3 ms·

Is it "Reading too much science" or is it just that human characteristic of confirming ones own biases (ie anchoring bias, confirmation bias or whatever they're

Is it "Reading too much science" or is it just that human characteristic of confirming ones own biases (ie anchoring bias, confirmation bias or whatever they're called)? Personally, I think its a lack of Baloney detection skills: https://www.openculture.com/2016/04/carl-sagan-presents-his-baloney-detection-kit-8-tools-for-skeptical-thinking.html https://www.openculture.com/2016/04/carl-sagan-presents-his-... Don't just leave it to the experts! Also, I'm not sure you can say that "Most published research is false". There are many degrees between true and false. In the physical sciences, with which I'm familiar, the papers are demonstrably 'mostly true'. For example, as a bi-product of an experiment last week we observed an Ekman spiral (first published in 1905 https://en.wikipedia.org/wiki/Ekman_spiral https://en.wikipedia.org/wiki/Ekman_spiral).

astrange 5y ago
> Also, I'm not sure you can say that "Most published research is false". There are many degrees between true and false. In the physical sciences, with which I'm familiar, the papers are demonstrably 'mostly true'. For example, as a bi-product of an experiment last week we observed an Ekman spiral (first published in 1905 https://en.wikipedia.org/wiki/Ekman_spiral https://en.wikipedia.org/wiki/Ekman_spiral). I agree it depends on the field. I was thinking of medicine, where there's a well known paper about this[1] that's led to some improvements like pre-registered experiments, although there's some newer ones that show ongoing problems in social sciences[2]. [1] https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0020124 https://journals.plos.org/plosmedicine/article?id=10.1371/jo... (Why Most Published Research Findings Are False) [2] https://advances.sciencemag.org/content/7/21/eabd1705 https://advances.sciencemag.org/content/7/21/eabd1705 (Nonreplicable publications are cited more than replicable ones)

astrange 5y ago

> Also, I'm not sure you can say that "Most published research is false". There are many degrees between true and false. In the physical sciences, with which I'm familiar, the papers are demonstrably 'mostly true'. For example, as a bi-product of an experiment last week we observed an Ekman spiral (first published in 1905 https://en.wikipedia.org/wiki/Ekman_spiral https://en.wikipedia.org/wiki/Ekman_spiral). I agree it depends on the field. I was thinking of medicine, where there's a well known paper about this[1] that's led to some improvements like pre-registered experiments, although there's some newer ones that show ongoing problems in social sciences[2]. [1] https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0020124 https://journals.plos.org/plosmedicine/article?id=10.1371/jo... (Why Most Published Research Findings Are False) [2] https://advances.sciencemag.org/content/7/21/eabd1705 https://advances.sciencemag.org/content/7/21/eabd1705 (Nonreplicable publications are cited more than replicable ones)