Science feels like the most reliable thing we have. The opposite of
belief. But it’s a belief system itself—a ritual, with all the failure
modes that rituals have. And the receipts are right there in the
Further reading
TheScientific Ritual — the article this lecture is based on
Problems withp-values — the technical companion: Fisher, Neyman-Pearson, the
hybrid mess
The trapof scientific evidence — on the “no evidence” tension and the
homeopathy/parachute paradox
Everythingis ideology — science as one belief system among several
In praise of thesage — other ways of knowing; the MD/PhD distinction
Scientificfact — on what science actually does
The value ofritual — ritual as a knowledge-production strategy
Meditation — on thedinner-table meditation example
BeyondSystem 1 and System 2 — on Kahneman’s dual-process framework
The placeboeffect — on why “works for some, not for others” is a feature, not a
bug
Grit —positive-psychology critique
Overengineeringcalming down (lecture) — the broader positive-psychology audit
Bias is good(lecture) — the cognitive-bias series
Life isworse (lecture) — the previous episode; a worked example of reading
a literature
References
The replication crisis
itself
Open Science Collaboration (2015), Estimatingthe reproducibility of psychological science, Science 349
(6251)
Wikipedia: replicationcrisis
American Statistical Association: Wasserstein,Schirm & Lazar (2019), Moving to a World Beyond “p <
0.05”
Statistical ritualism
Gerd Gigerenzer (2018), StatisticalRituals: The Replication Delusion and How We Got There, Advances
in Methods and Practices in Psychological Science
Philip B. Stark & Andrea Saltelli (2018), Cargo-cultstatistics and scientific crisis, Significance 15 (4)
Andrew Gelman & Eric Loken (2014), TheStatistical Crisis in Science — the “garden of forking paths”
paper
Andrew Gelman, WhyI don’t like so-called Bayesian hypothesis testing
p-values, Bayes factors,
and software
Wikipedia: p-value, Bayes factorRonald A. Fisher (1925), Statistical Methods for ResearchWorkers — where the 5% threshold appears as an illustrative
example
Harold Jeffreys (1939), Theory of Probability — where theBayes-factor thresholds (BF > 3 substantial, BF > 10 strong) come
from
JASP — the open-sourceBayesian statistics software with default priors
Specific
replication-crisis casualties
Cuddy, Wilmuth & Carney (2010) original power posingpaper; Carney’s later statement
withdrawing support
Hagger et al. (2016), AMultilab Preregistered Replication of the Ego-Depletion Effect
Bargh, Chen & Burrows (1996) original elderlypriming paper; failed Doyen
et al. (2012) replication
Brown, Sokal & Friedman (2013), The ComplexDynamics of Wishful Thinking — demolishing the 3:1
positivity ratio
Carol Dweck, growthmindset — replication concerns documented in Sisk et al. (2018) and
Bahník & Vranka (2017)
Angela Duckworth, grit —meta-analytic critique in Credé, Tynan &
Harms (2017)
Books cited in the lecture
Daniel Kahneman, Thinking, Fast and SlowStephen J. Gould, Adam’sNavel and Other Essays
Yann Martel, Life ofPi
Bill Mollison, Permaculture: A Designer’s ManualOther
RichardDawkins on militant atheism (TED) — the “evidence vs. faith”
framing
Reform efforts: preregistration, open data, multi-lab replication consortia(e.g. ManyLabs)