How We Can Interpret Data Better
My favorite read last year was Factfulness by Hans Rosling. It provides insight into how to approach data - particularly data related to hot button issues. I wrote about it here. A lot of the book can be boiled down to “ask the next question. “
For example, the other day I saw this headline - “Pasta may be good for your heart health.” I love pasta as much as the next person but I was skeptical.
Using Rosling’s blueprint, the next question you might ask is “say’s who?” Well, it turns out this study was conducted by Italian scientists. It might be that this was already sounding eerily similar to an argument my wife of Italian lineage might make, but I was just about ready to discredit it right there.
But I gave them the benefit of the doubt and asked the next question. “Who was this study done on.” Since stress testing heart attacks on humans is not legal, the study was actually conducted on mice!
Okay, this is getting ridiculous but I’ll ask the next question. “What exactly did these mice eat?” - It certainly wasn’t some al dente rotini. The pasta was constructed in a lab with supercharged Barley added in for good measure. Not exactly something you could pick up at the supermarket down the street.
Sorry - I’m not sure your pasta habit is doing you any favors.
Below are a few more fictitious examples –
Headline: “Breaking – Dow Jones Drops 200 points!”
Questions: How many points are there? There are over 35,000 so 200 ÷ 35,000 is roughly one-half of one percent! How often does that happen? Over the past 10 years, it’s occurred over 500 times, roughly once a week. Why again is this on the news?
----------------------------
Headline: “Market Guru Who Predicted 2008 Crash Predicts 2021 Crash”
Questions: Has he made any other predictions before or since 2008? Was he correct? Turns out he has been predicting a crash just about every year since 2011 and has been wrong. Are there other market gurus who predicted the 2008 crash that aren’t predicting a 2021 crash? Yes, the gurus are essentially split on their 2021 prediction, which they happen to be just about every year.
----------------------------
Headline: “Retail sales slump in July – is another economic collapse coming?”
Questions: How much did it slow by? By about 1% from the prior year. Were the months before July faster than normal? Yes, by about 5% from the prior year. They could have just as easily said “Retail sales are steadily higher from the year prior” instead, but that headline doesn’t sell.
I do not want to make light of the serious issues we face today, but rather help us all ignore media outlets broadcasting data without context simply to invoke fear. There are many economic and social issues that are serious and demand our attention. But for every one of those, there are 100 headlines to ignore.
Thank you for reading,
Alex
This blog post is not advice. Please read disclaimers.