David P. Miller
H-index:
53
According to FoxNews:
Independent journalist Christopher Rufo reported on Monday that, according to Mamdani’s full Columbia application, he scored 2140 out of 2400 on the SAT, which is below the median score of students admitted to the university in 2009.
There is literally a 50/50 chance of that.
Independent journalist Christopher Rufo reported on Monday that, according to Mamdani’s full Columbia application, he scored 2140 out of 2400 on the SAT, which is below the median score of students admitted to the university in 2009.
There is literally a 50/50 chance of that.
by Robert M. Califf — Reposted by David P. Miller
I am grieving for the many talented & dedicated FDA employees who have been mistreated & those left to do the work of protecting public health. Also for those who will be harmed by this among patients and the public. Those who are complicit in their silence will be judged harshly by history.
Why does the news always report fire burn areas in acres? Is that meaningful to anyone? NYT reported that the 27K acres burned so far is roughly equivalent to 20K football fields. That’s helpful, but why not say equivalent to approximately 2 Manhattan islands, or note that SF is 30K acres?
I hope BlueSky is able to respond to this. I also don’t feel like I have a good grasp of what’s tolerated or not tolerated and who is making the rules.
90% huh? Did you find your own bad AI to generate such a silly number?
That’s for the best. The text is obvious misinformation.
Waking up to get the paper from the driveway on the first Sunday in January is the right way to get a year-end list.
I’m going to read all these faker lists that are dropping now, but I hope they all missed something great that’s going to happen in December.
I’m going to read all these faker lists that are dropping now, but I hope they all missed something great that’s going to happen in December.
Final thoughts. Listen and build trust first. Don’t provide any data at all until you have trust. The real world is a little bit more complicated than Bayes Theorem (10/10)
In the end, you have a group of overly certain educated voters who don’t understand why their data and their candidates are being rejected by the less educated voters who don’t trust them (9/10)
Scientists do collect more and more data (e.g., more research is needed), but that data just causes them to have greater and greater confidence since they selectively publish data with which they agree (8/10)
Scientific peer review relies on reviewers who also hold subjective beliefs. They routinely reject papers that contradict their beliefs or the commonly accepted beliefs of their field. They readily accept papers that align with their beliefs (7/10)
What about educated voters? Are they objective and logical? I don’t think they are. What do those signs really mean that say this household believes in science? Nobody understands all the science. They mean they trust scientific peer review. (6/10)
In today’s world, where social media algorithms can be constructed to feed us information that aligns with our subjective beliefs, it is frighteningly easy for the subjective beliefs to win out over objective data (5/10)
In reality, especially for low-information voters, if subjective beliefs conflict with “the data”, one tends to distrust the data source and quits gathering additional data from that data source. The data can’t win if you don’t trust it (4/10)
The theory is that the subjective beliefs are supplanted by the data because, where disagreement exists, the data can always be increased such that the data in aggregate eventually wins out (3/10)
In Bayesian theory, one has some totally subjective beliefs, which can be represented as a “prior”. One gets in new data and updates their beliefs as a mix of the “prior” and the “data”. As one adds more and more data, the subjective prior beliefs become irrelevant (2/10)
In the context of Trump’s victory, I’ve been thinking a lot about Bayesian theory vs Bayesian practice. How does data actually change our beliefs, and how does that differ for so-called educated voters and low-information voters (1/10)
I’m tired of seeing posts about Trump’s cabinet picks being unqualified. They’re highly qualified for being the wrecking balls Trump wants them to be.
Focusing on qualifications reeks of coastal elitism. We should be talking instead about the value to regular people that will be lost.
Focusing on qualifications reeks of coastal elitism. We should be talking instead about the value to regular people that will be lost.
Strongly agree we need more human-interpretable measures in this arena.
Hall & Oates “I Can’t Go For That (No Can Do)” has been on heavy rotation at the grocery store.
I’ve only just realized that “No Can Do” is doubling down on “I can’t go for that”.
I always thought they couldn’t go for that “No Can Do” attitude.
Still learning. Never stop never stopping.
I’ve only just realized that “No Can Do” is doubling down on “I can’t go for that”.
I always thought they couldn’t go for that “No Can Do” attitude.
Still learning. Never stop never stopping.
Ummm… this might not be the right platform to engage Trump voters to learn what they think?
And Biden didn’t solve inflation. He slowed it, impressively and heroically, but telling people it wasn’t still a problem was and is bad politics.
And Biden didn’t solve inflation. He slowed it, impressively and heroically, but telling people it wasn’t still a problem was and is bad politics.
I love voting by mail in California and I completely trust the process. That said, there has to be something we can do about the pace of counting. All ballots had to be received 3 days ago. How can we still have 3 competitive congressional races with less than 90% of the vote counted? Fix this CA.
I suspect your preferred takeaway is that people who are comfortable with the log-rank test should be even more comfortable with Cox models.
I came to a different takeaway a couple years ago. People who are uncomfortable with the Cox model should be even more uncomfortable with the log-rank test.
I came to a different takeaway a couple years ago. People who are uncomfortable with the Cox model should be even more uncomfortable with the log-rank test.
Good advice on pushing back on p-value obsessed journal reviewers, but how many times does this same piece have to be written before we recognize that peer review needs to be re-invented entirely. I don't believe it can be fixed. www.acpjournals.org/doi/10.7326/...
The wordlebot is either bad at evaluating skill or bad at evaluating luck. I suspect it's both.
Among 6 re-elected incumbents since 1960, the average share of their party's NH primary vote was 82.5%.
Among 5 non-re-elected incumbents since 1960, the average share of their party's NH primary vote was 57.4%.
Biden will probably land around 65% once all the write-ins are tallied.
Among 5 non-re-elected incumbents since 1960, the average share of their party's NH primary vote was 57.4%.
Biden will probably land around 65% once all the write-ins are tallied.
Insufficient expertise and also insufficient time. When I think about the amount of time that goes into FDA reviews vs journal reviews, it's no wonder that I trust the drug label much more than the peer review. Sadly, there's no equivalent to the drug label for most peer reviewed research.
Journal-style peer review just doesn't work today. Maybe it never worked well, but it definitely doesn't today.
In alphabetical order (3rd 5 of 15):
Palehound - Eye on the Bat
Red Pants - Not Quite There Yet
Jeff Rosenstock - Hellmode
Slaughter Beach, Dog - Crying, Laughing, Waving, Smiling
Squirrel Flower - Tomorrow's Fire
Palehound - Eye on the Bat
Red Pants - Not Quite There Yet
Jeff Rosenstock - Hellmode
Slaughter Beach, Dog - Crying, Laughing, Waving, Smiling
Squirrel Flower - Tomorrow's Fire