The use of uncertainty measures is very common in scientific publications. However, the uncertainty measure or "±" such as SD, SE, CI is not systematically defined. This also affects error bars as their use is very heterogeneous in publications. ➡ This clearly impacts the interpretation of findings and their reproducibility. Several authors do not seem to know the difference between SD and SE and tend to report SE as it is conveniently smaller than SD (unless n=1, which would be ... (fill in the blank)). For inferential purposes, it is more informative to report CIs than SEs. 🔑 Clear universal guidelines in publications must be defined and implemented ensure reproducibility of scientific findings. Feel free to share your experience. For more info, consult this post: https://lnkd.in/eVw3_KDe We have been providing consulting and training services for over 30 years now. https://lnkd.in/eZ_5UFKG, but sadly the situation I described has not improved much. #correctly #reporting #uncertainty #better #science
Absolutely agree. Usually it is SD for a simple reason: the confidence interval would too large!
I have seen a few papers with such ambiguity in notation. They assume we know what they are trying to mean. This is a very weird practice.
PhD, Statistician
1moYou should check SAMPL guidelines for statistical reporting https://www.equator-network.org/2013/02/11/sampl-guidelines-for-statistical-reporting/