Thanks, Elise! Happy to hear you learned a lot!

No, I've missed it! Looks great. Thanks for sharing. I'll have a closer look at it later this week.

Please do so! The FORTT materials were truly inspiring for creating this course, so always happy to support others as well.

I’ve uploaded everything to @cos.io, and you can also reach me at amelie.godefroidt@kuleuven.be if needed!

osf.io/s8mxd/files/...

Btw: after a well-deserved long weekend, I’ll be sharing all course materials from the Open Science in the Social Sciences summer school @cos.io !

🎉 That’s a wrap!
Four days of deep dives into #OpenScience, #preregistration, #replication, #reproducibility, #multiverse, & more—applied to the social sciences.

Huge thanks to all participants for the energy, insights & thoughtful questions!
@ugentopenscience.bsky.social @gasparugent.bsky.social

🗣️ Fourth and final stop: Open Peer Review

We debated:
✍️ Signed reviews
🧾 Public reports
💬 Community commenting
✅ Platforms like @pubpeer.com @F1000Research & @publons

#OpenPeerReview

🔐 Third stop: how open can your data be under GDPR?

Enter:
✅ FAIR principles
✅ Personal data, special categories of data
✅ Anonymization vs. Pseudonymization
✅ Consent, licensing, & DPO support

More info: www.kuleuven.be/rdm/en/guida...

#FAIRdata #DataSharing #OpenData

💻 Second stop: #rMarkdown

📝 YAML -> Code chunks -> text -> Knit!
🔁 We did various exercises to knit dynamic, reproducible reports (HTML, pdf, word)
🛠️ Bonus: it’s the perfect gateway to @quarto.org

🔗 rmarkdown.rstudio.com/authoring_qu...

#rStats

🌌 First stop: Multiverse Analysis

We explored how analytic decisions shape results—and why reporting just one is like publishing from a parallel universe.

📚 Steegen et al. (2016): doi.org/10.1177/1745...
📚 Simonsohn et al. (2020): www.nature.com/articles/s41...
🔗 specr: masurp.github.io/specr/

🔥 Big day ahead at the #OpenScience summer school!

On the menu:
💻 Making your research reproducible with @rmarkdown.bsky.social & @quarto.org
🔐 Balancing Open Science with Data Protection
🧐 What’s the deal with Open Peer Review?
🌐 A quick dive into Multiverse Analysis

We ended with a #pHackathon 🎯

Students were challenged to “find significance” in a fake dataset using researcher degrees of freedom 😅 shinyapps.org/apps/p-hacker/

Key lesson: flexibility isn't bad—but without transparency, it’s dangerous.

2️⃣ Smart data sharing: as open as possible, as closed as necessary!

3️⃣ Reproducible reporting with @rmarkdown.bsky.social n & @quarto.org

4️⃣ Built better replication packages with solid README files, transparent code, and full workflows. social-science-data-editors.github.io/template_REA...

1️⃣ Consistency checker:

Want cleaner science? Use these tools:
✔️ GRIM test → check if reported means are possible
✔️ shinyapps.org/apps/p-check... → spot wrong p-values
✔️ michelenuijten.shinyapps.io/statcheck-web/ → scan entire papers for wrong p-values

🧰 Beyond (a priori!) power analysis, students got 5 more tools for their repro toolbox:

1️⃣ Consistency checkers (GRIM, p-checker, statcheck)
2️⃣ Smart data sharing (more tomorrow!)
3️⃣ Reproducible reporting (more tomorrow!)
4️⃣ Crafting solid replication packages
5️⃣ #ManyLabs5 on quality > prereg

We’d tackled HARKing & p-hacking—but what about statistical power? 🧠

We broke down how sample size, effect size & alpha shape our ability to detect real effects.

Median power in poli sci = 10% 😬 (Arel-Bundock et al. 2024). To solve this, we introduced:

🧮 G*Power and 📦 pwr in R (recommended)

First up: two terms that often get mixed up 👇

🔁 Reproducibility = same data, same code
📦 Replicability = new data, same design

We unpacked these (and more) with help from Brodeur et al. (2024):
📖 www.iza.org/publications...

🔥 Pro tip: join a #Replication Game via @i4replication.bsky.social