Uljana Feest (2019), Philosophy of Science, 86(5), 895–905 (December 2019) — DOI-10.1086/705451
Key takeaways
- All experimentation and replication rests on individuation judgments (what counts as the same effect, manipulation, measure), and much of experimental psychology consists in probing those judgments — so direct replication has a place but is less central than the replication-crisis debate assumes.
Seeded from the abstract alone, before the reading in NOTE-196. What follows is what the work says about itself.
Debates about psychology’s replication crisis assume that direct replication is valuable, mostly because it exposes questionable statistical practices. Feest looks more broadly at what replication is in psychological experiments. She argues that every experiment and replication involves individuation judgments, and that experimental psychology often turns on testing whether those judgments are adequate. Conceptual and material questions are therefore ubiquitous in research, and replication, though it has its place, is less central to psychology than is often supposed.
Standing in the record
Filed from the 2026-09-26 sweep of the papers-feed tracker for philosophy: 20 seconds of active reading over 1 session. It was filed Deferred, unread. NOTE-196 is the close reading of 2026-09-26, and it placed the work: Active — worth reading as a short, clearly argued statement of why replication cannot settle “what effect?” questions, built on one worked case (the Mozart effect). It is suggestive for construct-validity arguments about benchmarks, but its strongest lines are asserted rather than argued.
When seeded, before the reading in NOTE-196, the entry gave a priority for a deeper reading: medium — 20 s dwell, and a short, well-cited PSA paper whose thesis is clear from the abstract. The full text was unreachable, and the examples are what would transfer to evaluation practice.
What the seeding asked a deeper reading to check (NOTE-196 is that reading):
- There is a direct analogue in ML evaluation: whether a benchmark result “replicates” depends on individuation judgments about what the task, capability or construct is. That is the construct-validity point behind many evaluation critiques.
- Check the examples Feest uses (likely from memory or perception research), and whether she distinguishes exploratory from confirmatory contexts.
- Check the relation to her later work on exploratory experimentation and to Leonelli (LIT-189) on reproducibility.
Access when seeded: Fetched the Cambridge Core landing page via curl, which gave the abstract, vol. 86(5), pp. 895–905, December 2019. The full text is paywalled there. Cambridge shows “published online 01 January 2022”, which is a platform-migration date. published: uses 2019-07-16 from OpenAlex/Unpaywall, apparently the original University of Chicago Press online date; I could not verify that date at the publisher. The accepted manuscript on PhilSci-Archive (15219, “Feest 2018 Replication manuscript”) is behind a captcha, and PhilArchive is behind Cloudflare. Keywords are mine.