A pet peeve (HAR!!!): computer geeks and statistics experts reducing complex issue into “simple” charts and graphs that have apparent credibility because of their form rather than their substance. I encounter this seductive form of fake erudition—“You can’t argue with statistics!”—in every field I explore: baseball, politics (Sorry, Nate Silver), social science, science (climate change models are a spectacular example), education. “Simple, straightforward” arrays of statistics that hide biases, dubious assumptions, projections, value judgments, undisclosed definitions, and who knows what else are presented to persuade on the false representation that they are “hard” representations of fact.  Very frequently, they are not, and when they are not, they incompetent, irresponsible and dishonest. Also arrogant to the core.

You could find no better example of this than this dog chart, by David McCandless, which purports to summarize “big data”—read: “data that can be manipulated to show whatever you want it to show” indicating which dog breeds are “over-rated,” as well as how they score on a “costs and benefits” scale. The fact that anyone could take such a garbage graphic seriously is unsettling, but of course, it will only impress people who know absolutely nothing about dogs and dog breeds. That’s what all such arrays of statistics are for: to convince and mislead those who are too lazy or uninformed to really understand the topic at hand and its complexities, but who want to lay claim to an “informed opinion.”

Just look at this monstrosity (you can read it better here): Continue reading →