The US court may decide what the words "repeat scientific results" mean
Disputes in the turbulent and harsh world of science are settled over time, when overwhelming evidence accumulates in favor of one of the hypotheses. Up to now.
On April 10, 2006, economist John R. Lott, Jr., formerly of the American Enterprise Institute, filed a defamation lawsuit against economist Stephen D. Levitt of the University of Chicago and against the Harper-Collins publishing house that published Levitt's book "Freakonomics" in 2005 ". At the center of the discussion is Levitt's intention when he wrote that other scientists failed to "replicate the results of Lott's research*" that appeared in his 1998 book "More Guns, Less Crime". Lott used a sophisticated statistical analysis to analyze the differences between the gun carrying laws in different states in the US. He found that in countries that allowed their citizens to carry concealed weapons by law, there was a statistically significant decrease in the rates of robbery, rape and murder compared to countries that did not enact such laws.
As usual in studies like this, which are politically charged, a stormy controversy arose following the publication of Lot's book. During the flurry of lectures and scientific articles published on the subject, studies were presented that repeated Lot's results and other studies that failed to do so. For example, Lott confronted his critics about the results in a series of articles in the Stanford University Law Journal (the articles are available online at: http://papers.ssrn.com).
In his book "Freakonomics" Levitt offers his own explanation for the drop in crime rates in the US in the 90s - the ruling in the "Roe v. Wade" trial that allowed abortion. According to Levitt, there are many chances that children born in a poor family and in a hostile environment will end up in prison as adults. After the ruling, millions of single and poor women had abortions and did not give birth to future criminals. At the end of 20 years, the number of violent criminals has decreased and so has the crime rate. Levitt used comparative statistical analysis to show that in the five states that legalized abortion at least two years before the ruling, crime rates fell before they did in the other 45 states. What's more, the states where abortion rates were the highest in the 70s are also the states where crime dropped the most in the 90s.
One of the explanations that White Peter casually gave was Lot's explanation. In a single paragraph, in the middle of a 30-page chapter, Levitt writes: "Lot's interesting hypothesis is incorrect. Other scientists who tried to repeat the results of his studies came to the conclusion that laws allowing the carrying of weapons do not reduce the scope of crime."
According to the lawsuit, "the term 'repeat results' has an objective and factual meaning" and it is "that other researchers analyzed the exact same data that Lott used and analyzed them in exactly the same way that Lott did to check if they reach the same conclusion." And when Levitt wrote that they did not succeed he actually "accused Lott of falsifying his results."
I asked Levitt what he meant by the words "to repeat the results", and he replied: "I used the term in the same way that most scientists use it - to verify results." Verify, do not repeat. Did he mean to imply that Lot falsified his results? "No, I didn't mean that." In fact, others did accuse Lott of falsifying his data, so I asked Lott why he was suing Levitt. "When some anonymous people I've never heard of before write accusations on the Internet - that's one thing," Lott said. "But the accusations that appear in print, in the book of an economics professor, published by a reputable publishing house, and whose sales are already over a million copies - that's something completely different. In addition, Levitt is a famous person and unfortunately his claims have an impact. Many people asked me after reading Freakonomics, if the claim that others failed to repeat the results of my research is indeed true."
The verb "to return" has a meaning that depends on the intent of the sentence. "Repeating the research method" is a phrase that may correspond to the meaning that Lot is aiming for, but the meaning of the phrase "repeating results" is testing the conclusions that emerged from the research method, and in this case that the presence of weapons in society leads to a reduction in crime. The problem is that such an analysis is so complicated that it is likely that a failure to repeat the results points to sampling errors made during the original study or in the validation studies and not necessarily to falsification.
Mr. Lott: Please tear down the legal wall you built and let us go back to practicing science without lawyers. Replication of results is for testing hypotheses - not simply copying methods - and this mainstay of science can only stand up in an atmosphere of open peer review.
