Who should be training trainees on reproducibility and ethics?

A recent piece in Nature by Monya Baker, “Reproducibility: Seek out stronger science,” discussed the availability, or lack thereof, of training for junior researchers in reproducibility and ethics. Monya spoke with me while writing that article, asking if I knew what junior researchers were doing about training themselves in statistics, rigor, and reproducibility. I admit I was somewhat bemused by the question. Our conversation was very short. I told her that: training on how to do science seems like more of a priority for the agencies funding the science, and that junior researchers are too busy organizing training for their own career development, which is currently a much more pressing issue for them, given the scarcity of academic jobs.

However, my casual assumption that training in science is not the responsibility of trainees proved to be in contrast to that of Alison Gammie, from the NIH:

“Early-career scientists cannot expect to learn everything they need to know in their own laboratories, or even departments, says Alison Gammie, who administers research training programmes at the U.S. National Institutes of Health (NIH) in Bethesda, Maryland. “Science is changing at an incredible rate, and the current principal investigators and investigators were trained in a different era,” she says. Because many of the available training opportunities are new and not well known, scientists who want to improve their analytical skills must take the initiative to seek — or create — the resources they require (see ‘The learning hunt’).”

I want to address this claim: “Early-career scientists cannot expect to learn everything they need to know in their own laboratories, or even departments.” In particular, I want to highlight that this is in the context of rigorous scientific training.

I have been surprised to hear from numerous voices in NIH that its responsibility is not training researchers, but instead producing research. I take issue with this claim for two reasons:

1. The NIH funds trainees rather than stable scientist positions, mostly through research project grants and not through training grants. This system has perpetuated because trainees are cheaper. In the current debate about whether “trainees” get training and establishing a more sustainable enterprise, it would seem that part of the cost/benefit analysis of hiring a “trainee” versus an employee should be that trainees, who are cheaper to hire, are expected to be cheaper because you have to allow time for training.
2. Federal taxpayer dollars are spent on training scientists, and so it seems that there is a responsibility to the taxpayer to ensure that not only is science being done, but also that scientists are then being trained to benefit society.

What is most alarming about the claim that trainees cannot expect to receive training on how to conduct rigorous science is that this has the potential to fail in both missions: training scientists and ensuring that rigorous science is done in the first place. As David Glass notes in the Nature article, a drive for him writing an educational book on the topic was that, “It seemed odd that we weren’t teaching grad students how to actually perform science.”

In my work, I often talk about the conflict between institutions and federal agencies that insist that grad students and postdocs are “trainees,” while grad students and especially postdocs insist that they are not. One quote from our very first Future of Research symposium in Boston in 2014 was, “If I’m a trainee, then train me!”

One potential solution is that training for junior researchers could come from senior researchers in the department. Setting up a requirement and standard in the department for reproducibility and ethics and requiring younger scientists to follow them could create a culture around reproducibility, and a level of expectation that the department or institution takes responsibility for ensuring this standard is maintained. Training is one thing, but there also needs to be accountability and for people to put this into practice as a priority.

There is, however, open admission in the Nature article that current PIs often have not been trained in these methods themselves, so training cannot be expected to come from them. This begs the very serious question: How is the biomedical research enterprise ensuring that rigorous scientific research is taking place, if training in rigor and reproducibility is expected to be carried out in a non-standard, individual manner by the trainees themselves, despite a lack of protected time for any training? How do you find the balance among training, making sure that science advances fast enough to keep scientists competitive, and doing good science ethically and well?

Do you think training in rigorous science should be an essential part of training scientists, and who should be responsible for that training? Leave your comments below!

For further interest, a fascinating story, “Out of Bounds,” recently appeared in Science, part of which discusses how a training session on ethics and reproducibility was part of an incredible breakdown in an academic relationship.