College Basketball Study Tests a
Landmark Theory of Biodiversity
At the middle of this scientific tug of war is the question of randomness. To model species abundance and distribution, unified neutral theory assumes that the mechanisms that guide success or failure have no pattern.
Yet for Warren and other ecologists who disagree, it’s a vexing challenge to disprove, because identifying the mechanisms that affect a species’ success is so devilishly difficult.
A population of frog species, for example, could grow one year because of good rains but ebb in similar conditions another year because of a blight or the shift to a new region as a result of land use changes. Simply identifying the vast number of mechanisms affecting the frogs’ numbers is a challenge. Connecting these interrelated factors with effects is even trickier.
What if, Warren wondered, one could identify a population whose mechanisms of competition are well-understood and which also generated population dynamics consistent with neutral-theory methods?
This would prove, a priori, that mechanisms are not neutral. If mechanisms do matter, unified neutral theory would still be useful in describing population dynamics at a high level, but it would be inappropriate to use the theory to devalue the close study of ecological mechanisms in the field.
Enter the basketball test. A native of Indiana, home of Larry Bird and arguably the most basketball-crazed state, Warren considers himself a die-hard basketball fan. The idea for the study came to him one day in North Carolina’s Appalachian mountains, while he was counting seed-carrying ants as part of his research on herbaceous forest plants. Warren realized that the pattern of basketball wins is similar to the distribution curves that underpin unified neutral theory.
Warren isn’t the first ecologist to look to examples beyond biology to test aspects of neutral theory. Other researchers have identified even more eclectic sample groups where much is understood about the underlying dynamics. Stock prices, the occurrence of scientific citations and even set lists for the Cowboy Junkies have all been used to generate species abundance distributions, a core piece of evidence for neutral theory.
Basketball seemed like a particularly ripe prospect to Warren. What you’d need, he figured, is lots of species, and there are hundreds of basketball teams. In his approach, each team is analogous to a species, with each of their wins counting as an individual being born. Losses, therefore, are akin to an individual dying. And because basketball teams play many games each year, over many years, the resulting data set is big enough for solid statistical analysis.
Most of all, though, basketball makes an apt case, because the underlying mechanisms of success and failure are so well-understood by so many. As even casual fans can attest, winning in a season or over many years comes from a combination of strategies. Among dominant teams—such as Duke, North Carolina, UCLA, or this season’s winner, Connecticut—some combination of great coaching and strong recruiting, among other factors, has led to long runs of dominance.
For the statistical trial, Warren crunched won-lost records from 327 NCAA Division I men’s basketball teams, a pool of data covering some 20,000 games spanning 2004 through 2008. It may come as no surprise that, across so many games, the most competitive teams generated many more individuals, or wins.
As he suspected, Warren found that his data pool of winning basketball species produced abundance distributions pretty much identical to those observed in countless ecological communities. For most teams, wins are rare, but for a few they are very common.
In short, if basketball wins are not random, then neither is species success. Conversely, if the logic of unified neutral theory were applied to basketball, “then our findings would suggest that the top seed in the tournament is no more likely to win than the last seed in each bracket,” says Mark Bradford, a co-author and assistant professor of terrestrial ecosystem ecology at F&ES. “We know that’s not true.”
For ecology writ large, the implication of the basketball study isn’t to disprove neutral theory. The study does knock out a cornerstone piece of evidence, though, says Warren: “We show that a community, in this case college basketball—which is undoubtedly structured by competition—appears random when current methods for assessing biodiversity are used.”
Understanding what drives these dynamics is critical to both explaining species distribution today and guiding species conservation policy.
Reflecting on decades spent studying amphibian ecosystems—searching for causes of an ongoing population decline in frog, toad and related species—Dave Skelly, a co-author and professor of ecology at F&ES, acknowledged that the drive to infer process from pattern is one of science’s greatest animating forces.
Yet that search shouldn’t take away from on-the-ground study of mechanisms affecting species diversity. The hunt for a culprit in amphibian decline, for example, has led researchers to multiple causes. In the tropics, the chief culprit is a fungus. But in North America, most amphibians are immune to the fungus. Here, habitat loss is the greater threat.
“The upshot is that if you tried to ascribe a single cause to all this, you’d get it right in one environment, but dreadfully wrong in others,” says Skelly.
For conservation practices, the stakes for finding the balance are getting higher. In addition to fitting mathematical models to broad biodiversity patterns, “we need to put on our boots and head back to the woods to figure out why some species are common and so many are uncommon,” says Warren.
“Otherwise, we may find ourselves unable to manage species in the face of global environmental change.”