Learning to speak data

Statistics is the new grammar, writes Clive Thompson in the May issue of Wired. The statistically illiterate can’t understand public policy debates, which increasingly come down to what the data mean. Is the economy improving? Do childhood vaccines increase the risk of autism? Is global warming for real?  Is the latest political poll reliable?

Statistics should now be a core part of general education. You shouldn’t finish high school without understanding it reasonably well — as well, say, as you can compose an essay.

Schools teach probability — red, blue and yellow marbles in a bag — and “a bit of basic data analysis,” responds Mr. D of I Want to Teach Forever. But math teachers often gloss over “problem solving, finding reasonable answers and determining what data is needed to solve a problem.”

Aside from problem solving skills, we don’t spend enough time on proportional thinking (everything from using percents to measurement and scale) and just plain number sense that everyone could use on a daily basis. What we’re left with is a nation of people who fear math, who run to a calculator for the most rudimentary problems.

Some people live the data-driven life, writes Gary Wolf in the New York Times Magazine.

About Joanne

Comments

  1. I wonder what was the old grammar? It certainly hasn’t been, you know, grammar for decades…

  2. Michael E. Lopez says:

    Apropos of Rob’s comment:

    “Statistics should now be a core part of general education. You shouldn’t finish high school without understanding it reasonably well — as well, say, as you can compose an essay.”

    God help us if they do statistics as well as they write essays.

  3. Rob, Michael:

    Unfortunately, some of my biology majors can do the statistics, but fall flat on their faces writing the paragraph that interprets the numbers they just calculated. God help us indeed.

  4. Richard Aubrey says:

    Stat, which is to say, probability theory, if well understood, can help one against political BSing.

  5. Mike Skiles says:

    Just like Willingham says, teaching content is teaching reading, so the same can be said for stats. You have to know economics, chemistry, history, to really understand the stats.

  6. Greifer says:

    Learning “statistics” independently of subject matter is pretty much worthless. What’s an outlier depends on your content and your methodology. Which data points are due to experimental error depends on subject knowledge.

    But what is this “statistics” subject? Do people mean probability theory? combinatorics? bayes law? hypothesis testing? 1 sided and 2 sided tests? Usually, they mean some bizarre sampling from the above that isn’t a coherent subject.

    Given that the math community doesn’t even have one unified definition of probability, it’s a ludicrous idea to think we’re going to universally teach it to high schoolers. Learning the intricate details of t tests isn’t learning how to think about what question is being asked.

  7. Tracy W says:

    Greifer, as someone who learnt statistics at high school as its own subject, and then twice over again as part of communications engineering and econometrics, I think that studying statistics on its own is worthwhile. For a start, knowing the calculations side frees up brain power to think about the particular applications.

    As for your criticism of “what is this statistics subject?”, this could be applied to *any* subject studied at school. What is history? Do people mean what happened, why it happened, what could have happened, moral arguments, interpreting secondary sources? Given that the history community doesn’t even have one unified definition of the causes of WWII, is it really a ludicruous idea to think that we could teach the history of WWII univerally to high schoolers?
    Ditto for physics – is physics Newton’s laws? Einsten’s theory of relativity? The internal structure of an atom? Material science? Geometry? Calculus? How do you teach the concept of energy to high school students?

    Learning the intricate details of t tests isn’t learning how to think about what question is being asked. But that doesn’t mean it’s not worth learning at all, such information can give you a handle on whether a person proposing the answer knows what they’re talking about. I wouldn’t teach the intricate details of t-tests at high school level, but type I and type II errors sounds more plausible.

  8. Well, the Wiki page on statistics is pretty clear as to what the subject matter is all about. I worked with a programmer once who invented his own elaborate scheme because he had never heard of standard deviation.

    Simple statistics certainly should be taught in high school, it’s essential to understanding the modern (or ancient, for that matter) world. As Robert Heinlein said, “if you can’t apply numbers to it, it’s not science.” (quoting from memory here, give me a break)