Metascience Matters

Randy Ellis
  • 0.0 (0)
  • SCIENCE
  • UPDATED MONTHLY

I'm Randy Ellis, a computational biologist and neuroscientist who cares about metascience, reproducibility, and rigor in science. I started Metascience Matters because I believe science communication can be improved across the board for all fields of science, but most especially for the field of science dedicated to studying science itself.

Episodes

  1. APR 1

    Tim Errington: Replicating Cancer Biology, Academic Incentives, and Clinical Translation

    Tim Errington is the Senior Director of Research at the Center for Open Science. He led the Reproducibility Project: Cancer Biology, Systematizing Confidence in Open Research and Evidence (SCORE), as well as the implementation and evaluation of initiatives such as Registered Reports, Registered Revisions, responsible conduct of research trainings, and open science badges. CONTACT RANDY: Feedback: metasciencematters@gmail.com EPISODE LINKS: Investigating the replicability of preclinical cancer biology (Reproducibility Project: Cancer Biology led by Tim Errington): https://elifesciences.org/articles/71601 Bayer replication study: https://www.nature.com/articles/nrd3439-c1 Amgen replication study: https://www.nature.com/articles/483531a Reproducibility in Cancer Biology: Challenges for assessing replicability in preclinical cancer biology (Companion paper to Reproducibility Project: Cancer Biology): https://elifesciences.org/articles/67995 What is replication?: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3000691 Study comparing standard reports and registered reports in psychology: https://journals.sagepub.com/doi/10.1177/25152459211007467 Blog post on the seemingly magical success of revision experiments: https://rajlaboratory.blogspot.com/2013/06/the-magical-results-of-reviewer.html Google's AI co-scientist paper: https://arxiv.org/abs/2502.18864 Machine-readable documents: https://journals.sagepub.com/doi/10.1177/2515245920970949 How open science helps researchers succeed: https://elifesciences.org/articles/16800 ZBW's Expedition to Open Science Land: https://expedition-open-science.org/ OUTLINE: 0:00 - Introduction 4:47 - Tim's origin story as a cancer biologist 6:38 - Initial interest in metascience 9:24 - Starting the Reproducibility Project: Cancer Biology 12:07 - How were the studies that were replicated chosen? 14:41 - Publishing the Reproducibility Project: Cancer Biology as a registered report 17:26 - Results from the Reproducibility Project: Cancer Biology 20:28 - Tim's experience throughout the years running the Reproducibility Project: Cancer Biology 25:21 - The difficulty of running cancer biology studies 27:54 - Judging whether a replication is successful 31:23 - What has the response to the Reproducibility Project: Cancer Biology been? 37:52 - Why aren't replication rates higher? 40:26 - Challenges of running cancer biology replication studies 45:43 - Caveats of preclinical disease models 49:13 - The incentive for positive data in science 57:05 - Systemic intervention vs. Individual policing 1:01:04 - The value of registered reports 1:07:38 - Registered revisions 1:10:48 - Falsifying theories early at the preclinical stage 1:15:21 - Different institutions (e.g., academic, industry) conducting different studies (e.g., preclinical, clinical) 1:17:34 - New initiatives at the Center for Open Science (Replication project of social/behavioral sciences, automated tools for predicting replication success, LifeCycle journal) 1:23:02 - AI scientists are trained on biased literature; distrust of academic literature in drug discovery 1:28:46 - Peer review 1:32:51 - Narrative in science 1:35:06 - 100-200 years into the future 1:40:29 - Advice for high school/undergraduate listeners 1:42:51 - Metascience manifests in every field 1:44:46 - Philosophy of science 1:47:49 - Outro

    1h 48m

  2. APR 1

    Brian Nosek: Replicating Psychology, Founding the Center for Open Science

    Brian Nosek is the cofounder and Executive Director of the Center for Open Science. He co-developed the Implicit Association Test, a method that advanced the study of implicit bias. He then co-founded three non-profit organizations: Project Implicit to advance research about implicit bias, the Society for the Improvement of Psychological Science to improve the research culture in his home discipline, and the Center for Open Science (COS) to improve rigor, transparency, integrity, and reproducibility across research disciplines. He led the Reproducibility Project in Psychology, a replication of 100 studies from psychology, as well as multiple other replication projects, along with policy reforms such as open science badges, and the investigation of prediction markets for study replication. CONTACT RANDY: Feedback: metasciencematters@gmail.com EPISODE LINKS: Estimating the reproducibility of psychological science (Reproducibility Project in Psychology led by Brian Nosek): https://www.science.org/doi/10.1126/science.aac4716 False-Positive Psychology: Undisclosed Flexibility in Data Collection and Analysis Allows Presenting Anything as Significant: https://journals.sagepub.com/doi/10.1177/0956797611417632 What is replication?: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3000691 Metascience conference: https://metascience.info/ International Conference on the Science of Science and Innovation: https://icssi.org/ International Congress on Peer Review and Scientific Publication: https://peerreviewcongress.org/ MetaROR: https://metaror.org/ OUTLINE: 0:00 - Introduction 2:56 - Brian's origin story as a psychologist 6:35 - Initial interest in metascience  10:57 - "Scientific Utopia," starting the Reproducibility Project in Psychology, and major fraud cases 17:57 - How did the Reproducibility Project in Psychology come together? 23:40 - Results of the Reproducibility Project in Psychology 27:03 - Response from the field, institutional changes 30:39 - Advice for high school/undergrad listeners 34:03 - Outro

    34 min

  3. MAR 18

    Eugenie Reich: Litigating multimillion-dollar scientific fraud cases

    Eugenie Reich is an attorney committed to taking on scientific fraud, understanding the incentives that drive it, and recovering misdirected research funding. She is also a former investigative science journalist committed to correcting the scientific record. Her 2009 book Plastic Fantastic, details a major fraud cause in physics at Bell Labs. Two of the cases she has litigated cases were against Biogen ($900 million settlement), and the Dana-Farber Cancer Institute ($15 million settlement). CONTACT RANDY: metasciencematters@gmail.com EPISODE LINKS: Cargo Cult Science: https://calteches.library.caltech.edu... The PubPeer Conundrum: https://www.tandfonline.com/doi/epdf/... New Scientist: https://www.newscientist.com/ Skeptical Inquirer: https://skepticalinquirer.org/ Special thanks to Dylan Bouscher OUTLINE: 0:00 - Introduction 4:20 - How Eugenie became interested in science 6:25 - Interest in scientific fraud 7:38 - Deciding to become a science journalist 8:58 - Fraud in physics at Bell Labs (the subject of her 2009 book Plastic Fantastic) 20:15 - Bell Labs coming under private ownership 22:53 - Interviewing scientists for her book 26:50 - The response from the physics community 31:10 - Deciding to become a lawyer 33:28 - The False Claims Act 36:29 - The qui tam provision of the False Claims Act 37:46 - The $900M case against Biogen 44:02 - The day-to-day of working on this case 46:30 - Impact of the case on biotech and pharma 48:40 - The $15M case against Dana-Farber 50:42 - Do universities have an incentive to protect accused researchers? 54:15 - Is the scale of fraud too large to be fixed? 57:56 - Does fraud damage public trust in science? 1:00:44 - Tools/solutions to combat these problems 1:04:32 - Advice to practicing scientists 1:05:55 - Advice and resources for listeners 1:08:28 - Outro

    1h 9m

  4. MAR 4

    James Heathers: Forensic Metascience, the GRIM test, and technology for checking papers

    James Heathers is the Founder and Director of the Medical Evidence Project, a venture of The Center for Scientific Integrity. He aims to reduce medical harm and improve patient outcomes by identifying and publicizing errors and miscondcut in the medical literature. He uses forensic meta-analytical techniques to detect and deconstruct errors arising from low-quality science and fraudulent work in areas that involve large numbers of patients. CONTACT RANDY: metasciencematters@gmail.com EPISODE LINKS: The original GRIM test paper: http://www1.psych.purdue.edu/~gfranci... Machine-readable documents: https://journals.sagepub.com/doi/10.1... Chaos in the Brickyard: https://www.science.org/doi/10.1126/s... Books: The Man Who Only Loved Numbers: https://en.wikipedia.org/wiki/The_Man... The Emperor of All Maladies: https://en.wikipedia.org/wiki/The_Emp... OUTLINE: 0:00 - Introduction 5:27 - How James has been the last couple days, and his radio voice 6:40 - Writing the first book on Forensic Metascience 10:47 - James' PhD work on heart rate variability 14:12 - Starting to work on metascience 18:40 - The GRIM test 26:40 - Programmatically scaling statistical checks 33:25 - Centering quantitative results and peripheralizing narrative in scientific papers 39:36 - AI for metascience, unearthing data underneath narrative 48:11 - Pursuing cases of misconduct 52:45 - Founding and Directing the Medical Evidence Project 56:10 - Incentives for positive data, post-publication review, challenging the binary of positive and negative data 1:05:41 - Advice and resources for listeners 1:12:40 - Optimism over pessimism 1:16:58 - Outro

    1h 17m

  5. FEB 8

    Mu Yang: 300+ retractions, image manipulation, and why science should be boring

    Mu Yang is a behavioral neuroscientist at Columbia University, and a scientific sleuth responsible for more than 300 retractions. She led an effort that discovered more than 130 fraudulent papers in the publication record of Eliezer Masliah, former head of the Division of Neuroscience at the National Institute on Aging at the National Institutes of Health. Her sleuthing work has been documented in the book "Doctored" by Charles Piller, Science Magazine, and other outlets, and is unpaid. CONTACT RANDY: metasciencematters@gmail.com EPISODE LINKS: Books:  Doctored by Charles Piller: https://www.simonandschuster.com/books/Doctored/Charles-Piller/9781668031254 Unreliable by Csaba Szabo: https://cup.columbia.edu/book/unreliable/9780231216241/ Scientific integrity blogs:  Dorothy Bishop: https://deevybee.blogspot.com/ Leonid Schneider: https://forbetterscience.com/ Podcasts for critical thinking Plain English by Derek Thompson: https://www.theringer.com/podcasts/plain-english-with-derek-thompson The Gray Area by Sean Illing: https://www.vox.com/the-gray-area The Ezra Klein show (NYT): https://www.nytimes.com/column/ezra-klein-podcast OUTLINE: 0:00 - Introduction 2:58 - Mu's origin story 4:35 - Moving to Columbia 6:15 - How Mu became a sleuth 8:13 - Reporting her first case 13:09 - Red flags Mu looks for in papers 17:30 - Reductionism in behavioral neuroscience 18:04 - Standardization vs. Generalizability 19:58 - Data sharing standards across fields 21:09 - Difficulties of reporting irregularities in papers, university incentives 23:54 - Allocating time between images, numerical, other kinds of data 24:37 - How she searches through papers 25:45 - Examining the chemistry literature 31:10 - Types of misconduct vary by field, risks of reporting 35:43 - The case of Eliezer Masliah 40:31 - Why demonizing individual scientists isn't productive; the system isn't working 56:59 - Academic incentives for positive data 1:07:31 - Hard to publish null data; "unhealthy codependence" between academia and publishing 1:13:08 - Changing incentives 1:21:42 - Are we even making a dent in the scale of scientific misconduct?  1:27:35 - Mu's toolkit 1:29:38 - Mu does this work because it's fun! 1:34:38 - Protecting students; telling them that null data is ok 1:37:52 - Evaluating researchers  1:43:15 - Is peer review still relevant? 1:51:38 - How much better could science be? 1:55:14 - What will science look like in a century? 1:58:13 - Advice and resources for listeners 2:00:54 - Online presence 2:01:35 - Outro

    2h 2m

  6. JAN 23

    Florian Naudet: Esketamine for depression, registered reports, and alcohol use disorder

    Florian Naudet is a Professor of Therapeutics at Rennes University. As a metascientist and psychiatrist, his research interests lie in developing and evaluating methodological solutions to treatment assessment, primarily but not exclusively for mental health conditions. His work has also made inroads to quantifying and understanding research waste and the prevalence of substandard data-sharing practices. CONTACT RANDY: Feedback: metasciencematters@gmail.com EPISODE LINKS: Efficacy and safety of esketamine for “treatment resistant depression”: registered report for a systematic review with an individual patient data meta-analysis of randomized, double-blind, placebo-controlled trials https://link.springer.com/article/10.... Vibration of effects from diverse inclusion/exclusion criteria and analytical choices: 9216 different ways to perform an indirect comparison meta-analysis https://link.springer.com/article/10.... Pharmageddon (book) https://davidhealy.org/pharmageddon-i... A manifesto for reproducible science https://www.nature.com/articles/s4156... Using reporting guidelines to improve the reproducibility of cooking Christmas tree meringues: the “People tasting trees” cluster-randomised controlled trial https://pmc.ncbi.nlm.nih.gov/articles... OUTLINE: 0:00 - Introduction 0:36 - Florian's origin story as a psychiatrist and metascientist 4:46 - How Florian became interested in metascience 11:55 - Systematic review of clinical trials of Esketamine for depression 28:45 - Publishing this study as a registered report 40:49 - Vibration-of-effects analysis of double-blind randomized controlled trials assessing nalmefene and/or naltrexone for treating alcohol use disorders 59:58 - Advice for listeners interested in pursuing research like Florian's 1:03:38 - Outro

    1h 4m

  7. 12/15/2025

    Chirag Patel: Exposomics, Vibration-of-Effects, and the Future of AI in Health

    Chirag Patel is an Associate Professor of Biomedical Informatics at Harvard Medical School, renowned for his expertise in using computational methods to understand human health and disease from high-throughput data streams. He specializes in understanding the role in the intersection of genetics and environmental exposures (the exposome) in human health, as well as various disease areas such as diabetes, cardiovascular disease, and Alzheimer’s disease. CONTACT RANDY: Feedback: metasciencematters@gmail.com EPISODE LINKS: Chirag’s Lab: https://www.chiragjpgroup.org/ TEDx talk on the exposome:    • Exposome: decoding human health and diseas...   OUTLINE: 0:00 - Introduction 0:37 - How do you decide what to work on? 1:57 - Where does metascience fit into your work? 3:57 - Vibration-of-effects analysis 6:41 - How does VoE change how we see existing scientific work? 9:22 - The biggest challenges in the replication crisis within biomedical informatics 11:48 - Environment-wide association study of Type 2 diabetes 15:04 - The exposome 17:30 - Policy vs. precision medicine applications of the exposome 20:09 - The patient exposome 22:38 - HBA1C for diabetes as a bridge to the exposome 23:54 - Broader metascientific issues of the exposome 25:01 - The effects of extreme weather events on human health 29:35 - AI for biomedical informatics, the exposome, metascience 31:19 - Advice for listeners interested in pursuing research like Chirag’s 32:53 - Outro

    33 min
  • 7 Episodes

About

I'm Randy Ellis, a computational biologist and neuroscientist who cares about metascience, reproducibility, and rigor in science. I started Metascience Matters because I believe science communication can be improved across the board for all fields of science, but most especially for the field of science dedicated to studying science itself.

Information

  • Creator
    Randy Ellis
  • Years Active
    2025 - 2026
  • Episodes
    7
  • Rating
    Explicit
  • Copyright
    © Randy Ellis
  • Show Website
    Metascience Matters