This editorial in Nature Human Behaviour, whilst being aimed at publishing researchers, also serves as a basic checklist of what undergraduate students should be taught if using NHST. Adhering to it won’t solve all the issues but it will definitely help future research if good practice is taught early on! https://www.nature.com/articles/s41562-023-01586-w
@mcaleerp We make a point of introducing those and other important concepts (such as pseudoreplication, which is extremely problematic in our field) early in undergraduate years. Since doing this, I've noticed a constant improvement in the ability of undergraduate students to analyse and interpret experimental data, and in overall experimental design ability. The other day I was thrilled to hear a student in the lab correctly suggesting using a mixed effects model.
@nicolaromano Love that! 👏 We had our level 4 (final year) undergraduate conference yesterday and, likewise, to hear them confidently and accurately talking about analyses in ways that seasoned researchers commonly get wrong never fails to bring a smile! Long way to go but these small grassroots changes mount up I reckon!