Science

Current Picks

One Peptide Synthesis Reagent Swapped 11 of 20 Enzyme Activity Curves

By Jonas Eriksen / Jun 7, 2026

Substituting COMU for HATU in peptide synthesis shifted 11 of 20 enzyme activity curves. A cautionary tale about reagent trace effects in biochemical assays.

One Sediment Sieve Mesh Size Alters Three Paleoclimate Sea Surface Reconstructions

By Karim Osman / May 29, 2026

A single change in sieve mesh size—from 150 to 125 micrometers—can shift paleoclimate sea surface temperature estimates by up to 3°C, a new study shows.

Ten-Laboratory Replication Test Confirms Three of 18 Mouse Olfaction Studies

By Karim Osman / May 29, 2026

A ten-laboratory replication effort tested 18 mouse olfaction studies and confirmed only three. The findings highlight the role of incentives, funding, and infrastructure in shaping scientific results.

How Daniel Kahneman’s 1972 Lecture Stopped a Replication Crisis Before It Started

By Karim Osman / Jun 8, 2026

In 1972, Daniel Kahneman gave a small lecture on small samples that foreshadowed psychology's replication crisis. Decades later, his warnings ring true.

Room‑Temperature Superconductor Claims and the US$ 2 Million Replication Trap

By Jonas Eriksen / Jun 7, 2026

The LK-99 saga exposed a painful truth: verifying a room-temperature superconductor claim can cost US$ 1–2 million per lab. This article explores the replication trap in condensed-matter physics and how skewed incentives delay progress.

One Lab's Coated Stir Bar Contaminated 13 of 20 Organic Synthesis Runs

By Renu Shah / Jun 7, 2026

A routine stir bar leaching iron ruined 13 of 20 syntheses. The contamination went undetected for months, exposing how cheap consumables and publication incentives undermine reproducibility.

One Parameter Shifted 12 of 18 Muon g-2 Simulation Results

By Renu Shah / May 29, 2026

A single parameter in hadronic vacuum polarization calculations shifted 12 of 18 Fermilab muon g-2 simulation results, creating a 4.2 sigma tension with the Standard Model and sparking a debate between lattice QCD and dispersive methods.

One Grant Agency Rule Shift Changed Fourteen Computational Reproducibility Studies

By Renu Shah / May 29, 2026

In 2023, the NSF mandated code archiving for computational research. The rule triggered 14 reproducibility studies, revealing that only 30% of code ran without errors. This article examines the policy's impact on funding, workflows, and the future of reproducible science.

Seven-Laboratory Replication Confirms Two of 19 Primate Theory of Mind Studies

By Renu Shah / May 29, 2026

A seven-laboratory consortium replicated 19 primate theory-of-mind studies, confirming only two. The findings underscore the need for collaborative replication in comparative cognition.

Brenda Milner's Single Case Study Rewired Memory Circuit Maps

By Jonas Eriksen / May 29, 2026

How Brenda Milner's single-case study of patient H.M. dissociated memory systems, shifting neuroscience from lesion mapping to circuit dynamics and shaping modern connectomics.

Eight-Laboratory Replication Test Confirms Two of 20 Frog Electrophysiology Studies

By Karim Osman / May 28, 2026

A consortium of eight labs re-ran 20 frog electrophysiology studies, finding only two fully replicated. The project reveals hidden variability in model-organism research and offers lessons for improving reproducibility.

One Array Feed's Beamformer Resolved 11 Faint Fast Radio Bursts

By Renu Shah / Jun 7, 2026

CHIME's phased array feed and beamformer detected 11 faint fast radio bursts missed by earlier surveys, revealing a larger population and raising questions about FRB luminosity.

How One Sediment Grain Size Cut Shifts Two Paleoclimate Reconstructions

By Karim Osman / May 29, 2026

A tiny procedural choice—the sieve cut used to separate silt from sand—can produce opposite climate signals from the same lake sediment core. This article traces how a 63-micrometer boundary led two labs to different paleotemperature histories.

How One Drilling Rate Alters Seven Paleotemperature Records

By Jonas Eriksen / May 29, 2026

A study from Lamont-Doherty Earth Observatory shows that rotary coring speed biases foraminiferal Mg/Ca ratios, altering seven North Atlantic paleotemperature records by 0.3–0.8°C.

Belle II Data Cut Flips 8 of 19 Charm Decay Measurements

By Karim Osman / May 29, 2026

A single data cut in Belle II's charm decay analysis flipped the sign of 8 out of 19 measurements, tracing to an overlooked systematic bias in the detector's tracking algorithm.

Crowd-Sourced Replication Finds Three of 12 Rodent Learning Studies Hold

By Karim Osman / May 28, 2026

A large replication project found only 3 of 12 rodent learning studies held up. The $300,000 audit reveals methodological choices that shape the 25% success rate.

One Equipment Lease Cost Shifted Seven Paleoclimate Model Runs

By Renu Shah / May 29, 2026

A $12,000 equipment lease cost forced a University of Washington lab to cancel seven paleoclimate model runs, highlighting how infrastructure expenses shape research outcomes.

Funding Gaps Drive 18 of 40 Psychology Replication Attempts

By Renu Shah / May 28, 2026

A large replication consortium found that only 18 of 40 psychology studies replicated. Funding constraints per lab predicted success rates, highlighting how research economics shapes what we know.

One Catalyst Batch Batch Slowed 13 of 20 Hydrogenation Runs

By Jonas Eriksen / Jun 7, 2026

A single batch of a common palladium catalyst slowed 13 out of 20 hydrogenation runs. Trace iron impurities were the culprit, revealing gaps in standard characterization.

How One Ice Core Dating Choice Shifts Paleoclimate Reconstructions

By Jonas Eriksen / May 28, 2026

A single decision in how we date ice layers can alter our understanding of past climate shifts by centuries. This article examines the methods, uncertainties, and implications for paleoclimate science.