Will more people drown in the pool as Nicolas Cage's movies increase?
"The number of suicides in hanging" and "expenditure on science, space and technology in the United States", "the number of drowning in swimming pool" and "the number of film appearances by Nicolas Cage", "cheese consumption per American" There are also two things that seemingly unrelated at first glance, such as "the number of tangled with bed sheets and deaths"Correlation may be found. However, as a matter of course there is no causal relationship between the two things. Whether these correlations have led to the wrong scientific conclusion so far, movies that are easy to understand on YouTube are released.
This ≠ That - YouTube
A correlation of 66.6% is recognized for the number of people drowning in the pool during the year and the number released for Nicholas Cage's movies.
There is also a correlation of 99.26% between the divorce rate of the Maine state in the US and the margarine consumption rate per capita.
Furthermore, the correlation between the money spent on science and space technology and the suicide rate is 99.79%.
In this way, even if there is correlation between the two things, there is a situation that there is actually no causal relation, and hypothesis of the wrong causal relation by these correlations is called "logical error I will call it "Rikiki Giyou".
Logical errors are not just data that makes mistakes easy to understand as described above. For example, numerous studies have shown that menopausal women undergo hormone replacement therapy to reduce the risk of heart disease. Therefore, many doctors thought that hormone replacement therapy could reduce the risk of heart disease.
However, in a study of women as subjectsRandomized controlled trialIn fact, it turned out that hormone replacement therapy raises the risk of heart disease.
Analysis of the first data revealed that among the women who underwent hormone replacement therapy, those with reduced risk of heart disease had a correct diet and exercise. At first glance it seemed as if hormone replacement therapy had lowered the risk of heart disease, but in that shadow another element was hidden.
As another example, there is a research result that it is easy for a child wearing a night lamp when going to bed when he is a child.
In fact, however, my parents of myopia are often myopic, so I often put lights in my child's room.
These are examples of "latent variables" where there is an element called hidden C, rather than "A is causing B".
As chewing up, people who become lung cancer often carry writers, so we assume that 'writers cause lung cancer'.
In fact, however, there are causes for tobacco that can be put on fire by writers.
Scientists are making daily efforts to avoid such phenomena, but misunderstanding increases as famous media pick up the above-mentioned correlation. What made the revealed how media exaggerated the paper is "Chocolate is effective for dietingIt is a paper entitled.
In this paper, researchers with a doctorate in microbiology actually divided the subjects into "groups that reduce carbohydrates", "groups that eat chocolate by reducing carbohydrates" and "groups that do normal eating habits" I did it. As a result, the group that reduces carbohydrates and eats chocolate has succeeded in reducing the most weight. The media covered this greatly.
In this experiment, we examined the effects on 18 subjects including weight loss, increase / decrease of cholesterol level, sleep quality, blood pressure, etc., for 15 subjects. Normally, when examining a large number of things in a small group of people, a drastic change is necessary to say "it is affected." However, this trend was not seen in this research.
If the paper was peer-reviewed by other researchers, it should have come to light that something is wrong. So, we got a fake institution called "Diet & Health" and asked the publisher to pay 600 euros (about 71 thousand yen) so that the magazine posted the manuscript.
After that, when we gathered the release to a lot of media, the paper quickly appeared in the media and spread.
Evidence by correlation is an important element of science, and in scientific research all correlations can not be ignored.Double blind methodCan not be done in all studies for realistic and ethical reasons, and correlation often plays an important role. We should not forget the aspect that science progresses further by using correlation in research to investigate the causes and effects of things.