INFLUXUS RECIPROCI FALSUM - “The Correlation False”
The idea for this show started when we ran across the book Spurious Correlations by Tyler Vigen at the San Jose Tech Museum. Ed and Ron both picked it up and immediately loved it. And a show was born! So what’s up with the show title??? Literally translated, it means “the correlation false” and if you’ve been listening to the show for any period of time you probably picked up on the fact that Ed understands a thing or two about Latin.
Now that you’ve read this far, Episode 250 of The Soul of Enterprise was — predictably — about the confusion between correlation and causation. Folks - some of these examples are hilarious!
Let’s get a few key things out of the way before diving into the examples.
Correlation: two things vary together. Ron likes the example that wet streets cause rain.
Graph Paper Diaries has an excellent description of correlation/causation confusion along with six examples:
Thing A caused Thing B (causality)
Thing B caused Thing A (reversed causality)
Thing A causes Thing B which then makes Thing A worse (bidirectional causality)
Thing A causes Thing X causes Thing Y which ends up causing Thing B (indirect causality)
Some other Thing C is causing both A and B (common cause)
It’s due to chance (spurious or coincidental)
Data dredging: Provided enough data, it is possible to find things that correlate even when they shouldn’t. The world of big data and big correlations. Statistical significance increases as sample size increases. Every one of the correlations in Tyler Vigen’s book was discovered by a computer.
In terms of methodology, Vigen used Pearson’s correlation coefficient which is common for expressing linear relationships between variables. You can double-check any statistic in the book, and find many more charts: http://tylervigen.com/sources.
Spurious Correlations documents a statistically significant correlation between many humorous variables, including:
Earnings per share of Domino’s Pizza Group and Economic loss due to cybercrime (98%).
Undergraduate enrollment at U.S. universities and Injuries related to falling TVs (99.6%).
Customer satisfaction with Taco Bell and International oil production (79.9%).
Stay-at-home-dads and Walt Disney Company revenue (93.8%).
Beef consumption and deaths caused by lightning (87%).
Real stories of correlations ≠ causations
Mark Twain: “The difference between reality and fiction is that fiction has to make sense.”
The Economist had an article “Enough is never enough,” that discussed whether or not advertising is good or bad? A survey of 1 million Europeans who self-reported life satisfaction with variation in total advertising spending as a share of GDP and found a significant inverse relationship: a doubling of ad spending = 3% drop in life-satisfaction. So, North Korea should be most satisfied on the planet!??!
Adding one more woman in senior management or to a company’s board, raises its return on assets by 8-13 basis points (hundreds of a percentage point), according to one study. An IMF study shows a higher share of women on bank boards is associated with greater financial resilience, and greater financial stability.
The correlation between unemployment and inflation was postulated in 1958 by William Phillips, a London School of Economics professor. Milton Friedman falsified this theory but it still rules at the Federal Reserve Bank.
Discussing brushfires in Australia, The Arsonist by Chloe Hooper reports [from a book review in The Economist, “Into the inferno,” June 15, 2019.
“People are more inclined to destruction in places where “high youth unemployment, child abuse and neglect, intergenerational welfare dependency and poor public transportation meet the margins of the bush.”
Based on this book, Risk Savvy: How to Make Good Decisions, (Gerd Gigerenzer, 2014), the author believes we live in a risk-illiterate society. There is some truth to that. When meteorologists report a 30% chance of rain tomorrow, what does that mean?
Some think, it will rain 30% of the time tomorrow
Others, it will rain in 30% of the region
Others still, three out of ten meteorologists think it will rain and 7 don’t.
What it actually means: That it will rain on 30 percent of the days for which this announcement is made.
Many of us smile at fortune-tellers, but when they’re armed with computer algorithms rather than tarot cards, we take their predictions seriously and pay for them. One amazing fact Gigerenzer points out in his book is the 5-year survival rates of prescreening for cancer. This is achieved by screening at an earlier age (age 60, say, rather than 67 due to symptoms). Say the patient dies at 70 no matter what. In the first case, the 5-year survival = 100%; in the second case it is 0%.
Statisticians call this the lead time bias. The high survival rates don’t tell us if lives are saved. PSA screening detects both progressive and non-progressive cancers. It can’t distinguish between them. More men die with prostate cancer than from it. As such, Gigerenzer concludes that prostate cancer screening has no proven mortality reduction, only proven harm. It’s the same with mammography.
Gigerenzer First Law: The more the media report on a health risk, the smaller the danger for you.
In the book, The Cult of Statistical Significance: How the Standard Error Costs Us Jobs, Justice, and Lives (Stephen T. Ziliak and Deirdre N. McCloskey), Deirdre points out that statistical significance is not the same thing as a scientific finding. In fact, it can be misleading at best. Also,
“The mainstream in science, as any scientist will tell you, is often wrong. Otherwise, come to think of it, science would be complete.”
Science has stopped asking “How much is the effect? And What difference does the effect make? Fit isn’t the same as importance. W. Edwards Deming used to say: “Statistical significance provides no rational plan of action.”
The VeraSage Symposium
Use this link for more information: VeraSage DownUnder 2019
As you may or may not know, VeraSage meets biennially for a VeraSage Symposium [“to drink together,” and for the first time ever the Symposium this year is being held in Australia in November.
In conjunction with Australian Senior Fellow John Chisholm we have put together what we think is a terrific program covering two events—a one day “Transforming Your Firm” workshop in Melbourne on Tuesday, November 12, followed by the two-day VeraSage Symposium commencing Wednesday evening November 13 and running to the 15th, being held in Geelong about 70 miles from Melbourne on the wonderful Bellarine Peninsula.
Ron has visited Melbourne, Geelong, and the Bellarine Peninsula several times now and trust me they are wonderful places to visit and experience some of the best of what Australia has to offer.
In addition to the formal programs the Aussie contingent have put together some wonderful social events, with a special emphasis to showcase the local food and (of course) the wine. If you are at all interested in attending, register online or contact John Chisholm at john@chisconsult.com for further information, or any travel recommendations—especially if this is your first time visiting Australia.
You will see from the website that early bird pricing is available until end of August and attendance will be strictly limited. We look forward to seeing you Downunder in November!