Here’s what we got for ya:
- [2022] Toward a taxonomy of trust for probabilistic machine learning. {\em Science Advances}.
(Tamara Broderick, Andrew Gelman, Rachael Meager, Anna L. Smith, and Tian Zheng) - [2022] Pathfinder: Parallel quasi-Newton variational inference. {\em Journal of Machine Learning Research} {\bf 23}, 306.
(Lu Zhang, Bob Carpenter, Andrew Gelman, and Aki Vehtari) - [2022] Prediction scoring of data-driven discoveries for reproducible research. {\em Statistics and Computing} {\bf 33}, 11.
(Anna L. Smith, Tian Zheng, and Andrew Gelman) - [2022] The worst of both worlds: A comparative analysis of errors in learning from data in psychology and machine learning. {\em AIES ’22: Fifth AAAI/ACM Conference on AI, Ethics, and Society}, 335–348.
(Jessica Hullman, Sayash Kapoor, Priyanka Nanayakkara, Andrew Gelman, and Arvind Narayanan) - [2022] Selecting on statistical significance and practical significance is wrong. {\em Journal of Information Technology} {\bf 37}, 312–315.
(Blakeley McShane and Andrew Gelman) - [2022] How should scientific journals handle “Big if true” submissions? {\em Chance} {\bf 35} (1).
(Andrew Gelman) - [2022] No reason to expect large and consistent effects of nudge interventions. {\em Proceedings of the National Academy of Sciences} {\bf 119}, e2200732119.
(Barnabás Szászi, Anthony C. Higney, Aaron B. Charlton, Andrew Gelman, Ignazio Ziano, Balacs Aczel, Daniel G. Goldstein, David S. Yeager, and Elizabeth Tipton) - [2022] The development of Bayesian statistics. {\em Journal of the Indian Institute of Science}.
(Andrew Gelman) - [2022] “Two truths and a lie” as a class-participation activity. {\em American Statistician}.
(Andrew Gelman) - [2022] Criticism as asynchronous collaboration: An example from social science research. {\em Stat} {\bf 11}, e464.
(Andrew Gelman) - [2022] Beyond vaccination rates: A synthetic random proxy metric of total SARS-CoV-2 immunity seroprevalence in the community. {\em Epidemiology} {\bf 33}, 457–464.
(Yajuan Si, Leonard Covello, Siquan Wang, Theodore Covello, and Andrew Gelman) - [2022] Stacking for non-mixing Bayesian computations: The curse and blessing of multimodal posteriors. {\em Journal of Machine Learning Research} {\bf 23}, 79. (Yuling Yao, Aki Vehtari, and Andrew Gelman)
- [2022] Inference from non-random samples using Bayesian machine learning. {\em Journal of Survey Statistics and Methodology}, smab049 (Yutao Liu, Andrew Gelman, and Qixuan Chen)
- [2022] The Great Society, Reagan’s revolution, and generations of presidential voting. {\em American Journal of Political Science}.
(Yair Ghitza, Andrew Gelman, and Jonathan Auerbach) - [2022] Reconciling evaluations of the Millennium Villages Project. {\em Statistics and Public Policy} {\bf 9}, 1–7. (Andrew Gelman, Shira Mitchell, Jeffrey D. Sachs, and Sonia Sachs)
- [2022] A proposal for informative default priors scaled by the standard error of estimates. {\em American Statistician} {\bf 76}, 1–9. (Erik van Zwet and Andrew Gelman)
- [2022] Bayesian hierarchical stacking: Some models are (somewhere) useful. {\em Bayesian Analysis} {\bf 17}, 1043–1071. (Yuling Yao, Gregor Pirš, Aki Vehtari, and Andrew Gelman)
- [2022] A fast linear regression via SVD and marginalization. {\em Computational Statistics} {\bf 37}, 701–720. (Philip Greengard, Andrew Gelman, and Aki Vehtari)
- [2022] Mismatch between scientific theories and statistical models. {\em Behavioral and Brain Sciences} {\bf 45}, e15. (Andrew Gelman)
Thanks as always to my collaborators (Tamara, Rachael, Anna, Tian, Lu, Bob, Aki, Jessica, Sayash, Priyanka, Arvind, Blake, Barnabás, Anthony, Aaron, Ignazio, Balacs, Dan, David, Beth, Yajuan, Len, Siquan, Theodore, Yuling, Yutao, Qixuan, Yair, Jonathan, Shira, Jeff, Sonia, Erik, Gregor, Philip, and all the others not listed above), along with our funders.
Happy new year, everyone!
Regarding authorship, what are the rules for ordering the names? That is, what fields of endeavor or particular journals specify alphabetical order, primary author first, primary author last or some other sequence that I am unable to imagine? For example, Metropolis receives the naming rights because he was the first author of the famous paper. Ditto Pontryagin and the maximal principle of control theory.
Within a particular discipline, is the ordering well known and enforced? What does one learn (if anything) by reading the tea leaves of the ordering?
Paul:
It depends on the field.
In econ the norm is alphabetical order, but they don’t always do it that way. Also econ seems to have a tradition of single-authored articles, where many of the people who did the work are just listed in the acknowledgements but not as authors (see discussion here). In the natural sciences, it seems more standard to include everyone as authors, including the people who collected and analyzed the data, the people who built the equipment, etc.
In statistics, or at least in my own collaborations, the first author is the person who wrote the article, the second person is the person who did the second-most, etc. Sometimes we follow the traditional science rule that the head of the laboratory is last author, but we don’t always do that. In the very first article listed above, I think the five of us contributed equally, but maybe Tamara was the one to unify all our contributions into a single article. In the second article listed above, Bob had the original idea and then Lu took it and changed it enough so it was really her idea, and Aki and I made some contributions. And so on. Each article has its own story.
Listing in alphabetical order has a specific advantage in that it leads to ambiguity, thus avoiding future fights when Nobel Prizes in physics are being awarded. Not in alphabetical order sort of implies that the first author is in fact more important because the convenient ambiguity of the alphabet has been overridden. Unfortunately, along with difficulties of umlauts in various languages, in Norwegian, “aa” is actually the last letter of the alphabet as it decidedly is not in English.
In most of computer science its student authors first, and faculty who originated the idea/did the most overseeing/funded last. I’m last author on some papers that I wrote the majority of. But people in industry labs seem to do it differently, where closer to front matters more. And when you have a faculty lead author, but also students and other faculty on a paper, it gets confusing how to read the order of the rest and I’m not sure there’s any single convention.
I used to find all of this frustrating, since supposedly there is evidence from other fields of women who co-author with men getting less credit for that work than for work co-authored with only other women. It’s probably part of why I’ve continued to write first author papers even though its not necessarily so common in CS after becoming faculty.
Just before wishing all of us, “Happy new year, everyone!,” Andrew lists his collaborators by first name only, except for “Andrew Gelman” who comes between “Aaron” and “Ignazio.” As has long been suspected, Andrew is not an individual but a conglomerate instead.
Oops! That was a copy-paste error. Thanks for reading so carefully!
> As has long been suspected, Andrew is not an individual but a conglomerate instead.
That would explain his ridiculous productivity.
is the one in ppnas any good?