Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection

@bayesianboy Mathematistry! https://t.co/KR5R6u0xdv

@AlonZivony I found Dani Navarro's paper to be particularly useful when thinking about how computational models fit into simplified 'toy' environments. Those environments are an essential part of computational models and are shaped by theory. https://t.

@_fernando_rosas @PessoaBrain @R3RT0 @evantthompson As is this https://t.co/8jlT7ny94h

Este es uno de los papers que más han hecho huella en mi forma de pensar sobre la ciencia. Y además uno de los mejor escritos que conozco. Tiene un párrafo que desearía con toda mi alma haber escrito yo mismo. https://t.co/TtrSljQO9k

👀👀 Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection https://t.co/LxX86D3Jfq

RT @IrisVanRooij: 4. Also be sure to check out this paper: https://t.co/560KNwEBo5

@VandekerckhoveJ https://t.co/5p9T4D5nXR

The amazing Goddess of the Devil and the Deep Blue Sea @djnavarro ❤️‍🔥💜😍🙌 cc @StephStammel https://t.co/m0XTwvNP8X https://t.co/kv4C1p61YJ

@dougaparry @EikoFried @zerdeve And related from Navarro https://t.co/CAbV1fZijc

RT @vascobrazao: @RubinPsyc Reminds me of "mathemastistry" https://t.co/FO53PDGLca

RT @vascobrazao: @RubinPsyc Reminds me of "mathemastistry" https://t.co/FO53PDGLca

@RubinPsyc Reminds me of "mathemastistry" https://t.co/FO53PDGLca

#34: Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection (https://t.co/PGhoJXierH)

RT @socmathpsych: Link to the paper https://t.co/85SMSOVtyq

RT @IrisVanRooij: Check out @djnavarro’s brilliant paper if you haven’t yet: https://t.co/560KNwEBo5

Check out @djnavarro’s brilliant paper if you haven’t yet: https://t.co/560KNwEBo5

Link to the paper https://t.co/85SMSOVtyq

@Nate__Haines @o_guest @IrisVanRooij Getting some @djnavarro deep blue sea vibes ☺️ ICYMI https://t.co/KR5R6u0xdv

RT @Nate__Haines: h/t to both @djnavarro's great discussion of this: https://t.co/YF1Tr3mDi4, in addition to Busemeyer's writing on the gen…

Now we just need to figure out how to convince the undergrads. Highlighting two great papers that discuss what I would describe as being thoughtful with your statistical evidence. @djnavarro 's paper is a modern classic https://t.co/ja9Ac3K1vy.

RT @Nate__Haines: h/t to both @djnavarro's great discussion of this: https://t.co/YF1Tr3mDi4, in addition to Busemeyer's writing on the gen…

@aaronjfisher @SolomonKurz I agree 😊. But on the same point, I don't think fit stats, cross-validation, etc. are necessarily going to actually help us arbitrate between competing theories of the underlying DGP. Busemeyer does a good job of explaining why,

h/t to both @djnavarro's great discussion of this: https://t.co/YF1Tr3mDi4, in addition to Busemeyer's writing on the generalization criterion: https://t.co/pgHB2FcuRp

@mcarondiotte But the above chapter is mostly focused on quantitative model comparison, which is great but not the whole story. For example, check out: https://t.co/YF1Tr3mDi4

RT @chas_new: Just got an opportunity to read “Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistic…

Just got an opportunity to read “Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection” by @djnavarro Very thought-provoking! Link to the article: https://t.co/UvWSzpOxti https://t.co/UN2UHeKry0

10. "Between the devil and the deep blue sea: Tensions between scientific judgement and statistical model selection", by @djnavarro https://t.co/560KNwEBo5

RT @cobac_: @cantabile @djnavarro @DrVeronikaCH I remember feeling so insightful(?) the first time I read this article about how naive my p…

@cantabile @djnavarro @DrVeronikaCH I remember feeling so insightful(?) the first time I read this article about how naive my perspective about methodology was and about how clear and beautiful scientific prose could be. https://t.co/bWFR1xVtUy

RT @alexpghayes: @bayesianboy @mcmc_stan also, mandatory reading by @djnavarro https://t.co/5lpmIP8Y2l

@bayesianboy @mcmc_stan also, mandatory reading by @djnavarro https://t.co/5lpmIP8Y2l

"The problem for Bayesianism (...) is that scientists naturally appeal to qualitative theoretical criteria rather than probabilities when evaluating those theories" 23/ (I am of course reminded of "Between the devil and the deep blue sea" https://t.co/560

@jeblad I høve modellseleksjon kan jeg også anbefale denne: https://t.co/BMAHoNPrEQ Alle modeller er feil, noen er nyttige - hvordan skal vi velge modell, særlig med så kompliserte oppgaver som per i dag kun nevrale nettverk kan løse?

RT @DrSamirBhatt: New work led by @MellanTom and @iwonahaw and others https://t.co/hrknK4MnrH on model selection using what we call referen…

RT @DrSamirBhatt: New work led by @MellanTom and @iwonahaw and others https://t.co/hrknK4MnrH on model selection using what we call referen…

RT @DrSamirBhatt: New work led by @MellanTom and @iwonahaw and others https://t.co/hrknK4MnrH on model selection using what we call referen…

RT @iwonahaw: Check out our new preprint, well summarised in this thread by Sam. We proposed a new method of calculating model evidence usi…

Check out our new preprint, well summarised in this thread by Sam. We proposed a new method of calculating model evidence using "referenced thermodynamic integration". Special thanks to @MellanTom, @creswapi and @DrSamirBhatt for being amazing mentors and

Sam has summarized the work very well. Glad to be a part of it. Special mention to @iwonahaw and @MellanTom for this amazing work.

New work led by @MellanTom and @iwonahaw and others https://t.co/hrknK4MnrH on model selection using what we call referenced thermodynamic integration. Before reading, decide where you stand on Bayes factors and read the excellent work of @djnavarro https:

RT @IrisVanRooij: @wes_bonifay @o_guest @andrea_e_martin 2. By @djnavarro https://t.co/560KNwEBo5

@wes_bonifay @o_guest @andrea_e_martin 2. By @djnavarro https://t.co/560KNwEBo5

covariates in my linear regression won't help? Gasp. Also if you came this far and still haven't read Between the devil & the deep blue sea (https://t.co/KR5R6u0xdv) go do that. I'll wait. Then we can bask in our own ignorance together :) Another exam

@NeuroStats I knew of Box, and the notion of mathematistry, through @djnavarro 's paper "Between the Devil and the Deep Blue Sea" (https://t.co/560KNwEBo5), but realise I should read the original as well. Thanks again.

RT @IrisVanRooij: We are far from the first to point this out of course. It is well known in mathematical psychology that qualitative patte…

RT @IrisVanRooij: We are far from the first to point this out of course. It is well known in mathematical psychology that qualitative patte…

@wfrankenhuis1 @NicoleWalasek Perhaps of interest, too: https://t.co/JSftiw4IMq

RT @IrisVanRooij: We are far from the first to point this out of course. It is well known in mathematical psychology that qualitative patte…

We are far from the first to point this out of course. It is well known in mathematical psychology that qualitative patterns are more informative & more relevant to test for. See eg also @djnavarro’s work here: https://t.co/560KNwEBo5 and see also: ht

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

RT @zerdeve: 4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https…

4. @djnavarro has written on the interplay between science and statistics https://t.co/fcSThjbaUq and on preregistration https://t.co/rUEY5cOFO4

I love this paper with an important message about (my take) taking responsibility for your own scientific inference, and that it is getting deserved recognition.

If you haven't read Navarro's paper on model selection you're missing out on one of the best papers on model selection:

RT @Nate__Haines: There are so many good papers on these and related ideas. A few recent works: 1) Replications with the wrong measure per…

@DingemanseMark @hugh_rab @suzyjstyles @superlinguo @TrishaVZ @richarddmorey Here a related thread that I *just* see come by: https://t.co/qy4WnCtV8M

There are so many good papers on these and related ideas. A few recent works: 1) Replications with the wrong measure perpetuate wrong conclusions (https://t.co/q0KRfIY2zr) 2) Trade-offs between statistical testing and explanatory modeling (https://t.co/Y

@FullerofScience One of my favs: https://t.co/SPDtwmHHP9

4. Also be sure to check out this paper: https://t.co/560KNwEBo5

RT @_frycast: This article caused so many vague but important ideas on statistical modelling, ML, and explainability to crystallise and cli…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

RT @aggieerin: New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuW…

New submission for @Meta_Psy - @RealNathanEvans provides a comment on @djnavarro (2018). Original paper: https://t.co/YEYuWXXRkG Comment: https://t.co/zwKoPsLtwm

RT @IrisVanRooij: "In Between the devil & the deep blue sea https://t.co/560KNwEBo5, I suggested mathematistry is the practice of using for…

RT @IrisVanRooij: "In Between the devil & the deep blue sea https://t.co/560KNwEBo5, I suggested mathematistry is the practice of using for…

"In Between the devil & the deep blue sea https://t.co/560KNwEBo5, I suggested mathematistry is the practice of using formal tools to define a statistical problem that differs from the scientific one, solving the redefined problem & declaring the s

RT @IrisVanRooij: @JohnSakaluk @djnavarro Btw, Danielle already shared her expert views on the distinction between scientific and statistic…

@JohnSakaluk @djnavarro Btw, Danielle already shared her expert views on the distinction between scientific and statistical inference, e.g. here: https://t.co/560KNwEBo5 (Open Access) but either ppl don't read or don't understand. Either way, it underscore

@micahgallen @eegsynth Wonderful paper by @djnavarro on exactly this https://t.co/jYD7AR4Z9q See also her talk https://t.co/h8piZ2elk5

RT @JCSkewesDK: @chrisdc77 @RegReports Here is Danielle's paper which has her own example https://t.co/jYD7AR4Z9q We also filmed her talk a…

Danielle Navarro @djnavarro further illustrating the point https://t.co/uTG5pVitzo

@chrisdc77 @RegReports Here is Danielle's paper which has her own example https://t.co/jYD7AR4Z9q We also filmed her talk and will upload soon. I'm also hoping she'll write a follow up.

I haven't read this article yet but plan to do so. My view has always been that we must approach the truth and instead of thinking in discrete terms like correct and incorrect, we must think in terms of distance from the truth (continuous). Surely we can t

RT @CompBrainBeh: in a new @CompBrainBeh paper, @djnavarro asks "What goal does model selection serve when all models are known to be syste…

RT @CompBrainBeh: in a new @CompBrainBeh paper, @djnavarro asks "What goal does model selection serve when all models are known to be syste…

in a new @CompBrainBeh paper, @djnavarro asks "What goal does model selection serve when all models are known to be systematically wrong? How might “toy problems” tell a misleading story?” https://t.co/I78OxC3cTz

in a new @CompBrainBeh paper, @djnavarro asks "What goal does model selection serve when all models are known to be systematically wrong? How might “toy problems” tell a misleading story?” https://t.co/O5HKDgqn8P

Very interesting thoughts on scientifical vs. technical model selection https://t.co/nbhLtbzoMM

RT @djnavarro: It's been almost a decade since I had much to say on statistical methods in psychology (only my second paper since 2009) but…

RT @davdittrich: Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection https://t.c…

Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection https://t.co/tTaG5jAvOP https://t.co/Fcezlbsqlv ht @aortmannphd

There’s some bandwagon here that apparently every statistician is supposed to jump on to.

RT @djnavarro: It's been almost a decade since I had much to say on statistical methods in psychology (only my second paper since 2009) but…

RT @slderuiter: Great read! Food for thought for teaching model selection in advanced data analysis (something I really need to step up my…

Great read! Food for thought for teaching model selection in advanced data analysis (something I really need to step up my game on...) And also fun quotes and imagery.

RT @rlmcelreath: This paper is framed in terms of psych literature. But speaking as someone who has never taken a psychology course, I foun…

Between the Devil and the Deep Blue Sea: Tensions Between Scientific Judgement and Statistical Model Selection | SpringerLink https://t.co/D3yalP3IrQ

“Since all models are wrong the scientist must be alert to what is importantly wrong. It is inappropriate to be concerned about mice when there are tigers abroad.“ ~George Box

RT @rlmcelreath: This paper is framed in terms of psych literature. But speaking as someone who has never taken a psychology course, I foun…

RT @rlmcelreath: This paper is framed in terms of psych literature. But speaking as someone who has never taken a psychology course, I foun…

RT @rlmcelreath: This paper is framed in terms of psych literature. But speaking as someone who has never taken a psychology course, I foun…

RT @rlmcelreath: This paper is framed in terms of psych literature. But speaking as someone who has never taken a psychology course, I foun…