Greg Faletto
@gregoryfaletto.com
Statistician, Data Scientist.
Who thinks they should be reviewing twice as much, that they should wait twice as long for their articles to be reviewed, or that their papers should be reviewed by people who are half as qualified?
And without reviewers, you open the floodgates to trash research that adds less than 0 knowledge.
And without reviewers, you open the floodgates to trash research that adds less than 0 knowledge.
November 16, 2025 at 1:42 AM
Who thinks they should be reviewing twice as much, that they should wait twice as long for their articles to be reviewed, or that their papers should be reviewed by people who are half as qualified?
And without reviewers, you open the floodgates to trash research that adds less than 0 knowledge.
And without reviewers, you open the floodgates to trash research that adds less than 0 knowledge.
I get what you're saying, I'm just not sure this is the best way to deal with stuff that doesn't replicate. Preprint servers serve some of the functions that you're talking about in some fields. But *publishing* is costly, with the cost being reviewer time.
November 16, 2025 at 1:42 AM
I get what you're saying, I'm just not sure this is the best way to deal with stuff that doesn't replicate. Preprint servers serve some of the functions that you're talking about in some fields. But *publishing* is costly, with the cost being reviewer time.
We could change the metric and threshold for when we decide a measurement is meaningful if we want to. But whatever metric and threshold we agree on, when you plot a graph of the metric for published papers, there will be a big discontinuity at the agreed-upon threshold
November 16, 2025 at 12:09 AM
We could change the metric and threshold for when we decide a measurement is meaningful if we want to. But whatever metric and threshold we agree on, when you plot a graph of the metric for published papers, there will be a big discontinuity at the agreed-upon threshold
Research questions that are surprising or change our understanding of a topic when the answer is "no" are kind of rare. It's understandable that researchers wish every well-conducted study could be published, but time and attention are scarce and some findings are more worthy of them than others
November 16, 2025 at 12:06 AM
Research questions that are surprising or change our understanding of a topic when the answer is "no" are kind of rare. It's understandable that researchers wish every well-conducted study could be published, but time and attention are scarce and some findings are more worthy of them than others
Asking for evidence that sex scandals often derail the careers of politicians and swing elections feels a bit too clever by half. Google it if you want?
November 15, 2025 at 10:27 PM
Asking for evidence that sex scandals often derail the careers of politicians and swing elections feels a bit too clever by half. Google it if you want?
Using the partially linear model is just leaving money on the table when the AIPW estimator is right there. (Or the efficient influence function for the ATT or whatever other quantity you're estimating.)
November 14, 2025 at 10:25 PM
Using the partially linear model is just leaving money on the table when the AIPW estimator is right there. (Or the efficient influence function for the ATT or whatever other quantity you're estimating.)
"Actually, historically that title was first... whoops, never mind."
November 14, 2025 at 9:25 PM
"Actually, historically that title was first... whoops, never mind."
You're not imagining things--this is a real phenomenon. Here's why it feels that way:
November 14, 2025 at 1:38 AM
You're not imagining things--this is a real phenomenon. Here's why it feels that way:
All we can ever do is the best we can with the information we have at the time, and accept that some of it will turn out to be wrong. For some reason like solidly half of people can't and won't ever accept this
November 12, 2025 at 8:55 PM
All we can ever do is the best we can with the information we have at the time, and accept that some of it will turn out to be wrong. For some reason like solidly half of people can't and won't ever accept this
Maybe, but this table wouldn't tell you either way
November 12, 2025 at 8:44 PM
Maybe, but this table wouldn't tell you either way
Book publishers aren't stewards of other people's money or tasked with serving the public interest, it seems like they morally ought to have more liberty to allocate their money however they want
November 10, 2025 at 10:16 PM
Book publishers aren't stewards of other people's money or tasked with serving the public interest, it seems like they morally ought to have more liberty to allocate their money however they want
"Now that the election is over, we can ignore the stupid voters' opinions on what we need to do to get re-elected"
November 9, 2025 at 11:25 PM
"Now that the election is over, we can ignore the stupid voters' opinions on what we need to do to get re-elected"
This is the same impulse behind expecting squeaky clean behavior from Democrats on gerrymandering or candidate quality. "Democrats voted for a candidate who used violent rhetoric in a private conversation?? Truly a new low in American history!"
November 9, 2025 at 4:45 AM
This is the same impulse behind expecting squeaky clean behavior from Democrats on gerrymandering or candidate quality. "Democrats voted for a candidate who used violent rhetoric in a private conversation?? Truly a new low in American history!"
I feel like as you go back hundreds of years this gets more and more stark--storytellers were way more willing and able to go on random unimportant tangents
November 9, 2025 at 3:02 AM
I feel like as you go back hundreds of years this gets more and more stark--storytellers were way more willing and able to go on random unimportant tangents
One thing I've noticed even in the last 20 or so years is that storytelling has gotten tighter as there have been more demands for people's attention (especially smartphones I imagine). If you watch a comedy or variety show from the 60s, they are really not trying that hard to keep your attention
November 9, 2025 at 3:02 AM
One thing I've noticed even in the last 20 or so years is that storytelling has gotten tighter as there have been more demands for people's attention (especially smartphones I imagine). If you watch a comedy or variety show from the 60s, they are really not trying that hard to keep your attention
Most of the people in my circles are pretty well-educated, cosmopolitan etc., and among those the most Trump-curious people I've met have either been exactly who you'd think (jock/frat/small business owner's son white men) or first-generation immigrants of color.
November 9, 2025 at 12:14 AM
Most of the people in my circles are pretty well-educated, cosmopolitan etc., and among those the most Trump-curious people I've met have either been exactly who you'd think (jock/frat/small business owner's son white men) or first-generation immigrants of color.
Maybe so, but I think "debate kid" is the best angle he has to offer the general public. Other than that he's got "too online guy who's a weird kind of racist that isn't even relatable to most racists"
November 9, 2025 at 12:02 AM
Maybe so, but I think "debate kid" is the best angle he has to offer the general public. Other than that he's got "too online guy who's a weird kind of racist that isn't even relatable to most racists"
I mean Tim Walz did pretty terrible in that debate. And yeah I do think a debate is a pretty good environment for Vance--I wouldn't be shocked if he was a former debate kid in high school. But I think the Kamala Harris who showed up for the one debate with Trump would have cleaned Vance's clock
November 8, 2025 at 11:44 PM
I mean Tim Walz did pretty terrible in that debate. And yeah I do think a debate is a pretty good environment for Vance--I wouldn't be shocked if he was a former debate kid in high school. But I think the Kamala Harris who showed up for the one debate with Trump would have cleaned Vance's clock
He always looks like an 8-year-old who just told you his uncle works at Nintendo and he gets all the games before they come out, and he hasn't figured out yet how to not show it on his face that he's trying to figure out if you believe him
November 8, 2025 at 11:35 PM
He always looks like an 8-year-old who just told you his uncle works at Nintendo and he gets all the games before they come out, and he hasn't figured out yet how to not show it on his face that he's trying to figure out if you believe him
(mitigated, again, only by his sheer shamelessness to not acknowledge having done anything wrong, let alone panic)
November 8, 2025 at 1:18 PM
(mitigated, again, only by his sheer shamelessness to not acknowledge having done anything wrong, let alone panic)
Having done that, I'd say his campaign in 2024 was generally pretty bad, and most of the result can be explained by the fact that a ham sandwich with R next to their name would have won that election. He did terrible in the only debate, the last week of his campaign was a genuine disaster
November 8, 2025 at 1:18 PM
Having done that, I'd say his campaign in 2024 was generally pretty bad, and most of the result can be explained by the fact that a ham sandwich with R next to their name would have won that election. He did terrible in the only debate, the last week of his campaign was a genuine disaster
His (difficult, unprecedented) accomplishment in 2024 was precluding a serious threat of a primary opponent *despite having lost in 2020*. Mostly through sheer shamelessness--making it difficult for Rs to acknowledge he lost in 2020 prevented soul-searching that would lead to a different candidate
November 8, 2025 at 1:18 PM
His (difficult, unprecedented) accomplishment in 2024 was precluding a serious threat of a primary opponent *despite having lost in 2020*. Mostly through sheer shamelessness--making it difficult for Rs to acknowledge he lost in 2020 prevented soul-searching that would lead to a different candidate
Look carefully at your paycheck for going to the No Kings protest, he actually signed it
November 6, 2025 at 5:55 PM
Look carefully at your paycheck for going to the No Kings protest, he actually signed it
Thursday/Friday ask people about their plans for the weekend and talk about yours, Monday/Tuesday talk about the previous weekend, that alone gets you pretty far. If you're not doing anything remotely interesting over the weekend fix that (could even just be like "made pancakes" or "went to Ikea")
November 6, 2025 at 3:50 AM
Thursday/Friday ask people about their plans for the weekend and talk about yours, Monday/Tuesday talk about the previous weekend, that alone gets you pretty far. If you're not doing anything remotely interesting over the weekend fix that (could even just be like "made pancakes" or "went to Ikea")