By Michael Paladini, Vice President, Director of Analytics
Here are three funny quotes on measuring and counting, all of which provide insight into the pitfalls we face as marketing analysts (or any other kinds of analyst, as well).
The first, often mis-attributed to Yogi Berra, goes something like “Forecasting is difficult because we don’t know what’s going to happen in the future.” Beyond pointing to the banal observation that the future is always uncertain—and more uncertain than we care to admit—the remark also highlights our tendency to generalize from too few data points. As the brilliant John Elder of Elder Research entitled one of his 2014 papers, “It Is a Mistake to Extrapolate.”
Remember Arnold Schwarzenegger (bodybuilder, actor, governor)? Back in his acting days, his successive year-over-year earnings increases were huge. So large in fact that, to make a point, one statistical wag extrapolated them. When he did, “Ahnold’s” earnings exceeded the nation’s GDP after about ten years. Not sustainable, of course.
Here’s the shape over time that governs the world:
Or even this (although we don’t like to think about it):
Rather than simply extrapolating from present trends (and too few data points), build the unexpected into your plans. Anticipate discontinuities. Avoid excessive optimism. Have a little humility about tomorrow.
Then there’s this. Did you hear the one about “the statistician who had his head in the oven and his feet in a bucket of ice, and who said, ‘On average, I feel fine!’ ”? I’ve always liked that one. It points, of course, to how terribly misleading averages can be.
The problem, naturally, is that averages are skewed by outliers, by the extreme values in the data set you’re examining. One of the best examples comes from discussions about national income.
The average household income in the U.S. in 2011 (going back a few years, for some solid figures) was about $69,800. Not bad for a family. But the median income was only $50,500. The median income is the mid-point income, meaning half of households are above that figure and half are below. A family household making only $50k a year would be, in most parts of the country, just barely getting by. And half the country, a few years ago, was making less than that. Median is a better figure to use, presenting a truer picture of economic conditions.
As for the difference between the average and the median? The enormous concentration of wealth and income at the top of the economic ladder skews the average up. Income disparity is currently a subject of great national debate, and will likely be an issue in the next presidential election.
My third favorite funny remark on statistics comes from Ron Coase, the Nobel Prize–winning economist, who said, “If you torture the data long enough, they will confess to anything.” The idea here is to be wary of the misuse of statistics. Numbers, seemingly factual and objective, aren’t really neutral things. They are as subject to interpretation—and distortion—as words.
As you read studies or review findings, it pays to have a skeptical eye. Make sure the numbers are contextualized. Watch out for data that have been cherry picked. Ensure that factors said to be important aren’t being given undue weight. Ask whether the author isn’t leaping to a conclusion from what’s really too little information.
Analytics is best conducted by those who are always willing to question. As always, we welcome your comments.
The post Three Funny Quotes I Love About Statistics appeared first on Wilde Agency.
