#statisticalinterference ผลการค้นหา

@FBIDirectorKash @FBI @WhiteHouse Can you look into this? Statistics can be manipulated. x.com/MarioNawfal/st…

🇺🇸🇦🇫 FBI CRIME STATS ARE STRAIGHT-UP RIGGED They just classified the Afghan national who tried to bomb Texas as white. That’s not a glitch. That’s the rule. FBI still lumps every Moroccan, Syrian, Iraqi, Iranian, and Afghan into the “white/Caucasian” bucket because of a 1940s…

MarioNawfal's tweet image. 🇺🇸🇦🇫  FBI CRIME STATS ARE STRAIGHT-UP RIGGED

They just classified the Afghan national who tried to bomb Texas as white.

That’s not a glitch. That’s the rule.

FBI still lumps every Moroccan, Syrian, Iraqi, Iranian, and Afghan into the “white/Caucasian” bucket because of a 1940s…


And this is the big point, a lot of non statisticians applying statistics just try to find a model that gives them a significant result, without realizing the multiple testing, poor assumptions, ignorance of network effects/fat tails


If D and S don't follow the same distribution, then during finetuning S would override D's training, a concept known as catastrophic interference meaning pretraining on D would have no influence on model performance if the model is subsequently finetuned on an unrelated dataset S


The sample is what is (hopefully) random. Even with no true effect/association, a sample can be “unlucky” and look extreme. As n grows, large apparent effects due entirely to sampling noise become exceedingly rare, as they should. Statistical models just formalize this.


Um, how would you mitigate that? Robust statistics?


Tell me about it. LIke those statics used by those making unsubstantiated claims how certain minorities get killed by cops more than any other ethnicities, when it's provably false. Also statistics can be used to prove/disprove anything. They're utterly meaningless.

Sabamika1's tweet image. Tell me about it.
LIke those statics used by those making unsubstantiated claims how certain minorities get killed by cops more than any other ethnicities, when it's provably false.
Also statistics can be used to prove/disprove anything.
They're utterly meaningless.
Sabamika1's tweet image. Tell me about it.
LIke those statics used by those making unsubstantiated claims how certain minorities get killed by cops more than any other ethnicities, when it's provably false.
Also statistics can be used to prove/disprove anything.
They're utterly meaningless.

"When drawing a sampling frame, it should have a variable of interest. Without a good frame, your findings can never be comparable.", Dr Ssennono says. If the sampling process is faulted, the results can never be trusted; the future of statistics lies on the robust methodology.

Mramooti0's tweet image. "When drawing a sampling frame, it should have a variable of interest. Without a good frame, your findings can never be comparable.", Dr Ssennono says.
If the sampling process is faulted, the results can never be trusted; the future of statistics lies on the robust methodology.

Giorgos Bakoyannis, Aristofanis Rontogiannis, Ying Zhang, Wanzhu Tu, Ann Mwangi, Constantin T. Yiannoutsos. [statME]. Robustness intervals for competing risks analysis with causes of failure missing not at random. arxiv.org/abs/2511.20980…


Statistics are what happens when you try to merge pure mathematics with human biases, they borrow the credibility of maths only to bastardise it using poor data and unconscious bias before presenting it as truth.


"And manipulation of statistics is an underexplored angle." I've seen stats manipulated via trial designs where the trials were designed to fail. Statisticians had to be involved in those designs.


Matrix interference can obscure out-of-specification results, risking the release of noncompliant batches. ▶️Watch our on-demand webinar for a case study highlighting practical strategies to overcome matrix effects in pharma analysis. Click here: ow.ly/Osqm50Xt2n8


On Evolution-Based Models for Experimentation Under Interference ift.tt/3YBczeL


That is the problem. Statistics can be manipulated. If the police and the lawyers don't understand statistics they won't recognise when statistics have been misused.


“Although various approaches have been proposed to improve the presentation and interpretation of statistically nonsignificant findings, a widely accepted consensus has not emerged, as these approaches have yet to be systematically tested for their practicality and validity.…


🔓Misspecification-robust likelihood-free inference in high dimensions #OpenAccess article by Owen Thomas, Raquel Sá-Leão, Hermínia de Lencastre, Samuel Kaski, Jukka Corander & Henri Pesonen published in Computational Statistics doi.org/10.1007/s00180…

SpringerStats's tweet image. 🔓Misspecification-robust likelihood-free inference in high dimensions
#OpenAccess article by Owen Thomas, Raquel Sá-Leão, Hermínia de Lencastre, Samuel Kaski, Jukka Corander & Henri Pesonen
published in Computational Statistics
doi.org/10.1007/s00180…

🔍 Suspect citation manipulation? Learn how basic stats can help editors spot "citation stacking" and protect research integrity. Quick read: sciedit.eu/stat-tools-cit… #ResearchIntegrity #AcademicTwitter #PeerReview #ScienceEditors #OpenScience


ไม่พบผลลัพธ์สำหรับ "#statisticalinterference"

I pass this Baskin-Robbins sign everyday & can't help but think of @LifeAtPurdue STATS 301 class. #datacorrelation #statisticalinterference

ChelseaAmbrizTV's tweet image. I pass this Baskin-Robbins sign everyday & can't help but think of @LifeAtPurdue STATS 301 class. #datacorrelation #statisticalinterference

Loading...

Something went wrong.


Something went wrong.


United States Trends