Dan Kleinman
@dankleinman.bsky.social
Research scientist at Yale Child Study Center (formerly Haskins Labs). PhD. Interested in language, dyslexia, bilingualism, EEG, stats. Proponent of summary limericks. I probably drink more tea than you. Posts are ~50% science; opinions are 100% my own.
Me: It's nice not having a lot of meetings these days!
Zoom: You were saying...?
Zoom: You were saying...?
September 3, 2025 at 2:30 PM
Me: It's nice not having a lot of meetings these days!
Zoom: You were saying...?
Zoom: You were saying...?
This is a nice description of what I call “passing the smell test”. Fancy stats can be very useful (I occasionally rely on them!), but it’s critical to look for converging evidence that an effect is “real” rather than blindly trusting the results of a model (which could be misspecified).
August 15, 2025 at 6:07 PM
This is a nice description of what I call “passing the smell test”. Fancy stats can be very useful (I occasionally rely on them!), but it’s critical to look for converging evidence that an effect is “real” rather than blindly trusting the results of a model (which could be misspecified).
The process couldn't be easier. After getting packages installed and loaded, you read in the data using Julia command readjmp() (within R command julia_command()) and then transfer the result to the R workspace using julia_eval(). Magic!
August 6, 2025 at 3:56 PM
The process couldn't be easier. After getting packages installed and loaded, you read in the data using Julia command readjmp() (within R command julia_command()) and then transfer the result to the R workspace using julia_eval(). Magic!
In "news that will be very exciting to a small group of people", I discovered today that – a full decade after I first wished for this feature – it is now possible to read JMP files into R!
August 6, 2025 at 3:56 PM
In "news that will be very exciting to a small group of people", I discovered today that – a full decade after I first wished for this feature – it is now possible to read JMP files into R!
One of my favorite bands is crowdfunding to release a new double album. I kicked in to get a digital copy, but was gobsmacked by this singular perk, which awaits a very rich superfan. The FAQ includes this gem: "Q: Is the "Fateful Choice" incentive real? A: We dare you."
July 9, 2025 at 8:27 PM
One of my favorite bands is crowdfunding to release a new double album. I kicked in to get a digital copy, but was gobsmacked by this singular perk, which awaits a very rich superfan. The FAQ includes this gem: "Q: Is the "Fateful Choice" incentive real? A: We dare you."
In La Jolla, CA, you can find the intersection of La Jolla Village Dr. and Villa La Jolla Dr. Close by is the intersection of Nobel Dr. and Lebon Dr., which is just "Nobel" spelled backwards. Together, these streets form a wholly unnecessary square of confusion.
June 4, 2025 at 3:57 PM
In La Jolla, CA, you can find the intersection of La Jolla Village Dr. and Villa La Jolla Dr. Close by is the intersection of Nobel Dr. and Lebon Dr., which is just "Nobel" spelled backwards. Together, these streets form a wholly unnecessary square of confusion.
Having come of academic age during the social psych replication crisis, I’ve been overwhelmed by the scope of the structural changes to incentives that would be needed to reduce scientific fraud. Giving ~$1M to @jamesheathers.bsky.social et al. is an excellent step in the right direction.
June 4, 2025 at 1:58 PM
Having come of academic age during the social psych replication crisis, I’ve been overwhelmed by the scope of the structural changes to incentives that would be needed to reduce scientific fraud. Giving ~$1M to @jamesheathers.bsky.social et al. is an excellent step in the right direction.
I get a lot of spam from fake journals, but this one — which has a… unique subject line; has no journal title (only an acronym, here redacted); and for which the main argument seems to be that my submission within 24 hours would improve the sender’s (“my”) ranking — managed to set itself apart.
![An email with some names redacted. Subject line: “Shower us with your precious article”. Body: “Hi [redacted], To support [redacted], we kindly request you to submit your article within the next 24 hours. Your prompt submission will also contribute to my ranking this month.”](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:i3sq6zfcvzshxha3w3bifzzu/bafkreigx42i5zdspor3s5ou5yisygf5bazubkz7utg62rvhsaryxjoxdl4@jpeg)
May 6, 2025 at 1:28 PM
I get a lot of spam from fake journals, but this one — which has a… unique subject line; has no journal title (only an acronym, here redacted); and for which the main argument seems to be that my submission within 24 hours would improve the sender’s (“my”) ranking — managed to set itself apart.
Not an obscure place, but the Copenhagen Contemporary in Refshaleøen has a Ganzfeld installation (Aftershock) by artist James Turrell, who uses light as a medium. As with his other Ganzfelds, it’s sublimely disorienting (though absolutely verboten for anyone prone to epilepsy). Worth a visit.
April 26, 2025 at 8:31 PM
Not an obscure place, but the Copenhagen Contemporary in Refshaleøen has a Ganzfeld installation (Aftershock) by artist James Turrell, who uses light as a medium. As with his other Ganzfelds, it’s sublimely disorienting (though absolutely verboten for anyone prone to epilepsy). Worth a visit.
As a classroom exercise, they tasted some chili peppers, rated their spiciness, and tried to account for the data using a logarithmic function. Apparently, this did pretty well. A good reason not to use an interval chili scale!
March 21, 2025 at 2:45 PM
As a classroom exercise, they tasted some chili peppers, rated their spiciness, and tried to account for the data using a logarithmic function. Apparently, this did pretty well. A good reason not to use an interval chili scale!
Are... are these reverse-coded?
March 7, 2025 at 4:49 PM
Are... are these reverse-coded?
Of course, none of this is useful to me, a non-biologist watching an eagle cam wondering whether I'm looking at the mom eagle or the dad eagle. After all, I can't measure their hallux claw length or bill depth. But that won't stop me from enjoying the heartwarming family scenes.
March 7, 2025 at 2:38 PM
Of course, none of this is useful to me, a non-biologist watching an eagle cam wondering whether I'm looking at the mom eagle or the dad eagle. After all, I can't measure their hallux claw length or bill depth. But that won't stop me from enjoying the heartwarming family scenes.
Bortolotti devised a simple predictive model to discriminate between sexes using 2 criteria. It achieved 98.1% accuracy in his sample, and it showed validity via extension to 4 live (captive) eagles which weren't used to derive the formula – demonstrating the importance of out-of-sample testing.
March 7, 2025 at 2:38 PM
Bortolotti devised a simple predictive model to discriminate between sexes using 2 criteria. It achieved 98.1% accuracy in his sample, and it showed validity via extension to 4 live (captive) eagles which weren't used to derive the formula – demonstrating the importance of out-of-sample testing.
Me: “Long day, more terrible news for the future of science in this country, I’ll do the crossword to escape for a few minutes.”
The #nytxw:
The #nytxw:
January 28, 2025 at 4:49 AM
Me: “Long day, more terrible news for the future of science in this country, I’ll do the crossword to escape for a few minutes.”
The #nytxw:
The #nytxw:
More legalese: You still indemnify them for a whole set of potential claims, but they now only indemnify you for certain kinds of claims – specifically, "allegations that the OSF Software (but not the Content uploaded to it) infringes the patent or copyrights of a third party."
January 6, 2025 at 8:53 PM
More legalese: You still indemnify them for a whole set of potential claims, but they now only indemnify you for certain kinds of claims – specifically, "allegations that the OSF Software (but not the Content uploaded to it) infringes the patent or copyrights of a third party."
They now provide for single sign-on authentication, which makes sense given the addition of institutional affiliates as noted above.
January 6, 2025 at 8:53 PM
They now provide for single sign-on authentication, which makes sense given the addition of institutional affiliates as noted above.
The terms of use also clarify which open-source license OSF software is available under (Apache 2.0), and now note the (potential?) existence of an appeals process should your use of the website be terminated due to unauthorized use.
January 6, 2025 at 8:53 PM
The terms of use also clarify which open-source license OSF software is available under (Apache 2.0), and now note the (potential?) existence of an appeals process should your use of the website be terminated due to unauthorized use.
The terms of use have also changed. There is some updated language about the mission of the COS, which now describes itself as "a nonprofit culture change organization that aims to align scholarly values with scholarly practices". Nice!
January 6, 2025 at 8:53 PM
The terms of use have also changed. There is some updated language about the mission of the COS, which now describes itself as "a nonprofit culture change organization that aims to align scholarly values with scholarly practices". Nice!
The circumstances in which the Center for Open Science (COS, which manages OSF) may share information with third parties is no longer restricted to a "need to know" basis. Also, they retain the right to conduct surveys about site usage and open practices more generally.
. By providing us with your personal information during the user registration process and by agreeing to the terms of this Privacy Policy, you expressly consent to our storing, processing, and distributing your information for these purposes."
Updated text: "COS may share your personal information with our vendors, agents or contractors, such as a vendor that may host COS's servers to help us operate our Websites and Services. You can read a list of these third parties [here](https://help.osf.io/article/227-third-party-vendors). Additionally, we may utilize certain platforms for event management or automated workflows, which may collect and pass information between systems we ultimately use for data storage and management. By providing us with your personal information during the user registration process and by agreeing to the terms of this Privacy Policy, you expressly consent to our storing, processing, and distributing your information for these purposes.
COS may reach out to users to conduct research - whether that is internal user research to improve services or generalizable and externally shareable research on open practices. In doing so we will adhere to ethical guidelines when conducting generalizable and externally shareable research."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:i3sq6zfcvzshxha3w3bifzzu/bafkreicu5gj6pj26x6eilgayj3oragkvjukgwdin4ftabceziycqqlx6lm@jpeg)
January 6, 2025 at 8:53 PM
The circumstances in which the Center for Open Science (COS, which manages OSF) may share information with third parties is no longer restricted to a "need to know" basis. Also, they retain the right to conduct surveys about site usage and open practices more generally.
It appears that there will be new roles for project administration, including at the institutional level.
, who can access an insights dashboard for their institution, as well as request access to affiliated content in order to assist with metadata comprehensiveness. Administrators adhere to both the service agreement between COS and the Administrator's business, organization, or institution and to the COS's Terms of Use and Privacy Policy."](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:i3sq6zfcvzshxha3w3bifzzu/bafkreiauxotlad7ilyb3glq6ljwufgjqpdul3dofad25lukfyoa7l2gbzq@jpeg)
January 6, 2025 at 8:53 PM
It appears that there will be new roles for project administration, including at the institutional level.
First, the privacy policy. Some changes appear to be housekeeping: clearly delineating between the two documents, defining who counts as a "user" of the system. In a substantive change, it appears that pre-registrations will now be implemented with version control, permitting time-stamped updates.
January 6, 2025 at 8:53 PM
First, the privacy policy. Some changes appear to be housekeeping: clearly delineating between the two documents, defining who counts as a "user" of the system. In a substantive change, it appears that pre-registrations will now be implemented with version control, permitting time-stamped updates.
OSF is updating their privacy policy and terms of use! Looks like the biggest changes for most users will be (1) the addition of version control for pre-registrations, and (2) the addition of single sign-on, which will permit institutional affiliate roles. More details below:
January 6, 2025 at 8:53 PM
OSF is updating their privacy policy and terms of use! Looks like the biggest changes for most users will be (1) the addition of version control for pre-registrations, and (2) the addition of single sign-on, which will permit institutional affiliate roles. More details below:
Reminds me of this Achewood comic.
It takes expertise (or at least some savvy) to understand how easy it is to automate a process. Without that level of understanding, all software seems magical — and, pitched by the right salesperson, is worth arbitrary amounts of money.
It takes expertise (or at least some savvy) to understand how easy it is to automate a process. Without that level of understanding, all software seems magical — and, pitched by the right salesperson, is worth arbitrary amounts of money.
![A 6-panel Achewood comic by Chris Onstad featuring two cats, Ray and Roast Beef. Ray: “Roast Beef! Why are you programming that computer?” / Roast Beef: “I guess it’s the only thing that I’m really good at” / Ray: “What are you programming it to do?” Roast Beef: “This spreadsheet would tell a person when they needed to buy eggs or milk” / Roast Beef: “Based on how much eggs or milk they started with” / [A light bulb goes off above Ray’s head] / SOON: Ray, on a call center headset: “That’s right, sir. Based on how much eggs or milk you started with. Could I interest you in a trial version?” Roast Beef, to Ray: “I also wrote a function that calculates how much hash browns to make”](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:i3sq6zfcvzshxha3w3bifzzu/bafkreigu4kpwyumpe2emlutx552gfcucjkywby4z4piqiuswxsfbrwkvfa@jpeg)
December 30, 2024 at 5:23 PM
Reminds me of this Achewood comic.
It takes expertise (or at least some savvy) to understand how easy it is to automate a process. Without that level of understanding, all software seems magical — and, pitched by the right salesperson, is worth arbitrary amounts of money.
It takes expertise (or at least some savvy) to understand how easy it is to automate a process. Without that level of understanding, all software seems magical — and, pitched by the right salesperson, is worth arbitrary amounts of money.
Reading about orthography-phonology mappings and came across this gem by Kearns (2020) (doi.org/10.1002/rrq....). Language researchers know that almost anything can be an example; we should use this power for good more often!
December 20, 2024 at 5:51 AM
Reading about orthography-phonology mappings and came across this gem by Kearns (2020) (doi.org/10.1002/rrq....). Language researchers know that almost anything can be an example; we should use this power for good more often!
Boo on that editor. We once got away with a Highlights limerick:
December 19, 2024 at 2:32 PM
Boo on that editor. We once got away with a Highlights limerick: