Next steps after airing disagreement on a scientific issue with policy implications: a meta-analysis, multi-lab replication and adversarial collaboration

Abstract

Canadian policymakers are interested in determining whether farmed Atlantic salmon, frequently infected with Piscine orthoreovirus (PRV), may threaten wild salmon populations in the Pacific Northwest. A relevant work has been published in BMC Biology by Polinksi and colleagues, but their conclusion that PRV has a negligible impact on the energy expenditure and respiratory performance of sockeye salmon is disputed by Mordecai and colleagues, whose re-analysis is presented in a correspondence article. So, what is the true effect and what should follow this unresolved dispute? We suggest a ‘registered multi-lab replication with adversaries’.

Finding the right power balance: Better study design and collaboration can reduce dependence on statistical power

“…Such designs not only improve the overall power of estimates but also make them more biologically relevant and generalizable. Grant agencies, along with ethics committees, could encourage and specifically fund multi-institutional experiments, through which they could provide more opportunities to researchers from traditionally marginalized groups, spreading EDI in science [ 86 , 89 ] (for a related example of when and how such an experiment could be funded, see [ 90 ]). Such a multi-institutional experiment, combined with a later synthesis, can be seen as a “prospective” meta-analysis [ 91 ].…”