One of the biggest problems in science, in my opinion, is the dearth of reproducibility. When I was an undergrad, I realized that virtually none of the studies I read were verified by repeating the experiment. Sure, studies are peer reviewed, but the reviewing process does not account for false positives. On the one hand, it is understandable because repetition requires time and money which are admittedly in low supply for researchers. However, science relies on reproducibility because it adds credibility to a conclusion.
I recently came across an article that describes an idea to combat the issue: create markets where instead of asking the experts about their opinions about a study, have them bet on studies that they would deem as reproducible. It’s a fascinating idea that really does make sense when you think about it. This method essentially incentivizes people to do their due diligence because their money is at risk. It’s like a stock market for scientific studies; you could buy shares of a study and if it was reproducible you’d increase your value. If you deem a study to be flawed, you’d be able to short sell it. A group of researchers constructed a situation similar to what I just described and found some interesting results. The market predicted almost 75% verifiable replications, which is a significant increase in accuracy (up from 58% before the market initiation). To me, I wonder how repeated iterations of this scheme would pan out. Would the percent of correct predictions converge to a maximum or would it experience volatility similar to what we see in the stock market? Either way, I think that this approach is a super interesting way to deal with the reproducibility problem in science!