’t have time to advocate when their basic needs aren’t being met.

4 days in office

An interestingly timed post considering that the Trump administration is trying to remove the 15th Street bike lane in part because it says it causes more traffic for cars.

OH in-person requirements for DC employees

Streateries, tipped min wage, ...