Science and speculation

I recently read several articles about a visiting baseball player who was subjected to racial hazing in a game at Fenway Park. The sense of these articles is this attitude reflects on the city of Boston, and on America at large. This is an all-too-common tendency today, to extrapolate a statement, an incident, or even data, to have far broader applicability than the evidence warrants.


Science is much in the news, with accusations of “science denial” or climate change skepticism, Creationists disputing evolutionary evidence, scientist-celebrities making bold pronouncements, along with front-page scientific studies that were once lauded and have now been refuted (often on the back page).


Though the laws of science—gravitation, thermodynamics, the conservation of mass and energy—are fixed, for all practical purposes anyway, the interaction of influencing factors and forces in complex systems like the Earth’s climate, Lake Michigan, even local weather on a given day, can produce a variety of outcomes, some predictable, some surprising. Surprising not because the laws of science have been violated, but because the system, the combination of dozens or hundreds of factors and forces, couldn’t be adequately modeled, or the input to the model (data/design) was flawed or incomplete.


I’ve seen my share of bad science and bad data (sadly, guilty myself on occasion). I’ve learned that while we need to rely on data, honest skepticism is an important aspect of the scientific method. On many occasions, scientists—experts—have reached a consensus on something that was subsequently proven to be false. As Matt Ridley wrote in a 2013 Wall Street Journal article, “Science is about evidence, not consensus.” I’m with Mr. Ridley. I don’t care about consensus, no matter how passionate or morally indignant. I want to see the data and the evidence, and how it’s linked to conclusions.


Drawing broad conclusions from evidence or evidence-based models has inherent risks. This doesn’t mean we can’t (and don’t) rely on evidence and models, only that we should understand the limitations and risks of doing so. Some years back, The Wall Street Journal published my rebuttal to their news article entitled, Study Finds Global Warming Is Killing Frogs: “When science records what it observes, when it measures phenomena, and when it faithfully and accurately models that data, its findings are valid, useful and reliable. But when scientists…offer speculation…credibility and reliability are diminished, sometimes drastically. Thus, the observation that the frog population worldwide is declining…in combination with models that purport to demonstrate global warming, is not (yet) sufficient to assert the title of your article. This conclusion is speculative, as it is based on the assumption that warmer temperatures at higher elevations in Costa Rica are responsible for…the fungus that is infecting the frogs.”


If extrapolation of data/evidence is a problem with respect to the hard sciences, how much more so with the social sciences? What’s needed is a clear understanding of (1) how the evidence/data was obtained; (2) the extent to which this evidence/data applies to the system being studied, along with identification of any gaps or missing pieces; and (3) the extent to which the model faithfully describes the system being studied. Can speculative conclusions, such as Study Finds Global Warming Is Killing Frogs, be justified by the data and evidence? Stephen Hawking recently revised his “authoritative” conclusion that humankind has 1,000 years to escape the planet to 100 years. Hawking is a recognized expert on theoretical physics, but the fate of the planet is far too complex for 1,000 years, 100 years, or any number to be credible. Just because an authoritative individual or institution says something doesn’t make it so.


As to that fan, or handful of fans, at Fenway Park, what they said is on them, and based on the evidence, that’s what science would say too.