The Canseco Effect
Wherever he went, Jose Canseco made his teammates better power hitters. Can statistics be used to find juicers?
When Jose Canseco finally came clean as the "Godfather of Steroids" in 2005, his use of performance-enhancing drugs had already been a matter of speculation for nearly two decades. In his tell-all biography, Juiced, Canseco alleged that he schooled many of his power-hitting teammates on integrating steroids and growth hormones into their training regimens, including such baseball greats as Mark McGwire and Jason Giambi. Yet given his earlier denials of steroid use, his credibility was tenuous.
Now, a couple of Israeli economists, Eric Gould and Todd Kaplan, have released a study backing up Canseco's claims that he was "The Chemist" of baseball. Their work shows how the careful use of data can be an important tool in ferreting out drug use in baseball and other sports. They analyze the performance of Canseco's teammates—the players who had access to his locker-room pharmacy and expertise—and find that after playing with Canseco, players hit more home runs and otherwise boosted their numbers in the areas most affected by steroid use. Sometimes data speak more credibly—if not louder—than words.
Gould and Kaplan are labor economists, not baseball fans. Their primary interest in studying Jose Canseco's influence is in illustrating the sorts of pernicious behavior that can be learned in the locker room, by the water cooler, and in the workplace more generally. A few years ago, for example, there was a pair of scandals involving large numbers of false disability claims at the Long Island Railroad and the Boston Fire Department. (The Boston case broke when one of the purportedly disabled firemen placed eighth in a national bodybuilding competition.) In both cases, fraudsters possibly learned the benefits of cheating—and maybe even the details of how to do it—from their co-workers. Being around other cheaters also provides the ready excuse we've all heard before (and used ourselves): "Everyone's doing it." You'd be a chump not to get disability payments while everyone else is relaxing by the pool or to lose baseball games to competitors who have an unfair edge.
To provide a statistical assessment of Canseco's alleged influence, Gould and Kaplan compared the performances of every hitter and pitcher who played with Canseco, and analyzed how they changed after exposure to him. Focusing on the power-positions players—catcher, first base, outfield, and designated hitter—who would most benefit from extra heft and bulk, Gould and Kaplan found that contact with Canseco was worth an extra two home runs per year in the seasons that followed. Canseco's teammates also saw increases in other power statistics—half a dozen extra runs batted in per season, a one-point boost to slugging percentage, and a handful of additional walks. Meanwhile Canseco did not seem to help teammates in their fielding, base-stealing, and other nonpower areas. (In results not reported in the study, Gould and Kaplan also found that pitchers were able to put in more innings when exposed to Canseco, another indication of The Chemist's hand in helping his teammates work harder and longer.)
Of course, it's possible that Canseco's outsize influence could be benign—maybe he shared with his fellow power hitters a set of batting tips that proved effective. But if this is the case, Canseco's abilities as a hitting instructor were quite unique—Gould and Kaplan looked at the effect 30 other power hitters of Canseco's era had on their teammates and found that none of them had a statistically significant influence on the hitting performance of teammates. (Some of these were in fact Canseco's original disciples, suggesting, perhaps, that not all users become proselytizers.) What's more, the Canseco effect disappears after 2003, when baseball instituted random drug testing and punishments for those found guilty. If Canseco was merely offering innocent performance-enhancing advice, it stopped working with the advent of drug testing.
Can Gould and Kaplan's approach be applied to other sports? There are striking parallels between the Canseco saga and the ongoing doping scandal in competitive cycling. This year's Tour de France began under yet another cloud of controversy with the publication of former champion Floyd Landis' detailed description of alleged doping with seven-time tour winner Lance Armstrong. According to Landis, Armstrong lorded over a complex and multi-faceted operation that included bike sales (to fund the doping program) and clandestine roadside blood transfusions for him and the rest of his U.S. Postal Service team. Not surprisingly, Landis' interview, published by the Wall Street Journal, was followed by sharp denials from Armstrong and "no comments" from others implicated in the story.
Like Canseco, Landis himself initially denied the doping charges that led to his fall from grace, even writing a book-length defense titled Positively False. Many are using this flip-flopping to question his credibility—Armstrong points to Landis' repeated lying under oath as evidence that he can't be trusted.
The same approach employed by Gould and Kaplan could conceivably be applied to implicate—or exonerate—Armstrong. Cycling, like baseball, is a combination of individual and team effort. Each person rides his own bike, and at least on some days—when riders are competing in individual time trials, for instance—you could look for an Armstrong effect. If Armstrong's presence on the team improved the performance of his teammates, it wouldn't provide a smoking gun—just as Canseco's impact may have come in the form of batting wisdom, not drugs, Armstrong might have motivated his teammates to train harder and better. But smart researchers like Gould and Kaplan will surely find creative ways—as they have in uncovering Canseco's influence—to begin to cut through the lies that have come to characterize the debate on performance-enhancing drugs in baseball, cycling, and other sports.
How do these findings on what Gould and Kaplan term "ethical spillovers" leave us feeling about players who dope? While it's no excuse, it does matter if everyone else is doing it. One of the basic tenets of social psychology is the fundamental attribution error—we tend to attribute too much blame to the individual and not enough to his circumstances. It explains why we see baseball players and cyclists as dirty rotten cheaters and rail against unethical Boston firemen rather than asking whether we'd do the same in their circumstances. It also explains why it's so hard to reform a culture of cheating and corruption when everyone is doing it.
Ray Fisman is the Lambert Family professor of social enterprise and director of the Social Enterprise Program at the Columbia Business School. He is the co-author, with Tim Sullivan, of The Org: The Underlying Logic of the Office. Follow him on Twitter.
Photograph of Jose Canseco by Christian Petersen/Getty Images.