Why Most Discovered True Associations Are Inflated

By John P. A. Ioannidis. Epidemiology, 2008

1,355 citations recorded by the source corpus.

Subfields: Genetics, Statistics and Probability

View Paper: https://www.semanticscholar.org/paper/0373f18622b7b64e99c7e56340ab2011dbd00b5d

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Statistics and Probability

Citations in the corpus

Degrees of Freedom in Planning, Running, Analyzing, and Reporting Psychological Studies: A Checklist to Avoid p-Hacking (2016, background): “Second, their strategic use in research may inflate effect sizes (Ioannidis, 2008; Bakker et al., 2012; Simonsohn et al., 2014; van Aert et al., 2016).”
How Big Is “Big”? Interpreting Effect Sizes in L2 Research (2014, background): “Such experiments would tend to yield larger effect sizes (Ioannidis, 2008; Kline, 2004).”
On the Gap between Adoption and Understanding in NLP (2021, neutral): “3896 ner’s curse” (Ioannidis, 2008) of unreplicable re-”
P-curve: a key to the file-drawer. (2014, neutral): “As a result, published evidence is unrepresentative of reality (Ioannidis, 2008; Pashler & Harris, 2012).”
The Song Is You (2016, methodology): “…et al., 2008; Dunn et al., 2011; George et al., 2007; Langmeyer, et al., 2012; Ren-tfrow & Gosling, 2003; Zweigenhaft, 2008), and considering the comparatively large sample size used, the correlation estimates we observed should be quite stable (Ioannidis, 2008; Scho¨nbrodt & Perugini, 2013).”
Using the Student's "t"-Test with Extremely Small Sample Sizes. (2013, neutral): “From a Bayesian perspective, small sample sizes may still be problematic and may contribute to false positives and inflated effect sizes (Ingre, in press; Ioannidis, 2005, 2008).”
The statistical significance filter leads to overoptimistic expectations of replicability (2018, background): “We want to stress that none of the statistical points made in this paper are new (for similar arguments, see Button et al., 2013; Dumas-Mallet, Button, Boraud, Gonon, & Munafò, 2017; Frank et al., 2017; Goodman, 1992; Hedges, 1984; Ioannidis, 2008, among others).”
Does the Flipped Classroom Improve Student Learning and Satisfaction? A Systematic Review and Meta-Analysis (2019, background): “Furthermore, there is a risk that small, underpowered studies overestimate any effects reported, partly because statistical tests in such studies will reach significance only if the sample mean difference happens to be very large (Button et al., 2013; Ioannidis, 2008; Sterne, Gavaghan, & Egger, 2000”
Personality in 100,000 Words: A large-scale analysis of personality and word use among bloggers (2010, background): “These seemingly paradoxical results are easily reconciled if one supposes that the true effects under investigation were actually relatively modest in both studies, but that Hirsh and Peterson's (2009) results, which stemmed from a much smaller sample (N = 94), were more susceptible to sampling erro”
p-Curve and Effect Size (2014, background): “First, it shows the dramatic inflation of effect size generated by publication bias (Hedges, 1984; Ioannidis, 2008; Lane & Dunlap, 1978).”