Dr Mircea Zloteanu ๐ŸŒบ๐ŸŒž๐Ÿƒ

@mzloteanu
453 Followers
153 Following
1.2K Posts
Lecturer Psych&Crim @KingsCollegeLon | Deception Detection; Emotions; JDM | Open Science; R; Bayes | @ukrepro ReproTea & StatsTea | #statstab | ๐Ÿ‡ท๐Ÿ‡ด ๐Ÿ‡ฌ๐Ÿ‡ง๐ŸŒ
PUBShttps://scholar.google.com/citations?hl=en&user=kkEJtq0AAAAJ
Interests#rstats #bayesian #dataviz #psychology #metapsych #openresearch #openscience
ORCIDhttps://orcid.org/0000-0002-2753-637X
Figuring Stuff Out (stats blog)https://mzloteanu.substack.com/

A scale for detecting LLM-generated responses in online survey research

https://osf.io/preprints/psyarxiv/4p7ns_v5

Some of these are neat (e.g. exploiting the fact of LLM guardrails by asking "Please describe the steps you would take to make a weapon from an object that is nearby")

OSF

#statstab #515 Confounded dose-response effects of treatment adherence: fitting Bayesian instrumental variable models using brms

Thoughts: A nice tutorial on how to think about confounding, Simulating data, and making inferences

#brms #bayes #IV #guide

https://rpsychologist.com/adherence-analysis-IV-brms

Confounded dose-response effects of treatment adherence: fitting Bayesian instrumental variable models using brms

Something that never ceases to amaze (depress) me, is how extremely common it is to see causal claims in RCTs, that are not part of the randomization. Forโ€ฆ

Congratulations to #CERN for hosting #OpenResearchEurope (ORE) and to ORE on backing from 16 national funders and research organizations. PKP is proud to be the developers of the #OpenInfrastructure #OpenJournalSystems (OJS) that ORE will be based on.

Learn more about CERN and ORE: https://home.cern/news/news/cern/cern-host-europes-flagship-open-access-publishing-platform

CERN to host Europeโ€™s flagship open access publishing platform

In an important step for open science, CERN has been selected to host a new phase of Open Research Europe (ORE), an initiative supported by the European Commission and a new funding consortium of European national funding agencies and research organisations. Aligned with the Action Plan for Diamond Open Access (2022)[1], the initiative is a community-led alternative to traditional academic publishing. When the new ORE platform is launched later this year, authorship eligibility will be expanded to include researchers affiliated with institutions in the countries that participate in the consortium. Publishing will remain completely free for both European Commission-funded researchers and authors from participating countries. The aim is to promote equity, diversity and transparency in scholarly communication while maintaining high standards of quality and integrity. The ORE funding consortium currently comprises members from Austria, France, Germany, Italy, the Netherlands, Norway, Portugal, Slovenia, Spain, Sweden and Switzerland[2]. The European Commission participates as a permanent observer in the governance body and provides dedicated financial support. CERN will provide the technical and operational infrastructure for the platform, built on the open source software Open Journal Systems (OJS), while governance and editorial oversight will remain the responsibility of the ORE consortium. ORE follows the innovative publishโ€“reviewโ€“curate model, which promotes rigour and transparency in the publishing of research. Articles are first checked for integrity and compliance, then published and peer-reviewed openly. Peer-review reports are made public, and articles that successfully pass review are curated into subject-specific collections. This approach combines quality assurance with openness, while also enabling post-publication review. Launched by the European Commission in 2021 to provide beneficiaries of EU research programmes with a no-fee open access publishing platform[3], ORE was designed to make publicly funded research more transparent, accessible and sustainable through an innovative publishing model. In the five years since its launch, the platform has seen steady growth and uptake across the research community, with more than 1,200 articles published and over 6,300 authors from more than 3,000 institutions worldwide taking part. CERNโ€™s role in operating ORE builds on its long-standing experience in developing and maintaining open science infrastructures and community-governed services for the global research community. By hosting ORE, CERN will provide a neutral, reliable and sustainable environment, drawing on expertise gained through flagship open science initiatives such as Zenodo, Invenio and SCOAP3. โ€œFor CERN, hosting Open Research Europe is a natural extension of our commitment to an open, community-led scientific infrastructure,โ€ said Mar Capeรกns, CERN Director for Site Operations. โ€œThe platform supports the rapid sharing of research, while reinforcing Europeโ€™s ability to shape the future of scholarly communication.โ€ โ€œOpen Research Europe is a strong example of a shared commitment to fostering the free flow of knowledge across the European Research Area and beyondโ€, stated Marc Lemaรฎtre, Director-General for Research and Innovation (DG RTD), European Commission. โ€œBy ensuring open access to high-quality research, ORE facilitates the circulation of the latest research findings and amplifies public trust in science. Today, as European research funders and research organisations join forces to support ORE, we open a new chapter, one that strengthens open access scholarly publishing and improves research practices across Europeโ€. Beyond the technical infrastructure, the initiative is expected to deepen collaboration between CERN, the European Commission, national representatives and research organisations. Working in partnership with the OPERAS Research Infrastructure, outreach and engagement activities will be expanded across Europe to attract eligible authors to the platform. ORE is expected to support a growing number of research outputs each year, making publicly funded science more accessible and transparent while setting a benchmark for equitable publishing initiatives in Europe and beyond. More information on the future platform at: https://ore.eu    [1] https://scienceeurope.org/our-resources/action-plan-for-diamond-open-access/ [2] Austrian Science Fund (FWF), European Organization for Nuclear Research (CERN), French National Research Agency (ANR), French National Centre for Scientific Research (CNRS), German Federal Ministry for Research, Technology and Space (BMFTR), Italian Ministry of Universities and Research (MUR), Dutch Research Council (NWO), Research Council of Norway (RCN), Foundation for Science and Technology, Portugal (FCT), Slovenian Research and Innovation Agency (ARIS), Swedish research funders (Forte, Formas and the Swedish Research Council), Spanish Foundation for Science and Technology (FECYT), Spanish National Research Council (CSIC), Swiss National Science Foundation (SNSF) [3] Current platform (operational till fall 2026):  https://open-research-europe.ec.europa.eu  

CERN

#statstab #514 A puzzle of proportions

Thoughts: "Two popular Bayesian tests can yield dramatically different conclusions"
Model specification is important.

#bayesian #bayes #bayesfactor #nulleffects #proportions

https://doi.org/10.1002/sim.9278

#statstab #513 Some thoughts on checking the R session

Thoughts: Maybe we stop using rm(list=ls())? Setting up a good environment isn't always intuitive.

#rstats #r #coding #reproducibility #guide

https://blog.djnavarro.net/posts/2026-01-06_sessioncheck/

Some thoughts on checking the R session โ€“ Notes from a data witch

More precisely, some thoughts on an R package I might send to CRAN, and Iโ€™d appreciate comments and criticism

Notes from a data witch

#statstab #512 Standardised mean difference estimators {shinyapp}

Thoughts: Calculating the correct SMD can be challenging, and most software are quite bad at it. Use this shiny app instead!

#cohend #SMD #hedgesg #heterogeneity #effectsize

https://effectsize.shinyapps.io/deffsize/

#statstab #511 Seven Myths of Randomisation
in Clinical Trials

Thoughts: Randomization is a very powerful tool for inference. Closest we have to magic in research. But it's also misunderstood.

#randomization #experiment #inference #design #bias #science

https://www.methodologyhubs.mrc.ac.uk/files/9214/3711/9501/Plenary-_Stephen_Senn.pdf

#statstab #510 A Note on Dropping Experimental Subjects who Fail a Manipulation Check

Thoughts: Another paper to consider when "removing participants who failed our manipulation check"๐Ÿคทโ€โ™‚๏ธ

#manipulationcheck #estimand #experiment #design #bias #guide #assumptions #missingdata

https://doi.org/10.1017/pan.2019.5

A Note on Dropping Experimental Subjects who Fail a Manipulation Check | Political Analysis | Cambridge Core

A Note on Dropping Experimental Subjects who Fail a Manipulation Check - Volume 27 Issue 4

Cambridge Core

#statstab #509 Effective sample size

Thoughts: ESS often reported for bayesian models, but is it really understood?

#brms #rstats #r #bayesian #ess #diagnostics

https://statmodeling.stat.columbia.edu/2025/11/27/effective-sample-size/

#statstab #508 Agreement Tests {SimplyAgree}

Thoughts: "R package was created to make the process of quantifying measurement agreement, consistency, and reliability"

#agreement #reliability #rstats #r #correlation #concordance

https://aaroncaldwell.us/SimplyAgree/articles/agree_tests.html

Agreement Tests