User:Sj/scifraud

Retraction Watch

 * All-time leaderboard
 * COVID catalog

Distributed peer review

 * PubPeer - Catalog of public peer review. Little differentiation or structure annotation of individual comments

Fraud-analysis blogs

 * Data Colada - home of early p-hacking analyses. Simmons + Simonsohn
 * For Better Science: fraud and more. ELSE: pandemic profiteering

News overviews + monitoring orgs

 * The Disinfo Dozen
 * Anti-Vax Watch

COVID
Sources of popular COVID hoaxes, conspiracies, or statistical fraud that managed to get into mainstream journals.

General

 * RW's COVID catalog
 * CoronaCentral

Harald Walach
Superspreader event: June 25 – July 15, 2021. Two mainstream papers, millions of views.

antivax conspiracy:


 * Statfraud: lying about what can be inferred from side-effect reports; applying common antivax 'talking points' (misuse of stats) to public data; concluding vax harms are 5 magitudes higher than reality.
 * Retraction: retracted from MDPI Vaccines after 1 week.

COVID conspiracy: against masking in schools


 * Fraudulent presentation of flawed methodology, lying about implementation details where no underlying data could be provided. Misstating implications of experiment in the analysis.
 * Retraction: retracted from NEMJ Pediatrics after 2 weeks.

Ronald Kostoff
Superspreader event: ~Sept 20 2021. Elsevier Toxicology Reports paper. Tw by 9/20?, FB by 9/21. top 5% on Altmetrics by 10/15.

Background: Navy? Georgia Tech [partly removed late 2021, all old cranky PDFs still up w/ high grank]. Lately, published only on SS sites like DClout.

5G conspiracies: claims that 5G is dangerous, using a combination of statistical fraud, extended misinterpretation of data, and forking-paths


 * Statfraud: 5G intensities are 2 magnitudes lower than 3G; most radiation continues to be from uplinks and not downlinks (by 3-5 magnitudes); these are handwaved away.
 * Physfraud: claiming that resonant effects over short timescales could have some significant impact on cells. No such mechanism is known.
 * Experimental design fraud: numerous real experiments are done with cells, animals, and humans to evaluate potential impacts of exposure to radiation. (none have identified any impact, nor mechanism, for the kinds of harm suggested by conspiracy theories like this). This paper does no experiment, just carries out a hand-waving analysis, but presents itself as an experiment testing a hypothesis.

Contributing-factor fraud: exaggerating potential compound effects, searching a space of thousands of potential cofactors in disease, encouraging hypochondria.


 * Misrepresenting what can be inferred from text + data mining of a large corpus of papers.
 * A sequence of similar papers. Linking to any and all toxic exposure; linking to any other disease
 * Trying to link COVID to nicotine, complementing other work suggesting nicotine is beneficial as a prophylactic against COVID. (excluding nicotine from the list of contributing factors, in an analysis where nicotine showed up as the most commonly-suggested c.f.)

Antivax conspiracies: exaggerating potential side effects, lying about reported or measured harms, lying about the known morbidity and mortality of COVID


 * Statfraud: lying about what VAERS reports represent, misestimating implications for potential negative effects (by 4 magnitudes); lying about how deaths are attributed to causes, misestimating implications for mortality of COVID (by 2 magnitudes); hiding morbidities of COVID, pretending they don't exist.  Misrepresenting the idea of risk/benefit calcumations (by 1-2 magnitudes). concluding vax harms are 6-7 mags higher than reality.
 * Medfraud: presenting treatments with no significant measured positive effects as clearly beneficial; presenting treatments with extensively measured positive effects as harmful; encouraging use of fringe remedies and disregarding the recommendations of doctors.
 * COI: getting papers reviewed by friends or frequent co-authors
 * Recurrence: after 3 weeks of popularity in the public eye, published an update amplifying the fraud by 3x (on a news blog site)

Poulas
Nicotine (w/ Farsalinos): co-author of a series of ~ a dozen? articles all presenting the same sort of analysis of nicotine as potential ameliorating factor in COVID.


 * Post-hoc: No recurring data backs this up; all analyses rely on a post-hoc observation of less-than-expected incidence of smoking among early Chinese COVID cases.
 * Amplification: A single stats anomaly parlayed into a large # of papers and guided-confirmation studies
 * COI: authors never mention their funding via the NOSMOKE institute, an ironically named org whose funding comes from tobacco. Getting papers reviewed by friends or frequent co-authors.  Reviewing papers (Elsev Toxicology Rep./Notes), or co-editing special issues (MDPI COVID), presenting fringe COVID theories.
 * Retractions: One of the dozen duplicative papers was published in the European Respiratory Journal, then retracted for unstated COI. Possibly a reviewer of Walach's Vaccines paper; defended it vocally online post-retraction.

Farsalinos
Most distinguished research.

Nicotine: see Poulas. shared retraction; not actively part of the Institute

Other COVID: see Poulas


 * Washing: switching languages and countries. Most awards in Ru, writing in Gr, cited papers in En
 * Review + editing: regular editor for special issues
 * Commentating: CreteTV (Sahinis), Kontra24 (Savidis), Open TV (Savidis), Alert TV (Velopolos)

Tsatsakis
Toxicology: as Editor-in-Chief, making self co-author on dozens of papers published in own journals

5G: see Kostoff.
 * Self-dealing: adding self to papers submitted to own journal. Frequent co-authors: Kostoff [14 of last 17 papers], Poulas, Farsalinos.
 * Peer-review ring: arranging for frequent co-authors to have papers pass peer review in own journal, even when not up to statistical or methodological norms of the field [or past issues of the journal]. encouraging submitting authors to cite other papers by authors in the network
 * Editing ring: arranging for frequent co-authors to be handling editors, guest editors, or special-issue editors of own and other journals
 * Publishing ring: arranging for papers to be submitted to publishers run by friends, who benefit from APCs, citations, and volume.
 * Retraction:

Antivax: see Kostoff.

Ioannidis
From contrarian to outright fraud, including in his papers on the prevalence of fraud? Applied to COVID statfraud + commentating in many languages.


 * Statfraud re: US policy changes.
 * Review + editing: incl. bullying of opponents, profs and grad students.
 * Commentating: Fox, CreteTV

Briggs
Statfraud blog. Mixes good + bad observations with bad synth + motivated reasoning for uniformly false results.


 * Mostly COVID greatest hits : hiding disease morbidity + mask effects, inventing vaccine morbidity + mask harms, flu vogs.

Antivax pundits + searchers

 * Neil Z Miller - Runs own antivax institute; only affiliation. Regularly published in a handful of journals with av-friendly editors
 * Bret Weinstein - Popcast, brilliant brother, IM
 * Steve Kirsch (raised $3.5M for his own fund to invest in alt treatments, liz)
 * RFK Jr.
 * Rogan?

Undead antivax pundits

 * Herman Cain (reddit)

Antivax Doctors
This is its own broad category -- people with MDs in something who have bcome talking heads for antivax campaigns. Growing in size and scale as the Borgnet catches on, hooks them up with neutraceuticals + outrage-sites, both directly funding + monetizing their niche


 * RWMalone
 * Naomi Wolf: social + DailyClout
 * Zelenko [1]
 * Dr. Drew

Confusion more than fraud

 * ivmmeta and related sites for covid treatments -- stats confusion around how to conduct a meta-study, presented with great confidence. Similar approaches can find confirmation of effect of any highly-studied, low-morbidity treatment (most neutriceuticals).  Site maintainers seem particularly partial to ivm (in terms of study inclusion/exclusion and choice of metric to observe in each)
 * Slate Star Codex ran down a full overview of how it's possible to produce apparent meta-evidence for the benefits of any random effect (vitamins, home remedies) unless there's actually a significant negative effect, and concludes that for Ivermectin in particular the observed result may be a reflection of its actual widespread use to treat parasites. (so: a real small positive effect in some populations, misattributed + misgeneralized)

MOS:MIRACLE
"And it is at this point that we have an entirely new physics..."

Started on the LK-99 talk page

Reflecting on articles about various miracle-cure physics and science articles: we could use a style for tagging / talking about claimed advances that are not 'breakthroughs' in the sense of "biggest advance in years" but 'breakaways' in the sense of "sudden inexplicable advance, deviating from known models with limited details, intermediate stepping stones, or theoretical underpinning". For detailed thoughtsm see User:Sj/miracles.