Effect Size vs. Statistical Significance

Here’s a cheery one for our first episode of the year. Guess what happens when you give several sets of scientists the same dataset and ask them to answer the same question? Well, they all find the same results, right? Right!?

Sadly not. This “Many Analysts” problem has been analysed and debated in multiple different scientific fields and across several papers. We cover them in this episode. What does it tell us about the objectivity of science if different teams draw different conclusions from the exact same data?

The Science Fictions podcast is brought to you by Works in Progress magazine. Their excellent new article on how we’re living in “the golden age of vaccine development”, as discussed on the show, can be found (along with the rest of their articles on science, history, and technology), at worksinprogress.co. We’re very grateful that they support the podcast.

Show notes

* 2015 Nature commentary article on “crowdsourced research” (on racism in football)

* And the full 2018 writeup titled “Many Analysts, One Data Set”

* Gelman and Loken on the “Garden of Forking Paths”

* 2020 many-analysts neuroscience (fMRI) paper

* And the plan for the similar study on EEG

* 2022 PNAS many-analysts paper on the “hidden universe of uncertainty”

* 2026 critique on ideological bias from George Borjas

* 2023 critique on effect sizes vs. statistical significance

* 2025 ecology & evolution many-analysts paper on blue tits and eucalyptus

* 2025 economics many-analysts paper with results on data cleaning

* 2024 PNAS critique of many-analysts research

* Julia Rohrer’s critique of multiverse analysis

Credits

The Science Fictions podcast is produced by Julian Mayers at Yada Yada Productions.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit sciencefictionspod.substack.com/subscribe