A Better Way
Eve Sounder, Ph.D.
[abstract] Part of the reason we do science badly is that it's hard to do it well. Not "math is hard" hard, but filling an intro section with references we haven't read and pretending to be open-minded in a discussion section can be taxing. We can make science easier by discarding formalities and removing the layers of window-dressing we apply to our data. A simple plan is proposed. In this plan, the pressure to publish research findings would be reduced, and what is published would provide a better representation of one's value to academia than what we've been doing for decades, if not centuries.[\abstract]
The Facts of Life
This camel will proceed gracefully or awkwardly as it makes its way down the road. How well it does depends on who is inside. Their abilities and motivations, and their communicate and cooperation will matter tremendously. That is, they will determine how well the camel walks. However, those who watch the camel's progress will see a camel, not a collection of influences.
|All they see is a camel.
Yes, the emergent "camel" produced by two people and a two-person camel suit is a metaphor for science. We think a lot about the dynamics of our work environments, but all we should have to do is we have to produce information that's worth archiving. We agree on that, yet we still do bad science.
A Better Way
I'd like to propose something wild, which I think would appeal to most researchers, and that is that we not use the journals to publish data or statistical analyses of data. Let's reserve them for reviews, insights, hypotheticals, and thought experiments. And let's make those documents dynamic. Commenting skins could be activated for use by lone reviewers or mass ongoing discussion groups
Data should be self-published in raw and minimally transformed states, with accompanying information on how it was collected; enough information that another party wishing to use it will know what they have before them.
Instructive analyses of data should be archived for the public, and would be best presented on personal blogs, From there, the better studies might go on to aggregation sites that are as much social as scientific in purpose.
Human subjects should have the opinion of tagging any data they provide so that over the years, relationships among various outcomes that vary among individuals can eventually might eventually coalesce as syndromes (not necessary pathological, just related).
In this system, producing poor data wold be of little value except in agenda-pushing. Few are able to push agendas without being caught out in short order, though.
Rather than regarding others' use of one's data as parasitic, one could rightly take pride in collecting data that others want to analyze.
Tenure would no longer depend on lab access and funding. One's ability to think and write commentary that informs others is as I'm others view the problems they are trying to solve has never depended on those things, and job security and advancement should not, either.
We can be a much better camel than we have been. Freeing ourselves of The Format -- intro, method, results, distortion -- can only be for the best.