I’m jumping onto some of Cheryl’s recent posts on the “lab leak” hypothesis and the argument that the lack of a confirmed zoonotic host 17 months into the pandemic is a good reason to change our priors.
Science is often slow. Clinical trials and secondary data analysis can be fairly fast. COVID has led to pandemic related research to get prioritized review, decision-making, and web-only publication for some research so turn-around times for some types of science from submission to publication is short, but not all science is clinical trials, observational trials or clinical experience.
I’m in a field (health policy) that has a culture/expectation of frequent and relatively fast publication. On Friday, we got notified that one of our COVID related policy papers had received a soft acceptance. We formulated the question in March 2020, needed six months for the data to actually happen, and then the team spent a few months analyzing data and writing. It was under review and revision for just under six months. We are looking at fifteen months from an idea to acceptance, and given the journal in question, we are looking at likely eighteen to twenty four months from idea to official publication. Twenty four months is FAST.
I have another paper that was recently accepted. This was super-quick with perhaps five months from “hmm… this is a peer-reviewable idea” to acceptance. My co-authors and I performed an analysis on data that we already had for several years for other projects. We had been informally and casually talking through the idea for at least a year or more and all of us are publishing papers that are adjacent to this manuscript on a regular basis. This was an opportunistic publication that we think is timely, relevant and useful. It is only timely because we had pre-exisiting data, and we had been thinking hard about the problem in slightly different contexts for several other manuscripts and grant applications.
Those two papers are fast within the academic context because they are using administrative data that is fundamentally clean and we were applying fairly straightforward techniques.
Science is slow when new data needs to be collected, when new concepts and techniques need to be validated and explored. Science is slow when it is messy and complex. The belief that if nothing has been published in eighteen months after the barest parameters of the problem space were laid out means that we should significantly update our priors is likely to be a false belief. The belief that a year and a half is enough time for complex data collection to be complete and exhaustive for any problem greater than “what flavor Mac and Cheese is best for my eight year old” shows that the holder of that belief has never had to do raw data collection, cleaning and processing.
Some science can be fast. Some science is, necessarily, working on a time scale of years instead of weeks or months.