@MichaelColey @EricTopol @AnnalsofIM @HannahGrnwd @BondUniversity Everyone using CIs should read https://t.co/tyuTngPpkk by @richarddmorey and @EJWagenmakers.
@paulpharoah @KevorkJan The fallacy of placing confidence in confidence intervals https://t.co/1KaRju7N1a via @EJWagenmakers et al @dnunan79
@Brier_Gallihugh @zheer_kejlberg CIs don't really quantify uncertainty unfortunately https://t.co/Nlwhsywska
The fallacy of placing confidence in confidence intervals #maths #statistics #forecasting https://t.co/zRRMBaeTek
@statman_sean @JPhysiol "with narrower confidence intervals indicating more precision in the population parameter estimate". There is no necessary connection between the precision of an estimate and the size of a confidence interval. CI is much more confus
@beenwrekt I have liked reading this at various points. Might be of interest if you don't know it already. https://t.co/PwQ5E6mkf6
@bmwiernik Wish it was that easy. Start here: https://t.co/cnNw4n83eu and think very carefully as an exercise how your above CI definition differs from Bayes and fiducial. Each system has advantages & limitations. It is key to know all three when inter
@EpiEllie @kareem_carr I consider these articles must-reads about confidence intervals: https://t.co/DhQfaaROrk https://t.co/8HAWlPwBYW
link to paper: https://t.co/EGB6gWEsrJ
@octonion This paper from Morey et al (2016)The fallacy of placing confidence in confidence intervals, has some interesting examples https://t.co/M4UG1dWqIX
@TomBoylesID @MythoFobe @ChronicUTIAus @xxChronicutixx @KeatingRachelle @StephenTristram And somewhat tangentially, I have greatly enjoyed these on confidence intervals: https://t.co/DhQfaaROrk, https://t.co/8HAWlPwBYW.
RT @dylanarmbruste3: @nick_krontiris @MyNutritionSci @shivadas29 There's history behind the Confidence theory. There's push back against it…
@nick_krontiris @MyNutritionSci @shivadas29 There's history behind the Confidence theory. There's push back against it. https://t.co/3fWCc7e4Zu https://t.co/ihW5p5HIjz
@bertil_hatt @DavidBahry @MaartenvSmeden https://t.co/GZOyIKlBvt goes over the issues of confidence intervals
@bertil_hatt @MaartenvSmeden I think that CIs is misinterpreted in 99% of the case. At least in medical literature. Even from some statisticians. What a CIs really means is harder to understand than what is seems: https://t.co/JaFUhmn1qi.
@SquarrelZei @TimMcCulloch2 @MaartenvSmeden I can do it for you. This is an excellent paper about why CIs is not what most think it's. https://t.co/JaFUhmn1qi. The concept related to CIs is harder to understand for a non statistician so the idea to replac
RT @AleGiannini82: @irongrafFT Un realtà passare all'uso del CI non risolve alcun problema. Anzi probabilmente peggiora. Questo perché l'in…
@irongrafFT Un realtà passare all'uso del CI non risolve alcun problema. Anzi probabilmente peggiora. Questo perché l'interpretazione dell'intervallo di confidenza è pure più difficile di quello del P value e più soggetta a frequenti misconceptions https:/
@no_lying_online @ciphergoth @ArthurB @skdh @kareem_carr But wouldn’t “CI” then stand for “credible interval” rather than “confidence interval”? Not that I’d complain. 😉 https://t.co/DhQfaaSmgS
"Confidence intervals..are based on philosophy that doesn't allow inferences about plausibility &doesnt utilize prior info. Using CIs as if they were credible intervals is an attempt to smuggle Bayesian meaning into frequentist statistics, without prop
@GidMK (But the width of a confidence interval does not necessarily reflect precision either. Some specific ways to construct confidence intervals may happen to have that property, but it’s not inherent to CIs.) https://t.co/DhQfaaAd2K
@apsmunro Different theories https://t.co/4sALShndvy "Frequentist CI theory says nothing at all about the probability that a particular, observed confidence interval contains the true value" "Frequentist theory is a “pre-data” ... Bayesian theory, on the
Sometimes you strike gold! This is—without exaggeration—an absolute must-read for everyone but a select few. https://t.co/MbYLGJO08Q
Morey et al., 2015: The Fallacy of placing confidence in confidence intervals. https://t.co/eBSxNfLydz
@vamrhein But on the other hand, it can be argued that confidence intervals as a general concept are the wrong tool for this. https://t.co/DhQfaaROrk
@ulfgardleo @gchahal What I am arguing is: https://t.co/hHwLT98XWQ
@ulfgardleo @gchahal What do you mean not accessible? I am talking about this: https://t.co/hHwLT98XWQ
@ulfgardleo @gchahal Why should it matter that it’s not a CI? https://t.co/DhQfaaROrk
@kareem_carr As a tangential note (no quarrel with the core point), it is debatable whether confidence intervals per se are the appropriate tool for this: https://t.co/DhQfaaROrk
@benryanwriter @gregggonsalves @ADPaltiel In particular, what they *don’t* imply. https://t.co/DhQfaaROrk
20. אבל בהמשך יש שפע הגדרות מוצלחות, תובנות ומוטיבציות. מקור מוצלח נוסף הוא המאמר הבא, שנפתח בציטוט מתוך הנסיכה הקסומה https://t.co/ZOh18jomRV Morey, Hoekstra, Rouder, Lee, and Wagenmakers. "The fallacy of placing confidence in confidence intervals." 201
@dikla_oren חד משמעית credible interval. הנה מאמר חביב שמדגים למה רווחים אחרים מטעים. https://t.co/MtlWwQEIFF
@GidMK On a tangential note, confidence intervals are not great at this. https://t.co/DhQfaaROrk
@willemsleegers This appears to commit the precision fallacy: https://t.co/P42Npre11w. Narrow CIs != precise knowledge
@cantabile here’s that CI paper I was talking about by @richarddmorey https://t.co/fPJcuJ0ZGk
@UnEmpiriciste « L’estimation de ce contrefactuel sans pass sanitaire est assez précise statistiquement puisque l’intervalle de confiance (à 95 %) […] » 🙁 https://t.co/JqCgQXvglk
@Pseudorandom75 @BendersCloud @RyanMarino I think I would phrase it as small numbers leading to less precise estimates, wider credible intervals, or tending to have more extreme values. Confidence intervals are probably not that intrinsically relevant. ht
@JohnMashey @gorskon @CT_Bergstrom @profkeithdevlin — Here is a 95% confidence interval generated from this model and those data. — What is the probability that the parameter lies in this specific confidence interval? — ¯\_(ツ)_/¯ What kind of question is t
@veritanonbugia @TeoMiz @mrmattjackson @GidMK And if you can’t, that’s not enough to claim that there isn’t a difference. That’s the fallacy of acceptance: https://t.co/opOGlZ4V9K
@nickmmark @kate_sills @GidMK One must be careful with this sort of interpretation: https://t.co/gyJawapblA
RT @Benguimbis: Encore plus exhaustif ici https://t.co/B6iFFOWBWD C'est ma ressources preferee apres toutes les analyses que je fais. @EpiE…
Encore plus exhaustif ici https://t.co/B6iFFOWBWD C'est ma ressources preferee apres toutes les analyses que je fais. @EpiEllie a aussi ce magnifique blog tres illustré https://t.co/i0z1aSKSlL
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
RT @EssianeNelson: Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance.…
Un papier intéressant évoquant les erreurs d'interprétation commune lors de l'utilisation des intervalles de confiance. #MustRead #EconTwitter "The fallacy of placing confidence in confidence intervals" https://t.co/NvnXv71Jqu https://t.co/yhB43UKhjK
@federicolois @Akustronique As far as I can tell, you are committing the “likelihood fallacy” or “fallacy of acceptance” (fallacy 3): https://t.co/zZhQHSrrgd And your 1/ seems to be the “fundamental confidence fallacy” (fallacy 1).
@JamesMaloneLee3 It would appear that confidence intervals also have their share of problems: https://t.co/5qajqSP8Uo
@UnEmpiriciste Et même sans intention de faire ça, les p-values et intervalles de confiance sont très propices aux mauvaises interprétations en premier lieu. Ces articles sur le sujet sont très intéressants : https://t.co/J6K0on3T4i https://t.co/JqCgQXvgl
@AdamJKucharski (In)famously, it doesn't mean that https://t.co/xbfCSViJM0 https://t.co/arhHBQxdcr
@poppe_stephan Bin mir noch nicht sicher, ob da bei mir der Groschen gefallen ist. Wir das hier mitdiskutiert? https://t.co/UXSb0DO3ng
CIは不適切だから他の方法使えって主張されてるね... えぇ...じゃあなんでみんな使ってるんだ... https://t.co/TiAM3SKa15
RT @JumpXav: @PhDemetri You may also like: https://t.co/E9VDdKp5e4 - from @richarddmorey - particularly the lost submarine analogy!
@PhDemetri You may also like: https://t.co/E9VDdKp5e4 - from @richarddmorey - particularly the lost submarine analogy!
what happened to @richardmorey. Equivalence tests and confidence intervals. His evil twin suggest that we shouldn't place much confidence in these results. Long live ACE. :) https://t.co/lstetf8Sk4
@dnunan79 @ADAlthousePhD @stephensenn @MaartenvSmeden One problem: the interpretation of CI is just as baffling as the interpretation of p values. See, for example "The fallacy of placing confidence in confidence intervals" https://t.co/3Wms78UvPX
"If confidence int do not allow an assessment of the prob that an interval contains the true value, if they do not yield measures of precision, and if they do not yield assessments of the likelihood or plausibility of parameter values, then what are they?"
@TomChivers @Lord_Quantock @dave_chivers A minor point but will you please follow your own lead (w.r.t. item 11) and correct the "confidence interval" misconception in item 6? https://t.co/arhHBQxdcr [§ "Fallacy 1 (The Fundamental Confidence Fallacy)"]
@ESMDcan123 @drjgauthier @gary_lyman @stephensenn CIs are a tough concept that I will always struggle with. @richarddmorey’s https://t.co/7VmmnLX3BW is a great place to start. The paper you linked shows the beneficial effects of adjusting (filled circles i
@StuartJRitchie There is also this paper! https://t.co/BQTpfug5PS
@MakeUsway @Chaay این هم منبع حرفم: https://t.co/OQZNiiK2Ov
#SPH604_Fall2020 I think it should not be a white and black scenario, but it is a good point to think about! The fallacy of placing confidence in confidence intervals https://t.co/Uqe6m63NHn
this is a good read for anyone who wants their head to explode: https://t.co/LDhBlSOFZC
@ATBollands @jonathanberman @sapinker After diligent and skeptical enquiry, I found Prof Pinker is correct again. See https://t.co/LHLWGUAE87 for full explanation. For a normally distributed random variable, possibly distinction without much difference.
RT @phl43: Several people have already explained why Pinker is right in the comments, but in addition I would like to recommend this paper…
RT @phl43: Several people have already explained why Pinker is right in the comments, but in addition I would like to recommend this paper…
Several people have already explained why Pinker is right in the comments, but in addition I would like to recommend this paper by @richarddmorey et al. (https://t.co/xynmD5L9Mi), which clarifies many common misunderstandings about confidence intervals.
@sam_ereira Good explanation under Fallacy 1 here: https://t.co/VJyE7j7UDj I think in the end it boils down to the difference between long-run and subjective probabilities.
RT @frod_san: Nice thread which shows how slippery the concept of confidence intervals really is, and how often we misinterpret them. Remi…
Nice thread which shows how slippery the concept of confidence intervals really is, and how often we misinterpret them. Reminds me of 'the fallacy of placing confidence in confidence intervals' paper https://t.co/G83bvcg60x #Statistics
Stumbled upon this great 2016 article by @richarddmorey et al that explains why it is wrong to interpret confidence intervals as (1) a measure of precision, (2) a set of likely values, or (3) X% confidence that the interval contains the true value. https
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
A good read about confidence intervals and why they probably (95% chance at least) don't mean what you think they mean.
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
RT @aeryn_thrace: Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follo…
Can we please make this common practice? If you post an eye-roll about a misconception in stats or methods, please follow it up with an explanation of why and what the correct interpretation is, if possible, with links, as below. Helps make twitter a ha
@sadneurons @richarddmorey @statsepi Not co-authored with @statsepi, but do you mean like this: https://t.co/iSdRJrvxAY or this: https://t.co/k0VOl3Nbmp or this: https://t.co/fqjKgxedfw or this: https://t.co/9IdKkJbP6x ?
RT @rlmcelreath: @matt_zefferman You know this paper? https://t.co/Ra0cdYKMgX Submarine example seems nice for teaching, but I've not give…
@matt_zefferman You know this paper? https://t.co/Ra0cdYKMgX Submarine example seems nice for teaching, but I've not given it a try yet in a live class. @richarddmorey may have some good advice.
@unnombrealazar @patilindrajeets It's even more complicated than that - see https://t.co/qGKDPPg6fI
brilliant epigraph on a paper about misinterpreting confidence intervals: "You keep using that word. I do not think it means what you think it means." Inigo Montoya, The Princess Bride (1987). https://t.co/tcSoB1dxtv
RT @TheStatsGeek: @apsmunro You might find this blog post useful https://t.co/KRm8hUyZ2O (or not!) and @richarddmorey et al's paper https:/…
@apsmunro You might find this blog post useful https://t.co/KRm8hUyZ2O (or not!) and @richarddmorey et al's paper https://t.co/gpBs3gV9id which motivated it
The fallacy of placing confidence in confidence intervals. https://t.co/N1UQTX95ul
@gnatgoodman @giladfeldman Good summary! In case you don't know this paper by @richarddmorey yet ... I think you might like it. https://t.co/gfU5eTdqcT
@NiklasJuha Das Beispiel ist im übrigen von dem sehr lesenswerten Artikel "The fallacy of placing confidence in confidence intervals" inspiriert wurden (siehe Konstruktion "the trivial interval" bzw. "a trivial procedure"). https://t.co/snLrNPiNwh
@emilyandthelime Got lots of useful tips in this thread a few weeks ago: https://t.co/GGyP1UAulV
RT @katiecorker: Fans of confidence intervals - what's your best example for teaching to explain why 95% CI does NOT equal 95% chance that…
RT @katiecorker: Fans of confidence intervals - what's your best example for teaching to explain why 95% CI does NOT equal 95% chance that…