There are only two kinds of dichotomies.

Stop underselling your work. This may not be the first paper to study this policy with these data and empirical strategy. But it’s the first paper to estimate this policy effect for the seventh time.

Before the credibility revolution, everything was incredible.

Do questions in titles attract more attention? Evidence from you

I was hoping to trade this forthcoming paper for spices, but I have been told that it’s under embargo.

One of the most important roles of peer review is to chastise the authors for not thinking about power while simultaneously demanding heterogeneity analyses using 30% of the sample.

Is Claude or ChatGPT better at deciding whether a literature is vast, emerging, or burgeoning?

Trying to file my taxes using my ORCID.

We are excited to submit our paper for consideration to your journal (where we can just not think about it for 6 months).

A consulting firm which comes over to your house and tells you what that sound is.

Is Claude or ChatGPT better at expressing how much improved the paper is given the reviewer’s comments?

We don’t talk enough about how AI will rob future generations of reviewing a paper which misspelled “assess.”

Is Claude or ChatGPT better for asking editors about the status of your paper which has been under review for 6 months?

My strongest belief is that you should die at sea level, but it’s not a hill I will die on.

This insomnia study doesn’t replicate and people are losing a lot sleep over it.

This is the day that most authors like to write their “Data and code available upon request” statements.

Check out this interview with LDI Fellow @davidpowell.bsky.social on the Health Affairs "A Health Podyssey" podcast!

This paper provides critical evidence for policymakers to ignore.

This bag of lettuce has informed me that apparently lettuce can be used for salads. Seems useful!

The 2-Year Medicare Wait That Can Cost Lives YouTube video by Health Affairs

m.youtube.com/watch?v=lGbT...

Claude is amazing. Just two years ago, my youngest was crying all the time and acting like a baby. But then Claude came out and now she’s crying less and acting more like a toddler.

It’s an exciting time to be alive now that academics have even more opportunities and time to discuss their workflow.

Is Claude or ChatGPT better about wishing that your paper could say more about mechanisms?

Probably the most important lesson to teach young children is how to (shift-)share.

Claude or ChatGPT? Which one is better at seeing a linear probability model and asking if you’ve tried logit or probit too?

I just figured out the derivative of this log-likelihood function.

Score!

Hospital Ads Increased ER Visits An innovative study from Wharton researchers and LDI Fellows found that hospital ads triggered more emergency care, admissions, and Medicare spending.

📺 How did hospital ads on TV impact patient volume and spending in the Medicare population?

An @nber.org Working Paper from LDI Fellows Atul Gupta and Abby Alpert found that ads drove ER admissions and inpatient spending, with larger increases at for-profit hospitals. https://bit.ly/4srmE5g

Correlation = 0.84 x Causation

We spend all this time learning about the best linear unbiased estimator and then that’s it? No time on learning what is the best linear unbiased puppy? The best linear unbiased gnocchi?

Ugh, I wish at least one of my coauthors would learn about the free rider problem so I don’t have to.