“Correlation is not causation” is a methodological maxim invoked repeatedly across episodes of The Bob Zadek Show, where it serves both as a caution against overreading statistical association and as a recurring point of contention between host and guests. The excerpts show the phrase applied to police body cameras, federal dietary guidance, air-pollution epidemiology, mathematical economics, and bail reform. In each case the speaker uses the maxim to question whether an observed association between two variables licenses a causal conclusion, though the speakers differ on how easily that gap can be closed.
Body cameras and the Rialto study
Matthew Feeney raised the maxim in the context of police body cameras, discussing a study in Rialto, California, where a new police chief distributed body cameras to officers and compared use-of-force incidents and complaints against police in the year the cameras were worn with previous years. Feeney reported that the study found a dramatic decline in both use-of-force incidents and complaints, then immediately cautioned that correlation is not causation, noting that the chief was new and had implemented a number of reforms Matthew Feeney on Best Practices for Police Body Cameras (2016). He added that other studies on body cameras found similar effects, and that the reduction in complaints could probably be partially attributed to cameras capturing evidence. He described at least one incident in which an officer was falsely accused of sexual harassment by someone who did not seem to realize the officer was wearing a body camera.
Bob Zadek framed the discussion by asking about the anticipated benefits police departments had when purchasing the technology and the extent to which those benefits had been realized. After Feeney’s caveat, Zadek characterized the evidence as tangible and real rather than assumption, saying that the mere presence of a police body camera alters behavior and eliminates false allegations both ways. Feeney agreed in part but warned against overreliance: because a chest-mounted camera does not record what the officer saw when the officer moves his head, viewers should not assume they are seeing what the officer saw. Zadek summarized this as seeing what the officer’s chest is seeing, not what his eyes have seen. Feeney noted that at least one manufacturer deliberately does not make body-camera lenses wider than the human eye, and that while cameras should not be relied on for the whole picture, they can capture evidence vital in police misconduct investigations and investigations against citizens.
Dietary advice and the appeal to history
The maxim returned in a 2018 episode with Terence Kealey, where Zadek observed that the phrase “correlation is not causation” is one he has often used on the show, and that two events occurring around the same time does not automatically mean A caused B. Zadek then put it to Kealey that most thinking people of science had concluded the relationship at issue was more than mere correlation and was in fact causation “Breakfast is Good For You” and Other Dangerous Government-Sponsored Myths (2018). Kealey agreed emphatically that correlation is not causation, but argued that the discipline of history rests on the concept that someone does something at time A and something happens at time B, and that we assume some sort of causation or we would not have the discipline we call history. He cited the federal government’s 1977 dietary advice, said that advice caused type 2 diabetes and obesity, and noted that three years later type 2 diabetes and obesity took off. In his formulation the relationship was only correlation in a formal sense but was also history, a discipline he said we can put a lot of confidence in.
Zadek humanized the government decision by naming Senator George McGovern, describing him as a presidential candidate and a progressive in the Senate, and asking how the government came to offer bad advice. Kealey described an epidemic of heart attacks in America after the Second World War, said no one really knew what caused it, and invoked H.L. Mencken’s line that for every complicated story there is a solution that is simple and wrong. He described McGovern as a deeply bruised man who had lost badly against Richard Nixon and was going to lose his senatorial reelection campaign three years later, and who looked for something simple to capture the interest of the American people. Kealey described a debate in the scientific community between Ancel Keys, a professor from Minnesota, who argued that heart attacks are caused by fat in the arteries and that people should therefore avoid eating fat, and an army of professors of biochemistry, nutrition, physiology and health who said it was naive and simplistic to say the cause was simply fat in the diet.
Epidemiology and the misuse of correlation
Theodore Dalrymple addressed the maxim in a 2019 episode on negligence and unaccountability, in a segment on epidemiology and the misuse of correlation. He said many papers are epidemiological in origin, examining the distribution of diseases through populations, that they are highly mathematical, that not one in a thousand doctors understands the mathematics, and that they draw very large conclusions from very doubtful information Theodore Dalrymple on Negligence and Unaccountability (2019). His example was epidemiologists who looked at the all-cause death rate from air pollution, found a correlation between the level of pollution and the overall death rate in a population, and then concluded that permissible levels of pollution in the whole of the United States ought to be altered to take the correlation into account and produce equality of outcome in all areas. Dalrymple said this was done without realization of how difficult and expensive it would be, and without realizing that the logic would not allow any pollution at all, since nobody would say any pollution is good. He called it a very elementary thing that correlation is not causation, and the excerpt breaks off as he is saying that just because two things are correlated.
Zadek responded by asking what the thesis of the article was, saying it surprised him when he learned from Dalrymple’s book that this was a topic, and arguing that air pollution is beyond the competence of doctors to do something about or understand, relating instead to physical science and chemistry and other disciplines where doctors bring little more to the table than anybody else. He asked whether it surprised Dalrymple that the topic itself would appear in the New England Journal of Medicine. The excerpt then breaks off at a section heading on tobacco and lung cancer.
Economics, prediction and the charge against mathematical models
In a 2021 episode on Austrian economics, Zadek invoked the maxim against economists he associated with Paul Samuelson, saying that those who treat economics not as a social science but as a science look back and forget that correlation is not causation, see a couple of variables that seem to act in concert, and conclude that those variables can predict the future Austrian Economics Triumphs (2021). He contrasted this with the view that economic activity cannot be predicted except by acknowledging that human beings act in their self-interest, try to improve their lot, are to some degree generous, and respond collectively as each person responds individually. He cited Don Boudreaux of the Mercatus Institute, who he said has written with authority and passion that the lowest 10 or 20 percent of Americans today have a quality of life in every measurable respect better than that of Cornelius Vanderbilt or Andrew Carnegie, and attributed this to the byproduct of an active free market.
Jeff Deist largely agreed, saying most economists today do not understand much about economic history, that brilliant young people come out of PhD programs or business schools such as Wharton with superb number-crunching skills but without a sense of history, and that many under a certain age have never seen a bear market. He called the unwillingness to teach the past one of the big failings of the economics profession, and more importantly a failure of imagination. He said that in 1900 the United States was not wealthier than European countries per capita, that the country got really wealthy during the 20th century despite what he considered disastrous New Deal and Great Society entitlement programs and disastrous foreign policy, and that thinking this wealth is baked into the cake and will continue no matter what is done at the monetary or fiscal level is dangerous thinking. The excerpt then breaks off at a section heading on Operation Warp Speed.
Bail reform and the Comptroller’s report
The maxim appeared again in a 2022 episode on bail reform, where Zadek explained what he called the economist’s point that correlation is not causation: seeing a behavior between two independent events does not mean one caused the other or resulted from the other, and all that is known for the moment is that two statistics behave in concert Checking in on Bail Reform (2022). He then asked Joe Lancaster about a study by the New York City Comptroller on the effects of bail reform, which Zadek summarized as making it easier to release people who could not come up with money back into society pending trial or disposition.
Lancaster said that looking at data from 2019, 2020 and 2021, between pre-reform and post-reform numbers, there was a significant lowering in the number of people detained before trial and no corollary rise in crime among that group. He acknowledged other factors, including roughly a year and nine months of COVID during which there were far fewer hearings, but said that on a granular basis the people released who likely would not have been released beforehand did not go on to reoffend in large numbers, with the reoffending rate about equivalent before and after. He called it troubling that fewer people were assigned cash bail amounts while the amounts assessed and collected increased significantly, with the average bail amount approximately doubling, so that only about half of people who had bail set were ever able to secure bail and get out. Lancaster said the report concluded that when the reforms functioned as intended, allowing people to leave on their own recognizance, the reforms worked, and that the reforms should be strengthened and continued as written, making the newest weakening of those reforms contrary to the report.
Zadek added that the New York City bail reform package required courts for the first time to take into consideration the ability of the alleged perpetrator to pay, because one of the clearest abuses was the idea that a person is innocent until proven guilty unless he has no money, in which case he is stuck in jail. He noted that studies discussed on an earlier show found that low-income people accused of a crime who cannot make bail often serve more time awaiting trial than they would have served had they been convicted, and described the cascading damage to employment, family rent and permanent record. The excerpt then breaks off at a section heading on the bail bonds industry.
Across episodes
The excerpts show the maxim recurring across five episodes without a single settled doctrine. Feeney and Dalrymple state it as a straightforward caution against overreading associations, while Kealey treats correlation plus historical sequence as sufficient for confidence and Zadek presses the same maxim against Samuelsonian economics and against causal claims about bail reform. What changes between the earlier and later treatments is the target rather than the principle: the 2016 and 2019 episodes apply it to empirical studies of cameras and pollution, the 2018 episode tests whether history can close the gap, and the 2021 and 2022 episodes deploy it against prediction and against reading crime data as a consequence of reform.
What the sources do not cover
The excerpts do not define causation formally, do not present a statistical test or criterion for distinguishing correlation from causation, and do not resolve whether any of the disputed relationships were in fact causal. They also do not state the outcome of the bail reform legislation discussed, the fate of the federal dietary advice, or the conclusions of the tobacco and lung cancer segment, which breaks off at a heading. No excerpt supplies a citation to a study, a court holding, or a statutory text beyond the names and reports the speakers mention.