Again we find the
results not replicable.
False positives teem!
One of the core principles of research is that it should be reproducible, namely that someone else repeating your methods should get the same result as you. But there’s little resources available to reproduce work so it’s often hard to know just how reproducible a result is.
The results of a study by Camerer et al (2018) suggest that reproducibility (at least in certain fields) might be lower than expected. The researchers replicated 21 experiments published in the social sciences in the journals Nature and Science between 2010 and 2015. They found that only 62% of their replications showed evidence consistent with the original studies. Interestingly, they also found evidence to suggest that the research community could predict which studies would replicate and which wouldn’t.
Original research: https://doi.org/10.1038/s41562-018-0399-z