When Statistics Mean Something

I just got back from a lecture featuring Stephen Dubner (coauthor of Freakonomics and Superfreakonomics). If you’re not familiar with those books, they try to strip away how we ‘feel’ about a particular topic (for example, teachers cheating on standardized test scores, or the hand hygiene of doctors in hospitals) and they address those sorts of topics “with neither fear nor favor, letting numbers speak the truth.” When we think of economics, we usually think of budgets, currency, the stock markets, etc., but what Dubner and his coauthor Levitt do is look at what some call ‘behavioral economics.’

A few things stood out in his talk. One was the reminder about how we are much more able to perceive traits (good and bad) in others than we are at seeing them in ourselves. Another is how hard it is to change human behavior. Finally, when collecting data, how you collect it is really important. Self-reported data, according to Dubner is usually pretty useless (especially if you ask people on a survey to identify themselves – even as a group). He gave an example of a headline that went something like this: “Recent survey shows that favor of nuclear power has declined.” Hmmm, a survey taken right after a tsunami destroying a nuclear reactor. Dubner mentioned how these surveys/polls are everywhere and those are not the kinds of statistics he is attracted to.  He also warned everyone about using incentives to try to change behavior. They have a tendency to backfire. Dubner also made a case for thinking outside the box. “Be a heretic,” he said, “but remember that most were wrong, many were executed, however, those who were right and lived, changed the world.” In terms of education, this talk reminded me of the importance of keeping our rigorous curriculum balanced between learning basic skills and fostering natural curiosity and creativity. For me, It was interesting to compare his talk with that of David Brooks who I saw last week when he was in town promoting his book. I was fascinated by many of the similarities.

Speaking of numbers, if you haven’t visited the site Gapminder by Swedish Statitician Hans Rosling, you really ought to. Can numbers be fascinating? They certainly can, and he does an incredible job on his interactive website which visualizes the data on world development. He’s done numerous TED talks advocating that one of the ways to stave off world population growth is to create wealth in all nations. His latest talk “The Magical Washing Machine” is only about 6 minutes long and well worth watching till the end when he makes his point. It makes me want to take more statistics courses.



One thought on “When Statistics Mean Something

  1. “Recent survey shows that favor of nuclear power has declined”

    The news (http://www.themalaysianinsider.com/world/article/iaea-suggests-japan-widen-exclusion-zone-near-nuclear-plant/#IDComment138789273) today quotes

    “The Japanese government says nuclear power will remain an integral supplier of power. Before the disaster, Japan’s nuclear reactors provided about 30 per cent of the country’s electric power. That had been expected to rise to 50 per cent by 2030, among the highest in the world.”

    Perhaps the people are afraid but the Japanese seems determine to continue. There might be a backlash because, if people’s statistic says no, then no nuclear unless there is a security that Fukushima’s episode won’t repeat. As humanitarian calamity is bad enough to accept, the repeat of nuclear leakage would a speechless disaster.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s