John V. Kane
banner
uptonorwell.bsky.social
John V. Kane
@uptonorwell.bsky.social

Political Scientist. Professor at NYU’s Center for Global Affairs. Experiments, data analysis, guitar, drums, fan of comedy. Make guides for @statacorp.bsky.social users. Two boys and exhausted all the time. More at www.johnvkane.com .. more

Political science 48%
Physics 27%
Pinned
🚨It's finally out AND #OpenAccess!!!

Do you do survey experiments? This article is for you! 7 things that increase the risk of null/non-significant results & how to detect/prevent them. It's one of my fav things I've ever written so I hope you enjoy ☺️ polisky

cup.org/3OQhKNv
More than meets the ITT: A guide for anticipating and investigating nonsignificant results in survey experiments | Journal of Experimental Political Science | Cambridge Core
More than meets the ITT: A guide for anticipating and investigating nonsignificant results in survey experiments
cup.org

Thanks so much! 😁

Not sure Daniel has anywhere near enough time (and not totally sure I have anywhere near enough experience doing meta science). That said, thinking about null results is kind of an obsession of mine, so I’ll keep it in mind for the future! 🙏

Really enjoyed this, as always!👏

FWIW, one argument I find persuasive is that, yes, high SES kids have advantages on the SAT that are unrelated to the underlying trait.

BUT—banning it would give more weight to things that high SES kids are *even more* advantaged on (rec letters, HS quality, etc) 🤔

Indeed! Would love to see a side-by-side comparison of (1) z/t-statistic distributions for hypothesized effects (we'd see the classic drop right around 1.96ish), vs. (2) z/t-statistic distributions for any placebo tests--just how different would these distributions look? 😬

The more I think about it, p-stacking might be esp common for robustness checks & placebo tests, wherein the goal is to show a non-significant effect.

For example, rule out an alternative explanation via showing it's non-significant, which serves as more evidence for one's significant effect. 🤔

😂 that one crossed my mind as well.

Ultimately I thought better to go with something that evokes “going up” (and maybe something less violent lol)

Really enjoyed this!

I esp loved the discussion of reverse p-hacking as a means of purposely generating null results. I could picture this happening more as null results become more acceptable--it'd be yet another way of creating a "clear story." Might I suggest calling it: "p-stacking"? 😉

I didn’t even notice—I just feel that last sentence deep in my soul every time I have to grade a new batch of assignments 😭

Amen to this, Andrew!
Wrote up a little intervention post/explanation for my class about why using LLMs for trying to learn programming (as first time learners!) is bad and detrimental datavizf25.classes.andrewheiss.com/news/2025-11...
ANES Data Release! electionstudies.org/data-center/...

The 3-wave ANES panel is now available. It merges data from 3 election studies (2016-2020-2024), the first time the ANES has collected interviews of the same respondents across 3 presidential elections.
2016-2020-2024-panel-merged-study - ANES | American National Election Studies
electionstudies.org

Halloween fun fact 🎃: "Skeleton" has its own IMDB page. 💀

In fact, it stars in one of my favorite horror movies as a kid: House on Haunted Hill (1959 version). Not necessarily the best performance in the film, but also not bad given that it lacks a central nervous system.

So nice of you to say that, Andrew. And thank you for sharing about your children—they’re lucky to have a dad as kind as you, and it makes the paper all the more touching.

And your graphs look terrific! 🤩 (I quickly spotted those bolded headings on the y-axis 😊) So glad the guides are useful! 🙏

Thanks so much for this, Andrew. Before the Ph.D. I was a social worker for 6 years, working with developmentally disabled adults. This research hits close to home 😔

LOVE this book ❤️🙏

@sgadarian.bsky.social and @polpsychangel.bsky.social, in case it’s useful for you! 👍

Much appreciated, Sam! 🙏
🚨 I created a new guide on how to make nice dot charts. These can be really informative, yet remain weirdly underutilized. The examples are with @statacorp.bsky.social but the logic applies to all
software. Link in thread. 👇

Made it mostly for my students but hope you find it useful, too! 😁

You are most welcome! I’m a huge fan of writing practical guides for researchers and students, so it was wonderful to read yours.

(That said, apologies for the unintentional spike in requests! 😬)

@meganakpeters.bsky.social I really loved this article. I definitely will share with students.

FYI I posted it on "the other site" and it kind of blew up (as it should!). 😁

A short-and-sweet guide to developing research questions published in @nature.com.

Lots of good advice here 👍

Really appreciate it, Rikio! My primary goal for that paper was for it to be as practical and useful for researchers as possible, so this is really nice to hear! 😊🙏

Reposted by Winston T. Lin

Are you or one of your students considering doing a Ph.D. in a social science? I've spent a lot of time talking about this w/ students & finally wrote something up.

IMO, there are only 3 good reasons to do it. One of them needs to be true--otherwise, don't.

medium.com/the-quantast...
The Only Three Reasons to Do a Ph.D. in the Social Sciences
If none are true, don’t do it.
medium.com

Just heard about this from a friend. Wow. It’s like the Refused, Hives, Minus (circa “Jesus Christ Bobby” album), and At the Drive-In all combined together in the best possible way. Loving it.

My goodness—hope the recovery keeps going smoothly, Raul! 🤞

Reposted by David Darmofal

🚨 Very flattered to have a piece, w/ the terrific
@miacosta.bsky.social, in the new @apsa.bsky.social Experiments Newsletter!

In it, we offer some thoughts on how to think about how "realistic" experimental results are. Might be useful for students 😁

connect.apsanet.org/s42/newslett...
Newsletter: Main Page
connect.apsanet.org

Many thanks for the s/o, Arieda! I’m really glad you liked it and hope folks find it useful 😊

Reposted by John Kane, Yue Qian

If you use Stata and you want to make nice coefficient plots, this article by @uptonorwell.bsky.social is really fantastic!

medium.com/the-stata-ga...
Making Regression Coefficient Plots in Stata
Traditionally, researchers reported results of regression analyses using tables. A more visually appealing way of presenting these results is by using a coefficient plot. Unlike a typical scatterplot…
medium.com

Just submitted book revisions 🥳 and wanted to write something a little more "fun."

I've been wanting to write this one for a long time and just posted it on Medium (no paywall 👍). For the profs out there, it might serve as a useful class reading (?). Hope you enjoy!

medium.com/@jvk221/3-st...
3 Statistical Problems with Fate
We’ve all thought about whether particular events in our lives might have been due to Fate. Here are 3 statistical problems with doing that.
medium.com

Agreed, yes. I think what we're showing is that info about the "inflation rate" might not be interpreted the way we assume. There's clearly a lot of public confusion about what it means.

Lots more to do on this, so thank you for the feedback! 😁