In the documentary, we watched in small group today white doctors fell victim to the confirmation bias. Confirmation bias is "the tendency to interpret new evidence as confirmation of one's existing beliefs or theories." Hoffman who looked at mortality rates determined that the "negro race must be destined for extinction" in his extinction thesis but failed to look at other factors such as poverty that could contribute to poor health. Because of his previous belief that blacks are inferior when he saw the higher mortality rates he looked for an explanation that confirmed his existing bias. This emphasizes the importance of data analysis. Hoffman a statistician should have known that correlation does not always equal causation. But instead, he falsely interpreted the statistics and was blind to other factors. This especially important today with the rise of fake news. Many statistics are presented as causation when in fact that is not the case. On an episode of Last Week, Tonight comedian John Oliver emphasizes this. The episode covers the topic of Bad science. They covered faulty studies that stated that chocolate is good for a healthy pregnancy, late-night snacking is bad for your brain, and, most importantly, smelling farts could prevent cancer. The problem is that not only are these studies published but most of the times they are not replicated. Additionally, Oliver said "scientists can feel pressure to churn out positive results in high-profile journals. And when those results do come out, they’re often translated to the public in the form of press releases, which can lead journalists to overhype findings or wrongly convey the science behind them." Most of these faulty studies are sensationalized across local news and why should we not believe them. Shouldn't we trust scientists? Moving forward what can be done to encourage replicating studies? As Oliver said "there is no noble prize for fact-checking." Additionally as a part of the public how do we know what to trust?
https://en.wikipedia.org/wiki/Frederick_Ludwig_Hoffman
https://www.statnews.com/2016/05/09/john-oliver-bad-science/
While I agree confirmation bias is a dangerous thing, I believe that there is more to this. The smelling farts study you bring up was also mentioned to have been taken out of context from a study to be used as click bait. Not only do studies exist that are skewed to favor certain results (say those on obesity paid for by the soda industry), but also perfectly reliable data being misconstrued in order to make it more interesting. Another way to solve this issue besides more funding for replication is further government regulations on what is ethical when executing an experiment (which will hopefully have means to make sure that the government officials handling such matters remain objective).
ReplyDelete