Here is the job ad:

We do not find evidence that psychotherapy is used as a substitute, suggesting many women are effectively untreated for mental health issues during this critical period. Here is the link to the full paper with more details: https://t.co/UAod71UJvP

In a large US sample, approximately half of all women discontinue antidepressant usage during pregnancy (no change in usage for spouses). This is in a new paper that I just published with @claireeboone and @ccolina_pe that is primarily descriptive.

New Working Paper. LEGO movie summary of main result 😀 For more results see NBER Working Paper ( https://t.co/tfYVgRRr7P). Amazing team: @campos_mercade @meier_armando @Devin_G_Pope @f_h_schneider @ErikWengstrom

I wonder if in-game betting odds right as the 3rd quarter ends are properly pricing in who will have the ball at the beginning of the 4th or if there is inattention to the possession arrow.

The correct answer is 5%+! This went against my intuition (and it looks like most of yours too), but the value of 1 possession in the 4th quarter of a close basketball game is bigger than most people think (5-6pp).

I did a podcast with @LeadingSaints. If you follow me and you are LDS, you might enjoy it. If you follow me and are not LDS... haha, it will make no sense. https://t.co/2KAemcyBio

Consider an NBA basketball game that is tied going into the 4th quarter. How many percentage points more likely to win is the team that gets the ball first to start the 4th quarter?

As always, it was super fun to work with a great set of coauthors: @campos_mercade, @meier_armando, @meier_steph, @f_h_schneider, and @ErikWengstrom.

Many other results such as the impact of individually-targeted incentives, long-run vs. short-run effects, unintended consequences (none!), and family spillovers can be found in the paper. https://t.co/TYh9Nsxz0s

I'm excited to release a new working paper today. When we paid people ~$20, they were 13pp more likely to take the Covid Booster shot. This guaranteed financial incentive worked better than lottery and charitable incentives in our large, pre-registered field experiment.

Of course, the best scenario would be that researcher's findings do not predict political or other attitudes a priori, but then do correlate some with their beliefs ex post due to researchers changing their minds!

Of course, it might be the case that research results DO predict personal beliefs quite accurately (e.g. finding evidence that gun control is effective predicts someone who is pro gun control). I hope that is not the case, but worry that it is...

I'm surprised by how people make assumptions about me based on my research. For example, many people have made comments about how I must hate religion, which if you knew me you would know is far from the truth. Has this happened to others?

Oh, and I just want to thank @lymanstoneky and many others once more for engaging with the paper (both praise and critiques). There is nothing more fun than having people pay attention to something you've been working on for a long time.

Overall, my view is that the cellphone data does a reasonable job of predicting visits in a variety of settings and gives us a useful estimate for visits to places of worship - even though there are clear limitations (just as surveys also have clear limitations).

It is totally fair to argue that places of worship are different than Olive Gardens. Church walls might be thicker, people use their phones more at dinner than at worship, etc. These are all reasonable concerns and something that will always be hard to fully address.

One quick note is that part of the reason for the underprediction (explains ~50% of the underprediction) is that the data does not contain the total number of AMC theaters as "Places of Interest". This doesn't have to do with phones pinging, but just a data construction issue.

The cellphone predictions are quite accurate for many business establishments. The estimates are clearly too high in a couple of cases (e.g. Target) and clearly low for at least one (AMC Theaters).

In the paper, I compare predicted visits from the cellphone data to actual visits in a variety of locations. Here is a screenshot of the table: