Good question! I suspect that the first thing that’ll come to mind for a lot of people is paywalls and publishing fees – particularly as they limit the access to science for researchers in the Global South. As we discussed in last week’s thread on peer review, I suspect that it would take quite the economic overhaul to adequately address this problem but it’s worth mentioning anyway.
One big thing would be for journals to get more comfortable with published ‘null results’, i.e. papers that don’t have big flashy results but which nonetheless contribute something really valuable to the literature. The literature in many fields is definitely plagued by publication bias and the current norms – where we de-facto know that we’re more likely to get published if we get “significant” results – also incentivises even well-meaning researchers to engage in things like p-hacking.
Journals committed to publishing ‘negative results’ are an interesting approach (e.g. in psychology and biomedicine) but unless those journals get sufficiently prestigious, I’m not sure they solve the incentive problems. I also quite like the idea of registered reports, where journals basically commit to publishing your research on the basis of the quality of your theory and methods rather than on the ‘significance’ of your results. If some big journals started doing that (I know it’s happening in some econ and psych journals), I think that’d be great!