Reproductions are time-consuming and often require detective work. Better documentation and archiving could simplify reproductions. We provide recommendations for authors, journals, and data providers.
Read more about this (open access): doi.org/10.1098/rsos...
Openness and reproducibility are problematic across disciplines. While some disciplines (e.g. political science) fare better than others, openness and reproducibility are far from perfect across the board.
…deviations are usually small and random. Most original and reproduced results are stable in sign and significance, and we find no bias towards exaggerated effect sizes in original research.
…reproducibility problems are widespread. Only one in three authors (35%) share code, and even when code is available, only every second main result (51%) is numerically reproducible.
Do researchers share their code upon request? Does running their orginal code on the original data produce the original results? We provide evidence in a new Royal Society Open Science publication. Studying more than 1,000 articles which use data from the European Social Survey, we find that... 🧵
Billions of regressions ≠ robustness.
My new PNAS Letter shows how unjustified models can drown out justified ones — and why thoughtful model selection matters👇
www.pnas.org/doi/10.1073/...
This is an extremely relevant and interesting study that shows that behavioral intentions measured in survey experiments should not be mistaken for a measure of actual behavior. And it is a nice example of studies that replicate themselves by using different research designs!
If you are interested in why many-analyst studies may overestimate uncertainty, you might want to take a look at our recent open access article: Toward a more credible assessment of the credibility of science by many-analyst studies | PNAS www.pnas.org/doi/10.1073/...
I am happy that my paper with Sabine Düval is now published. We use survey experiments & causal mediation analyses to explore gender ideologies & norms underlying the division of housework. We argue that traditional ideologies may have been overestimated in previous research...
Sociology is behind on ** reproducibility ** but we can get together to catch up!
Join our 3 sessions at the ECSR conference in Barcelona (ecsr2024.com) on what journals can do.
At @sociologicalsci.bsky.social , we require data & code or a good reason.
This year we host in October the META-REP 2024 an international conference focusing on meta-science and replicability in the cognitive, behavioral, and social sciences. The deadline for submissions is April 30. Contribute to vital discussions on meta-science and replicability. #metarep #openscience
Sociology is behind on transparency but we can get together to catch up!
In Europe, the ECSR conference in Barcelona will host two sessions on reproducibility, with special attention to what journals can do:
ecsrnet.eu/other-events/
At Sociological Science, we're now requiring data and code.
Christmas Card with wind turbines (Frohe Weihnachten)
The GLEN team sends Season's Greetings to you and your loved ones! Enjoy the holidays!
We are happy to announce the new DFG-funded longterm German Longitudinal Environmental Study (GLEN).
@glen-study.bsky.social will provide representative, individual level panel data on human-environmental interactions and climate issues.
glen-studie.de
#sciencenews #Klimawandel #climatechange
Ross et al. (Nature, 2022) claim that in most fields women are disadvantaged, but their results show that in only 2 out of 13 fields women are less likely to be granted authorship. For women, it seems that science is not as dismal as reported in the Nature article.
"If your data contradicts the common wisdom: ignore your data." It seems that Ross et al. followed this advice when they concluded in their recent Nature article that women are credited less likely authorship in "most scientific fields". See our comment: osf.io/preprints/so...
Thanks to AMCIS for inviting me and organizing a wonderful workshop. It was great to learn that there are so many interesting projects with survey experiments planned and carried out at the UvA!
Yesterday’s workshop ended with a keynote by @kauspurg.bsky.social on robust design of factorial survey experiments, with proposals for how to incorporate causal inference. We’ve been very fortunate to have Katrin visiting UvA during the last week, giving valuable input to multiple projects.
Our study, which uses a longitudinal field experiment during the "European refugee crisis," comes to null- results, see sociologicalscience.com/articles-v10.... Great that we could publish that!
More immigration, more ethnic discrimination? I am happy that the joint article with Renate Lorenz and Andreas Schneck has now appeared open access in Sociological Science.