← POLYGENIK ISSN PENDING

Methods+Madness

The Journal of the Unpublishable Published twice yearly. Numbered. On paper. Vol. 1 · Edition of 1,000
The First Issue · Date to be announced

journal of negative results, failed experiments, and the things this collective did on purpose.

Peer reviewed where it matters. Numbered, capped, printed on paper that outlasts a server. Publication bias is the best-documented pathology in science. This is where the results that were supposed to disappear are kept instead.

Read free Buy the object
Title
Methods + Madness
Volume
One
Edition
1,000 numbered copies
Frequency
Two issues per year
Publication date
To be announced
Editor
To be named
Publisher
methodsandmadness.press
Model
Free online. Scarce in print.
Peer review
Negative Results section only
Front Matter

The Position

p. 003

Every collective of one-night events leaves behind an archive it does not own. This is our attempt at the opposite: a record that is ours, that is scarce, and that admits when the experiment failed.

Negative results have almost nowhere to go. A journal of failures has no prestige to trade, so nobody builds one, so the failures are repeated at scale, forever, by people who never learned they were failures. That is publication bias, and it is expensive.

A journal of failures produced as a fashion magazine — on heavy stock, with real peer review, in a numbered and capped run — has a different kind of prestige. It is the only kind we could stomach charging for. The science stays open online. The object stays rare. We think both facts are load-bearing.

We will not launch this issue in print until an editor who is not the collective agrees to sign it. A journal edited by the collective it documents is a zine. A zine is a fine thing. This is not that. The masthead therefore reads Editor: to be named, and it will stay that way until it is true.

Standing Section 01 · Peer Reviewed

Negative Results

Two to four papers per issue
pp. 010–041

Real submissions, really peer reviewed, really published. The results that were true, uninteresting to a career, and therefore never written down anywhere else. Each paper below carries the reviewers' names and its own rejection history.

NR·01

No detectable effect of a widely marketed nootropic stack on working memory across 214 healthy adults: a pre-registered null

A. Okonkwo, L. Serrano, R. Feld · Community Lab Cohort · Received under C7 · Two rounds of review

Abstract

We tested the eight-ingredient formulation most frequently discussed in self-experimentation forums against placebo in a pre-registered, double-blind crossover (n = 214, twelve-week washout). Across n-back accuracy, reaction time, and a delayed-recall composite, we observed no effect distinguishable from zero at any dose (all 95% confidence intervals spanning the null. Largest standardized effect d = 0.04). Subjective alertness rose in both arms identically. We report the full negative in the interest of the several thousand people currently buying this compound monthly. The pre-registration predicted a moderate effect. It was wrong, and we are the only people who will tell you so.

Pre-registeredNull resultn = 214Open data
NR·02

Fourteen attempts to reproduce a headline gut-microbiome–mood correlation, and why none of them worked

M. Bright, S. Oduya · Depositor-linked cohort · Received under C7 · One round of review, one dissent

Abstract

A frequently cited 2019 finding reported a strong association between a single bacterial genus and self-reported low mood. Using three independent sampling protocols and two sequencing pipelines across fourteen re-analyses, we recovered the association exactly zero times. The original signal appears to have been a batch effect confounded with collection season. We publish the fourteen failures in full, including the two where we were briefly convinced we had found something and had not. The dissenting reviewer argues our sampling window was too narrow to close the question. We agree it is not closed. We argue it is a great deal more open than the literature currently admits.

Failed replicationBatch effect14 attemptsContested
Standing Section 02 · Reprinted in full

Verdicts

From the Polygenik IRB
pp. 044–052

The Polygenik Institutional Review Board's determinations for the period, reprinted in full with dissents. The board has no power and issues its opinions anyway. Its verdicts carry no legal weight whatsoever.

Case 001 · Submitter: Polygenik (self-referral) Sitting: First Quarter

Determination on "Communion" — a shared-culture drop

Approved with Conditions

The panel finds the proposed work outside the Common Rule and within its own advisory jurisdiction. It is approved subject to four conditions: written informed consent from every participant in plain language; a named on-site clinician for the duration; exclusion of anyone under 18; and destruction of all biological material within 72 hours. The conditions were imposed and, per the collective's own reporting in Case Files this issue, followed. The panel notes that a collective publishing an adverse condition against its own drop is the only reason to trust the next verdict.

Dissent · one seat

I would have withheld approval pending a written destruction receipt rather than a stated intention. Intentions are not evidence. The receipt now exists. My objection stands as a matter of sequence.

Case 007 · Submitter: withheld Sitting: Second Quarter

Determination on a proposed self-administered peptide protocol

Not Approved

The submitter proposed an unsupervised self-experiment involving repeat self-injection of a compound with no human safety data and a plausible immunogenicity risk. The panel declines to approve. It notes explicitly that this determination is an opinion, carries no legal force, and cannot compel the submitter to stop. It observes only that the submitter's own reversibility answers on the form were blank, and that the act of leaving them blank is the finding.

Standing Section 03 · Documented properly

Case Files

The drops, on the record
pp. 056–071

The events, documented as protocols rather than press: what was planned, what happened, and the specific place where reality diverged from the run-of-show. The gap between the two columns is the article.

Communion — Drop 002

One night · 88 participants · IRB Case 001, Approved with Conditions

Protocol

Participants contributed a cheek swab on arrival. Samples were pooled, cultured on-site behind glass, and the resulting shared plate was photographed hourly and destroyed at 72 hours per the IRB conditions.

What went wrong

Contamination. A single unmasked contributor introduced a mold that overtook four plates by hour nine. The "shared culture" everyone watched grow was, for the back half of the night, mostly one person's kitchen.

What we learned

The failure was more legible than the success would have been. Guests understood contamination the moment they saw it eat the plate. We kept the ruined plates in the photographs. We did not reshoot.

The Celebrity Meat Tasting — Drop 003

One night · 40 seats · Withdrawn menu item, documented

Protocol

A tasting built around cultured cell lines with consented, non-identifying provenance. Twelve courses, each plated with a printed provenance card and a consent reference number.

What went wrong

Course seven never left the kitchen. The provenance chain for one line could not be verified to the standard we had published, so we pulled it mid-service and told the room why, out loud, between courses.

What we learned

Announcing the pulled course was better theatre than serving it. The rigour was the piece. This issue reprints the empty provenance card as the course-seven plate.

Standing Section 04 · Reader-submitted, anonymous

Reviewer 2

The worst real comments
pp. 074–079

A column of the worst real peer-review comments any of you have received, submitted anonymously. This is the most-read page in every issue. We print them verbatim and we print nothing that identifies the reviewer. Feeds directly off Reviewer 2 (reviewer2.wtf).

"The authors have clearly not read the foundational literature in this area, by which I mean my 2011 paper, which they have not cited."
Cell biology, revise and resubmit
"I did not have time to read past the abstract, but I am confident the conclusions are overstated."
Neuroscience, reject
"The English is fine. The science is the problem. I cannot say more without revealing who I am."
Genomics, major revision
"Reviewer 1 recommends acceptance. I do not know Reviewer 1, but I disagree with them on principle."
Immunology, reject
"The requested experiment would take approximately three years and a facility the authors do not have. I look forward to seeing it in the revised manuscript."
Structural biology, major revision
"This is a negative result and therefore, respectfully, not a result."
Submitted for this issue, framed and hung in the office
Standing Section 05 · From The Freezer

The Manifest

Sealed contents, public register
pp. 082–087

Selected entries from The Freezer — the public minus-eighty — reprinted with the depositors' own statements about why they sent what they sent. The contents stay sealed. The reasons do not.

SlotDepositorContentsDepositedIn their words
017WithheldSourdough starter, est. 2016Y1·Q1 "It is older than my marriage and has been more reliable. If everything else fails I want the bread to survive me."
044D. MarloweA dead cat's whiskers, sealedY1·Q1 "I could not bury all of him. This is the part I kept back for reasons I still cannot explain, which is why it belongs somewhere that keeps things without asking."
061WithheldCord blood, single donorY1·Q2 "The private bank quoted me a mortgage. Two dollars a month and a number I can read on a website felt more honest than a brochure."
088R. Iqbal (commissioned)A vial of tears, collected over one yearY1·Q2 "An index of a bad year. I want it archived, not felt. Sealing it is the point. Retrieval rights are the mercy."
Standing Section 06 · The fashion editorial

Retracted

Twelve pages, shot on textiles
pp. 090–101
Look 04 · "Duplicated Bands" — a floor-length coat printed with a lane repeated six times, the retraction's original evidence worn as pattern.

Camouflage cut from the evidence

Twelve pages of fashion editorial, shot on textiles printed with figures pulled directly from retracted papers. The recurring motif is the duplicated Western blot — the same band spliced across lanes it was never run in, the single most common fingerprint of image manipulation — rendered at fabric scale as camouflage. Worn, it does exactly what it did in the paper: it hides in plain sight until you know what you are looking at.

The garments are unwearable in the useful sense. They are meant to be photographed, printed on heavy stock, and kept. The styling notes on the reverse cite the case number and the date of retraction for every figure used.

Figures sourced with the work of the retraction watchdogs. Credited by name and case in the printed edition.

Standing Section 07 · In memoriam

Obituaries

Failed biotech, written properly
pp. 104–109

Failed biotechnology companies, written as proper obituaries: cause of death, survived by, in lieu of flowers. We do not use real company names. The deaths, however, are composited from real ones.

Halcyon Therapeutics, Inc.

Founded Y-9 · Wound down Y-1 · Aged 8

Halcyon Therapeutics died peacefully in its Series C, surrounded by a syndicate that had stopped returning its calls. Once valued at nine figures on a single surrogate endpoint, it is survived by a confirmatory trial that read out negative eleven weeks after the term sheet lapsed.

Cause of death
Accelerated approval, then withdrawal. The drug worked on the biomarker and not on the patients.
Survived by
Four hundred participants, a landlord, and a mechanism of action that remains, technically, elegant.
In lieu of flowers
The family asks that you read the confirmatory data before you write the press release.

Meridian Genomics

Founded Y-6 · Dissolved Y-1 · Aged 5

Meridian Genomics passed after a long illness widely misdiagnosed as product-market fit. It leaves behind a platform that could sequence anything and a market that wanted only one thing, cheaper. The pivot deck, its ninth, is being held in lieu of a headstone.

Cause of death
A platform in search of a product, outrun by an incumbent who simply lowered the price.
Survived by
Its intellectual property, now grazing quietly in a patent troll's portfolio.
In lieu of flowers
Please do not build another platform. Build the one product first.
For contributors · Two paths

Submit

Different rules apply
see C6 & C7

There are two ways in, and they are not the same door. Negative results are peer reviewed. Everything else is commissioned. You keep your work: contributor agreements license, they do not assign (C6).

Negative Results

Peer reviewed · Policy C7

Real methods, honest nulls, failed replications. We review for rigour, not for excitement. A well-run negative belongs here even — especially — when it is dull.

Demo form. Nothing is sent, logged, or stored. Real submissions open with Vol. 1.

Received (as a demo). No data left your browser. Under C7 a real submission would enter peer review with two reviewers and a published determination.

Everything Else

Commissioned · By pitch

Obituaries, case files, the Reviewer 2 column, editorial. We commission these. Send the idea, not the finished thing — a paragraph is plenty.

Demo form. Nothing is sent, logged, or stored.

Received (as a demo). No data left your browser. Commissioned pieces begin with a short conversation, not a contract.
The object · Free online, scarce in print

Subscribe

Two issues a year
Numbered · Capped

The science should be open and the object should be scarce. Read every paper online, forever, for nothing. Or hold the reason we will still exist in ten years. Edition of 1,000. When they are gone, they are gone.

Checkout is a demo. No payment is taken and nothing is stored. Vol. 1 goes on sale on a date to be announced. The online archive is free now and always.