RT @PloederlM: @JDaviesPhD https://t.co/qC74uoiNl9 So much evidence that nearly everybody has difficulties understanding statistical signif…
RT @PloederlM: @JDaviesPhD https://t.co/qC74uoiNl9 So much evidence that nearly everybody has difficulties understanding statistical signif…
I suspect this article is significant, but I distrust anyone who actually understands it! (Is there a P value implied in that?) https://t.co/R1WPmTHk6Q
For the numbers nerds
RT @PloederlM: @JDaviesPhD https://t.co/qC74uoiNl9 So much evidence that nearly everybody has difficulties understanding statistical signif…
@JDaviesPhD https://t.co/qC74uoiNl9 So much evidence that nearly everybody has difficulties understanding statistical significance.
RT @kaz_ataka: 大事。あまりにも多くの人が検定と統計的有意性(という言葉)に対する過信があることにリスクを感じる。ビッグデータ時代にp値を議論してどれほど意味があるのか、を学部までに考えさせておくべき。 Scientists rise up against st…
Scientists rise up against statistical significance. https://t.co/jX2dmeL2uV
Statistical inference in the 21st century: a world beyond P < 0.05’. “don’t say ‘statistically significant’” https://t.co/nm0DEF8hml
@dnunan79 @MaartenvSmeden Another useful reference: https://t.co/S8vogMDIV1
@crabb_vicki @andrew_croxford @jasonlschwartz @robinmonotti @jesscataldi @George3031301 @JohnTal6 The concern raised in that Nature article is EXTREMELY important. EVERY medical researcher should be aware of it. It is indeed pervasive: it results in live-s
@crabb_vicki @andrew_croxford @jasonlschwartz @robinmonotti @jesscataldi @George3031301 @JohnTal6 This issue of dismissing differences seen in the data is a serious one. Rather than saying, “There were differences observed, but there are not sufficient num
@crabb_vicki @andrew_croxford @jasonlschwartz @robinmonotti @jesscataldi @George3031301 @JohnTal6 Thanks for that. I’ve seen these types of comparisons also done with the same group but prior to vaccination. Either way two quite serious adverse events grea
RT @AazamVirani: When all is said and done, someone out there will examine a random sample of scientists and claim to reject the null hypot…
@JorgeLarangeir1 A bright-line fallacy em RCTs pequenos consiste no erro que você não cometeu, mas tantos cometeram e cometem: interpretar como "negativos" melhoras empíricas (do grupo tratado em relação ao placebo) porque p>0.05: https://t.co/8DAoyrlp
@allergyPhD In terms of critiques from "inside the house," this one got talked about among my quantitative colleagues. https://t.co/QNmRmKGFt1
Found this oldie; it's nice to know that scientists are grappling with the same kinds of issues that plague media coverage of polls. #polls #auspol Source: https://t.co/zGoL2VDkMw https://t.co/ChbiqtiEDW
@MaxCRoser @TomChivers Erros in statiscal reasoning seem to b very common even among scientists. Errors like using a "negative" trial as evidence of absence (when sometimes should be even evidence of effect) have real world implications in science (like ki
“Compatibility interval” sounds like a Tinder feature, but otherwise I love this article. This is why I say “statistically reliable” rather than “significant.” Even if we don’t reform p values, we can promote statistical understanding #academictwitter http
@paulg @kenshirriff Additionally, many (maybe most) smart cientists still make a mess with basic statistics, which might affect many lives, as discussed here: https://t.co/ET0pEz4HLj
RT @hispamed: “Scientists rise up against statistical significance” https://t.co/RbBU4v9JHo https://t.co/8tED09gm6D
Enjoying re-reading this accessible article from a couple of years ago about problems with how p-values are often used and interpreted: https://t.co/bvc46kM0fc H/T: @verena_rieser
@smarrupe @filipe_rafaeli @seleno_glauber @leandrotessler @UlissesPSampaio @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD Boa parte dos erros nas msgs incorrem na bright-line fallacy, já explicada aqui: https://t.co/8DAoyrlpBH
@smarrupe @coronaloover @filipe_rafaeli @seleno_glauber @leandrotessler @AbdenurFlavio @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD E ainda caíram no erro do paper da nature: "paper underpowered negativo = indicio de ausência de evidê
@seleno_glauber @smarrupe @leandrotessler @AbdenurFlavio @filipe_rafaeli @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD Caros, podem dar uma lida nesse texto? É bem curto porém muito relevante ao tema. https://t.co/ET0pEz4HLj
@smarrupe @seleno_glauber @leandrotessler @AbdenurFlavio @filipe_rafaeli @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD Caro, cuidado com a interpretação do p https://t.co/ET0pEz4HLj
@leandrotessler @seleno_glauber @smarrupe @AbdenurFlavio @filipe_rafaeli @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD Caros, podem dar uma lida nesse texto? É bem curto porém muito relevante ao tema. https://t.co/ET0pEz4HLj
@smarrupe @seleno_glauber @leandrotessler @AbdenurFlavio @filipe_rafaeli @Celso_Lemos @marivarella @clorexiduda @bcaramelli @MBittencourtMD Amanhã dou uma lida. No meio tempo sugiro leitura desse texto também. Fala sobre erros comum na interpretação do p e
https://t.co/1JMyBLW8G9 This applies to the reporting on the JAMA IVM study. The language of the study itself (w/o being technically inaccurate) encourages a negative interpretation…
Scientists rise up against statistical significance https://t.co/3lY4N7gte8
Fascinating “Factors such as background evidence, study design, data quality and understanding of underlying mechanisms are often more important than statistical measures such as P values or intervals.” https://t.co/jfYE3xrEGf #research #data #science #
RT @VandyCORE: Scientists rise up against statistical significance https://t.co/F5lkB6GGhF
Y (afortunadamente) cada vez se cuestiona más su uso en las publicaciones científicas. https://t.co/RVS3HV3Zn5
RT @SalfordHMR: Looking forward to today's journal club - Dave Howard on his favourite topic, the dreaded curse of p<0.05! 'Scientists ris…
Looking forward to today's journal club - Dave Howard on his favourite topic, the dreaded curse of p<0.05! 'Scientists rise up against statistical significance' (see link below). https://t.co/S6lpEBIEsh
@mc_hankins @AndrewPGrieve @stephensenn @lakens @uygun_tunc I'd call 800 signatories 'many'. https://t.co/3AvPtA3jwN Another plot twist: There are more than two sentences and some points that might be actually worthy of discussion in the paper. But I gue
@neely615 You might be interested in https://t.co/kqrD1uxMf8 ;)
@mc_hankins @AndrewPGrieve @uygun_tunc @tunc_necip Thht was your takeaway? Interesting. https://t.co/kqrD1uPnDI we are calling for a stop to the use of P values in the conventional, dichotomous way — to decide whether a result refutes or supports a scienti
RT @HopeFeldmanMD: @juliomayol https://t.co/3x9JWFttFm. Many of my mentors prefer Bayesian analyses.
@motazqadan Has crossing the Atlantic made you so cold hearted? https://t.co/WR3RUFwyB8
RT @HopeFeldmanMD: @juliomayol https://t.co/3x9JWFttFm. Many of my mentors prefer Bayesian analyses.
@juliomayol https://t.co/3x9JWFttFm. Many of my mentors prefer Bayesian analyses.
@mab_sp125 De qualquer forma, do pouco que aprendi até hoje por aqui, me parece que há um abuso na forma como do conceitos estatísticos como p são ensinandos e utilizados. Interpretação binária pode por vezes promover uso dos dados em sentido oposto: https
@sbenefield90 @AlexBerenson @pfizer @moderna_tx By the way statisticians have objected to this practice of ignoring the SIZE of an effect, based SOLELY on what the statistical significance says:
@federicolois @BallouxFrancois Read and learn. https://t.co/yN6wi6EOnE
RT @baumfran: Very important perspective - talks to limits of #epidemiology in #Publichealth Scientists rise up against statistical signif…
@schmittpaula Texto da nature abordando o tema: https://t.co/ET0pEz4HLj
Somehow I missed this great article. It seems to have been the most talked about scientific paper in 2019. https://t.co/Hy8wqydQhC
@tpahern @erikamedwards @DebJakubowski Here's a brilliant new paper by @Lester_Domes on multiple comparisons: https://t.co/smC5eWgN1A ... and, if you don't know it, you may check: https://t.co/CTesqcX8XV
RT @gary_lyman: Scientists rise up against statistical significance https://t.co/IOiwMJQMIn @NicoleKuderer @HarvardBiostats
@ProfMattFox @dianabuist @TimothyLash I quote some combination of: ASA statement: https://t.co/3WZ2m9v0w6 Amrhein et al Nature letter: https://t.co/uWR3kvt8Hr Greenland et al 2016: https://t.co/V8GOwywjp4
RT @gary_lyman: Scientists rise up against statistical significance https://t.co/IOiwMJQMIn @NicoleKuderer @HarvardBiostats
Scientists rise up against statistical significance https://t.co/IOiwMJQMIn @NicoleKuderer @HarvardBiostats
@nntaleb Shame that so many doctors and "experts" don't read papers and some who do still fail to understand some basic statistics/probability concepts (sometimes even the people who write the papers). This text from Nature exemplifies a common confusion:
@bumaBAgyo Because they're arbitrary? Because they're dead? 😂🤷🏻♀️ I don't understand either but I find comfort when I find something that disputes what I don't understand haha! Maybe @john_t_ormerod can help us understand? 🤭🙈 https://t.co/2urMaVFmKT
@DrToddLee @pedrocaiado3 @drjohnm @kaulcsmc @djc795 @PulmCrit @AndrewFoy82 @raj_mehta @mikejohansenmd @First10EM I don't know if the positive effects seen are real or not. But the p-value should not be the sole determining factor in that: "How do statistic
@EdLatimore Basic probability and logical concepts get misinterpreted even by experts who read and produce scientific papers. https://t.co/ET0pEz4HLj
@DrToddLee @drjohnm @brophyj .@boulware_dr If by “Bayesian” you mean going beyond just statistical significance then I’m all in favor. Many researchers now want to go beyond the hegemony of the p-value: https://t.co/xIcnBS1Xzw
How often do we get obsessed with p <0.05! Scientists rise up against statistical significance https://t.co/5waedn0X56
Scientists against the idea of statistical significance: https://t.co/ISuCDQwG3L
@Crude_Future @Clara_Locher @philippejuvin J'ai lu il y un but ds temps des articles justement intéressants sur l'utilisation malfaite du p-value https://t.co/IPfSMJHwtQ https://t.co/s0sZuoxB5e Et le top des articles https://t.co/1Vxm28HLFQ
RT @JeanPatrickT: @EssiColo @clicjf Cet autre article paru dans Nature en 2019 était également fort intéressant https://t.co/tQRub64ab7
@EssiColo @clicjf Cet autre article paru dans Nature en 2019 était également fort intéressant https://t.co/tQRub64ab7
Fort éclairant, en effet. Ah! le fameux "statistiquement significatif" versus "biologiquement significatif"!! D'ailleurs, article fort intéressant paru dans Nature en 2019 sur la question....👇🏾 https://t.co/tQRub64ab7
Scientists rise up against statistical significance https://t.co/LFIYCugaxi
@Sonic_urticant La p-value, c'est en effet pas mal pour prendre une décision. Par contre, ce n'est pas adapté si on veut savoir si quelque chose est vrai ou faux, car on n'a alors pas de décision à prendre En réalité, je pensais à cet article quand j'ai é
Old but current. Have we all embraced this? Any flaws? https://t.co/RFODdxr5gY @NatureNews @NatureComms
Supergod artikel om hvorfor vi skal droppe dikotome p-værdier (> el. < 0,05) og i stedet diskutere point-estimater og den kliniske betydning af alle udfald indenfor konfidensintervallet. Tak @Per_Damkier!
Scientists rise up against statistical significance https://t.co/oup7M6ea15
@Medicinmanden En pæn lang historie som har været genstand for en til tider *host* ophidset debat. Et godt udgangspunkt https://t.co/KYHvOqWvh3
@WithoutDepth I think the scientific community is somewhat aware of the issue though - see the discourse around misuse of p-values https://t.co/ClMqArec4a or the psych replication crisis https://t.co/I6nxVqG9X2 or clinical trial design
@tfleury4 @luannavargas @mab_sp125 @jadervinicius_ @AbdenurFlavio Assim como no caso da vacina, o p alto quer dizer apenas a evidência possui alta incerteza para a métrica em questão. É o que explicam nesse texto da Nature: https://t.co/ET0pEz4HLj
@mab_sp125 @luannavargas @tfleury4 @AbdenurFlavio Entendo que não seja. Mas acho também que nesse ponto há um erro de interpretação quanto ao significado do p. Um erro que vi muito na mídia e inclusive na comunidade científica. Pra mim essa é o ponto centr
@AkshaySharmaMD Did you mean #phacking? 😉 You probably saw this, something to re-read every once in a while... @vamrhein @Lester_Domes https://t.co/QU7YeDG4Y1 A Bayesian perspective @StatModeling https://t.co/WYeL3klbLq
RT @luannavargas: This text (Source: Nature) should be mandatory to read for anyone who wants to use statistics to make decisions. In Medi…
RT @juan_francomd: Se sigue citando este artículo sin entenderlo en su contexto. La preocupación de los estadísticos del artículo de Nature…
"The misuse of statistical significance has done much harm to the scientific community and those who rely on scientific advice."
"Let’s be clear about what must stop: we should never conclude there is ‘no difference’ or ‘no association’ just because a P value is larger than a threshold such as 0.05 or, equivalently, because a confidence interval includes zero."
"[...] las inferencias deben ser científicas, y eso va mucho más allá de lo meramente estadístico." Discusión interesante sobre el valor P y su interpretación distorsionada en la toma de decisiones. Scientists rise up against statistical significance http
RT @juan_francomd: Se sigue citando este artículo sin entenderlo en su contexto. La preocupación de los estadísticos del artículo de Nature…
RT @juan_francomd: Se sigue citando este artículo sin entenderlo en su contexto. La preocupación de los estadísticos del artículo de Nature…
RT @juan_francomd: Se sigue citando este artículo sin entenderlo en su contexto. La preocupación de los estadísticos del artículo de Nature…
Se sigue citando este artículo sin entenderlo en su contexto. La preocupación de los estadísticos del artículo de Nature es la sobrevaloración de la p para tomar decisiones, sí, pero principalmente cuando <0.05 (sesgos, p hacking, etc.). 👇
RT @talithafs: Relendo https://t.co/F6ah6QvFXw
RT @norabar: (En inglés) Marzo de 2019, en @nature. Más de 800 investigadores advierten sobre el mal uso del "p-valor" 👇 https://t.co/eLr0…
@norabar @nature Si, lo leí pero no encuentro el ejemplo que preguntaba. https://t.co/moJZRMzo7b
RT @norabar: (En inglés) Marzo de 2019, en @nature. Más de 800 investigadores advierten sobre el mal uso del "p-valor" 👇 https://t.co/eLr0…
RT @norabar: (En inglés) Marzo de 2019, en @nature. Más de 800 investigadores advierten sobre el mal uso del "p-valor" 👇 https://t.co/eLr0…
RT @norabar: (En inglés) Marzo de 2019, en @nature. Más de 800 investigadores advierten sobre el mal uso del "p-valor" 👇 https://t.co/eLr0…
RT @norabar: (En inglés) Marzo de 2019, en @nature. Más de 800 investigadores advierten sobre el mal uso del "p-valor" 👇 https://t.co/eLr0…