David Powell
@davidpowell.bsky.social
I wish mathematicians would consider more interesting questions. I mean, sure, there exists an epsilon greater than zero. But is it happy? Has it known love?
November 17, 2025 at 3:43 PM
I wish mathematicians would consider more interesting questions. I mean, sure, there exists an epsilon greater than zero. But is it happy? Has it known love?
I misuse “orders of magnitude” millions of times per day.
November 17, 2025 at 12:28 PM
I misuse “orders of magnitude” millions of times per day.
Reposted by David Powell
The 1975 Voting Rights Act extension lowered mortality for most nonwhite groups but raised it for white adults and older nonwhite men, patterns consistent with status threat, from Atheendar Venkataramani, Rourke O'Brien, Elizabeth F. Bair, and Christopher A. Lowenstein www.nber.org/papers/w34421
November 4, 2025 at 8:30 PM
The 1975 Voting Rights Act extension lowered mortality for most nonwhite groups but raised it for white adults and older nonwhite men, patterns consistent with status threat, from Atheendar Venkataramani, Rourke O'Brien, Elizabeth F. Bair, and Christopher A. Lowenstein www.nber.org/papers/w34421
This model refers to a representative agent, but I don’t remember voting for them.
November 3, 2025 at 5:51 PM
This model refers to a representative agent, but I don’t remember voting for them.
Remember, you under-use “How dare you?!” when responding to reviewer comments.
October 15, 2025 at 6:40 PM
Remember, you under-use “How dare you?!” when responding to reviewer comments.
What would have happened if they never made Sliding Doors?
October 15, 2025 at 1:40 PM
What would have happened if they never made Sliding Doors?
There should be more top 5 journals.
October 14, 2025 at 2:33 PM
There should be more top 5 journals.
Conditional acceptance!
I just have to write a paper that referees and an editor like!
I just have to write a paper that referees and an editor like!
October 10, 2025 at 1:04 PM
Conditional acceptance!
I just have to write a paper that referees and an editor like!
I just have to write a paper that referees and an editor like!
To our knowledge, this is the first paper to ignore the prior literature.
October 8, 2025 at 7:53 PM
To our knowledge, this is the first paper to ignore the prior literature.
Confused about what inference method to use? What level to cluster at?
Based on a review of thousands of published empirical studies, we develop a useful heuristic: a good approximation for your standard error is your point estimate divided by 1.96.
Based on a review of thousands of published empirical studies, we develop a useful heuristic: a good approximation for your standard error is your point estimate divided by 1.96.
October 6, 2025 at 10:42 PM
Confused about what inference method to use? What level to cluster at?
Based on a review of thousands of published empirical studies, we develop a useful heuristic: a good approximation for your standard error is your point estimate divided by 1.96.
Based on a review of thousands of published empirical studies, we develop a useful heuristic: a good approximation for your standard error is your point estimate divided by 1.96.
Reposted by David Powell
New research alert: Medicaid covers nearly 40% of treatment for opioid use disorder (OUD), but recent “unwinding” of pandemic-era enrollment protections led many to lose coverage nationwide. How did this affect life-saving buprenorphine treatment?
September 30, 2025 at 2:10 PM
New research alert: Medicaid covers nearly 40% of treatment for opioid use disorder (OUD), but recent “unwinding” of pandemic-era enrollment protections led many to lose coverage nationwide. How did this affect life-saving buprenorphine treatment?
Do I really have to rue the day? Can I just rue a couple hours?
September 26, 2025 at 5:12 PM
Do I really have to rue the day? Can I just rue a couple hours?
Pro tip: If the journal has space constraints, replace “unintended consequences” with “whoopsies.”
September 14, 2025 at 10:06 PM
Pro tip: If the journal has space constraints, replace “unintended consequences” with “whoopsies.”
My front-of-the-envelope estimate is that it costs 78 cents to mail a letter.
September 12, 2025 at 4:25 PM
My front-of-the-envelope estimate is that it costs 78 cents to mail a letter.
You get more positive reviews if you call it warm and fuzzy regression discontinuity.
September 8, 2025 at 8:50 PM
You get more positive reviews if you call it warm and fuzzy regression discontinuity.
A true labor economist uses Labor Day as a negative shock to labor.
September 1, 2025 at 9:15 PM
A true labor economist uses Labor Day as a negative shock to labor.
The optimal number of times saying, “The optimal number of [terrible thing] is not zero” is zero.
August 30, 2025 at 11:54 PM
The optimal number of times saying, “The optimal number of [terrible thing] is not zero” is zero.
Quick tip: if your paper status says “Pending Decision,” it’s likely to be positive. If it’s likely to be negative, it’ll say “Impending Decision.”
August 28, 2025 at 4:11 PM
Quick tip: if your paper status says “Pending Decision,” it’s likely to be positive. If it’s likely to be negative, it’ll say “Impending Decision.”
Lowering drug prices by 1500% is just hyperbolic discounting.
August 28, 2025 at 12:07 PM
Lowering drug prices by 1500% is just hyperbolic discounting.
A triple-blind study where you also don’t tell the readers which was the treatment arm.
August 27, 2025 at 8:49 PM
A triple-blind study where you also don’t tell the readers which was the treatment arm.
The replication package is your opportunity to add snarky comments throughout your code about your reviewers.
August 25, 2025 at 11:59 PM
The replication package is your opportunity to add snarky comments throughout your code about your reviewers.
I wish there were a word that paid homage to the silent h.
August 25, 2025 at 7:48 PM
I wish there were a word that paid homage to the silent h.
You just think raisins are healthy since they’re disgusting.
August 23, 2025 at 4:58 PM
You just think raisins are healthy since they’re disgusting.
I recommend constructing a file called omitted_variables.dta so when a reviewer asks about omitted variable bias, you can just merge it into your dataset.
August 22, 2025 at 10:47 AM
I recommend constructing a file called omitted_variables.dta so when a reviewer asks about omitted variable bias, you can just merge it into your dataset.