The Better Government Lab - a joint research center at @McCourtSchool & @fordschool - has a fresh website.

www.bettergovernmentlab.org

You can now follow us on LinkedIn --> www.linkedin.com/company/bett...

@pamherd.bsky.social @donmoyn.bsky.social @jbarofsky.bsky.social @giannella.bsky.social

Gorgeous! Meanwhile on the eastcoast

Coding people's experiences w applying for #SNAP --this anonymous response depicts the Kafkaesque application maze some applicants find themselves in.

Better Government Lab retreat at @fordschool - excellent 2 days full of research and org strategy!

The 🥯is rolling…

@donmoyn.bsky.social @giannella.bsky.social @pamherd.bsky.social @jaeyeonkim.bsky.social @jbarofsky.bsky.social

And this is exactly what we found across (almost) all our pre-registered outcomes!

signaling good test scores (vs no info). This is a more clean test than high vs low performance bcs our theory of discrimination assumes that the absence of such information leads to discriminatory practices bcs ppl use statistical stereotypes abt black students w/o direct info.

-Monetary incentives did not increase vaccinations among those who received incentives
-Cohabitants of individuals who received incentives were 0.3 percentage points LESS likely to get a booster!

(10/16)

-1 individual in half of all households gets a letter that offers ~$23 for getting vax’ed at the events and ~$23 more if at least 900 residents get vax’ed across the four events

(9/16)

Experimental design:

-All ~40,000 residents (in 10,000 households) of Ravensburg, Germany were mailed letters with information abt four upcoming, public vax events

(8/16)

In short, we propose adding step 5 (in blue) to existing evidence-based policy-making processes.

(4/16)

New working paper with

@johnt1.bsky.social Florian Keppeler @dominikvogel.bsky.social

“Scaling Behavioral Interventions in the Presence of Spillover: Implications for Evidence-Based Policy Making”

osf.io/k5uap/

PAsky polisky econsky policysky scipol science4policy

/Thread

Local policy-makers strongly overpredict the efficacy of 💸!

WHY?

We suspect one problem is a lack of policy education about the potential pitfalls of financial incentives.  For instance, when it comes to degrees typically pursued by German policy makers:

Here’s how policy makers predictions stack up against the actual effects of 💸:

- Prediction: Mean of 15pp; over 80% expected a positive effect
- RCT: Precise null

We used the same materials in a city-wide RCT (N=41,548 individuals in 10,032 households).

One member per 5,016 households randomly selected to receive 20 (+20) Euros incentive, or not.  

(Designed to measure spillovers... more on this in an upcoming thread.)

German local policy-makers (N=88) predicted the efficacy of 💸 incentives of 20 Euros (+ an additional 20 if city-wide vax goal is reached).

Here are the materials they evaluated (and which we also used in the RCT):

This is not to say 💸 does never work, just that it is not a silver bullet. A recent meta-analysis for COVID vaccination shows substantial variation in efficacy of 💸 to motivate COVID19 vax, small fx, and similar fx sizes as nonfinancial interventions.

tl;dr:

-Prediction study among local policy-makers (mayors, etc.) evaluated same materials in RCT testing whether financial incentives increase vaccination
-Policy-makers predicted 15pp increase, BUT the RCT found a precise null
-We discuss why, and what to do about it

📣 New WP alert 📣

osf.io/jq28n

“Policy makers believe money motivates more than it does”

With Flo Keppeler, John Ternvoski, Dominik Vogel & Erez Yoeli

/Thread 🧵