Short answer
NIH expects institutional training grants, institutional career awards and individual fellowships to describe formal instruction in rigorous experimental design, covering well-controlled experiments, relevant biological variables such as sex, authentication of key resources, appropriate statistics and transparent reporting. The strongest programs pair a course with recurring lab-based practice.
What does NIH expect?
NIH expects training programs and fellowship applicants to describe formal instruction in scientific rigor and transparency: designing well-controlled experiments that consider relevant biological variables, using authenticated resources, applying appropriate statistics and reporting methods fully enough for others to reproduce the work. NIH announced this in December 2015 (NOT-OD-16-034) for institutional training grants (including T32, T34 and T35), institutional career development awards (K12/KL2) and individual fellowships (F30, F31, F32 and others).
Reproducibility showed up as a session at the 2015 TWD meeting, and by the 2017 meeting NIGMS was presenting it as a pillar of modernized predoctoral training.
What should the instruction cover?
| Topic | Practical content |
|---|---|
| Experimental design | Controls, randomization, blinding, sample size and power |
| Biological variables | Sex as a biological variable, age, strain, and when they matter |
| Resource authentication | Cell line identity, antibody validation, reagent tracking |
| Statistics | Choosing tests before collecting data; avoiding p-hacking; reporting effect sizes |
| Data management | Lab notebooks, version control, data sharing |
| Reporting | Methods detail, reporting guidelines, preregistration where it fits |
Which formats work?
- A required course early in training for shared vocabulary.
- Design reviews where trainees present an experimental plan to peers before starting.
- Reproducibility journal clubs that pick apart methods and statistics, not just findings.
- Reanalysis exercises using public datasets.
- Committee meeting prompts that ask about controls, power and authentication.
How should it appear in an application?
Summarize the plan in the training program description, then attach the detailed curriculum where the announcement asks for it. Name who teaches each part, when trainees take it, how it's reinforced in labs and how you measure whether it changed practice. For poster presentations, the same habits apply: our poster and abstract guide covers reporting methods clearly in a small space.
Questions people ask
Is a single course enough?
A course gives everyone the same foundation, but reviewers look for reinforcement. Pair it with recurring activities such as design reviews before experiments start, reproducibility-focused journal clubs, and rigor questions at committee meetings.
Are there ready-made training modules?
Yes. NIGMS supported a clearinghouse of training modules to enhance data reproducibility, and many universities share course materials. Adapt them to your fields rather than using them unchanged.