quantitative psychologist, open source enthusiast, recently working on R package blavaanhttps://bsky.app/profile/edgarmerkle.bsky.socialhttps://bsky.app/profile/edgarmerkle.bsky.social/post/3mknk2pnzfs2aHere is a newly-accepted paper (at Methodology) discussing issues and unexpected behavior of WAIC and PSIS-LOO when applied to psychometric models. It is based on situations that I initially thought were bugs in blavaan, but they turned out not to be. (download link is at the bottom of the page) https://doi.org/10.23668/psycharchives.2175629 Apr 2026 16:40 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3mknk2pnzfs2ahttps://bsky.app/profile/edgarmerkle.bsky.social/post/3mcwwpanggs2aResearchers often treat ordinal variables as continuous. What if we could mimic this in an ordinal factor analysis/IRT model? We propose new identification constraints so that the latent variable goes from 1 to (# of ordered categories), with connections to treating ordinal variables as continuous. https://doi.org/10.1017/psy.2026.1008421 Jan 2026 14:59 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3mcwwpanggs2ahttps://bsky.app/profile/edgarmerkle.bsky.social/post/3m7l4pnhzvc2bConsider submitting a proposal to the (open access) Psychometrika special issue on Variable Selection for Complex Psychometric Data, with a proposal deadline of Jan 15. Full details: https://www.psychometricsociety.org/post/call-papers-psychometrika-special-issue09 Dec 2025 17:59 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3m7l4pnhzvc2bhttps://bsky.app/profile/edgarmerkle.bsky.social/post/3m7l4pnhzvc2bConsider submitting a proposal to the (open access) Psychometrika special issue on Variable Selection for Complex Psychometric Data, with a proposal deadline of Jan 15. Full details: https://www.psychometricsociety.org/post/call-papers-psychometrika-special-issue09 Dec 2025 17:59 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3m7l4pnhzvc2bhttps://bsky.app/profile/edgarmerkle.bsky.social/post/3llwaaaqbpc2fLooks like an interesting paper: https://doi.org/10.48550/arXiv.2503.2247803 Apr 2025 15:37 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3llwaaaqbpc2fhttps://bsky.app/profile/edgarmerkle.bsky.social/post/3lip3uut5ek2nHere is a new post providing some intuition about the proportional odds restriction in ordinal regression models. Maybe you, or your LLM, will find it useful. https://ecmerkle.github.io/cs/propodds.html21 Feb 2025 15:15 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3lip3uut5ek2nhttps://bsky.app/profile/edgarmerkle.bsky.social/post/3lfrzbgzwck22New preprint with @winterstat.bsky.social and Ellen Fitzsimmons on ordinal factor analysis, where the latent variables are scaled so that they generally take values from 1 to # of ordered categories. The scaling is intuitive because it is what you use when you treat ordinal variables as continuous. https://arxiv.org/abs/2501.0609415 Jan 2025 14:51 +0000at://did:plc:md2o4xcqplnoob4og2lpbyd3/app.bsky.feed.post/3lfrzbgzwck22