Dr. Mark Gordon on science: what the evidence says · JRE #2262
SUBJECT: SCIENCE
Not a true/false call. Every claim is logged with its sources; read the exhibits below.
70 percent of them can't be reproduced. And when you looked at the actual scientists who did the original work, goes back and tries to reproduce it, 70 percent failure rate. So there are major publications that have talked about this reproducibility error.
What the evidence says 01 / RECORD
The 70 percent figure traces to a 2016 Nature survey of 1,576 researchers, but it measures something different from what is claimed. More than 70 percent of respondents said they had tried and failed to reproduce another scientist's experiment (rates ranged by field, from about 87 percent of chemists to 62 percent for other fields), which is a statistic about researchers' experiences, not a finding that 70 percent of all published studies cannot be reproduced. For reproducing their OWN work, the survey found more than half (roughly 50 percent), not 70 percent, of researchers had failed. The survey itself was nuanced: most respondents said they still trust the published literature, and fewer than 31 percent thought a failure to reproduce meant the original result was probably wrong. Conflating a researcher-experience survey with a blanket 70 percent study-failure rate, and attributing it to original authors, overstates and misframes the source.