Login

ADVERTISEMENT

BRENDA LOOPER: Without cause

Doesn't necessarily follow

Dealing with a stomach bug, a cold, and the fever associated with them has made the last week a little surreal. I'm pretty sure there weren't dancing munchkins in the newsroom. But surely the president didn't spend part of the weekend urging people to sue ABC News because of stock losses.

Oh ... but the munchkins still weren't there, right?

Stocks were already volatile Friday morning on news of Michael Flynn's expected guilty plea in federal court, and weren't helped by Brian Ross' erroneous report tying Donald Trump the candidate to Flynn's overture to Russia. (ABC's response, by the way, was appropriate--unpaid suspension for a month is a serious consequence for sloppy reporting, as is being pulled from specific stories.)

I'm no expert on the stock market, but historically speaking, news reports (and the occasional ill-advised tweet) have the tendency of having only a temporary effect on stocks. Indeed, the 350 or so points the Dow lost earlier in the day were almost completely made up by close of the day. Likewise, Trump's tweets about companies that displeased him were found by the Wall Street Journal to have had only limited effect, with nearly all of the 12 companies studied having recovered quickly, some gaining significantly; Toyota was the only one that had not completely made up the loss.

But then again, was the Ross report the reason for the dip Friday? (And that was all it was, especially considering the volume traded on the New York Stock Exchange; 350 points out of 24,000 is much different than 350 out of 7,000.) Maybe, maybe not.

Aaron Brown of Bloomberg View wrote in July: "What about all the political turmoil, the populist revolts and terrorism? By and large, the market anticipates news events about 18 months in the future. It's not perfect, of course, but it's a lot better than experts and commentators. It's silly to expect today's news headlines to affect today's stock prices in a large way, with the exception of truly unanticipated events such as earthquakes."

Flynn being charged wasn't exactly unexpected, now, was it?

Sue Chang of MarketWatch talked to Richard Hastings, macro strategist at Seaport Global Securities, who expected Friday's Flynn news would be just a speed bump. "We've seen some dips on Trump legal risks before," Hastings told Chang, "and the Flynn situation could trigger a new outbreak of Trump allegations. But going from minor stuff to bigger steps seems like a big chasm, and we doubt Trump is doing things that totally [go] into the disaster bucket."

Stories like those specifically tying Ross' report to the stock dip provide an excellent illustration of the old correlation versus causation logical fallacy. And you know how I love me some logical fallacies.

Essentially, just because two things happen in succession or consistently correlate with each other, it does not mean that one was caused by the other. It can, but it's not guaranteed.

One of my favorite spurious correlations comes from the Church of the Flying Spaghetti Monster (the Pastafarians). Pirates (not the modern version) are revered as the first Pastafarians, and the "religion" holds that the decline of swashbuckling brigands caused global warming. Sure enough, graphs show that as the number of pirates fell, global temperatures rose. So that must mean it's true, right? And that if we just create a new generation of Capt. Jack Sparrows we could reverse climate change altogether, right?

Except that with a little statistical tomfoolery you can make even the most disparate things seem to correlate (such as per capita margarine consumption and the divorce rate in Maine). Between spurious correlations and cherry-picking, sometimes it's hard to know what statistics to trust.

Thus the old saw, "Correlation does not imply causation." There may be causation, but in the scientific sense, without sound methodology, controlled variables and clear-cut evidence, you can't say for certain that something caused something else, no matter the degree of correlation. Confounding factors--unseen variables--are, according to RationalWiki, major reasons that "correlation does not equal causation, as the causality could be the confounding factor, and may well be counterintuitive." It could also be sheer coincidence.

But that doesn't serve those who wish to further an agenda, especially when the people targeted aren't intellectually curious enough to dig deeper than the surface.

Harvard Business Review, in its June 2015 issue, advised caution when confronted with graphs that compare dissimilar variables (apples and orangutans), skewed scales (such as weighting to get lines to match), or unrelated data sets (margarine and divorce) plotted together that imply a cause and effect.

So please, be cautious ... or you can keep believing whatever people tell you about cause and effect. And lay off the cheese--the higher per capita cheese consumption is, the more deaths there are of people tangled in bedsheets. Don't look at me; that's what the graph told me ... and the dancing munchkins.

------------v------------

Assistant Editor Brenda Looper is editor of the Voices page. Read her blog at blooper0223.wordpress.com. Email her at blooper@arkansasonline.com.

Editorial on 12/06/2017

Log in to comment