iSixSigma

Numb3rs

You don’t have to read much of the daily paper, in the US at least, to see datapresented in very interesting ways.

“Gas price SKYROCKETS to $4 a gallon!”

“The Dow Jones Industrial Average PLUMMETS to 12,500!”

“Pistons [Basketball Team] Have the EDGE Now!”

“Kid Obesity Rate STEADY”

Now, part of the reason for this hyperbole isthat exciting headlines get more people to buy the paper, andso you may think thattheexaggeration is just a way to get people to read the accompanying story.

But when you look more closely, thegas price moved from$3.97 the week before; the DJIA had been 12,600on the previous day, the Pistons were tied 2-2with the Celtics, and buried in the paragraph about the kids was this statement: “…it’s too soon to know if this really means we’re beginning to make meaningful inroads… it may simply be a statistical fluke.”

Well, that putsa little different spin on the headlines. I worry about this for two main reasons. First, we are all at the mercy of first impressions, and while newspapers need to sell, they sometimes do it by presenting data in a way that is easy to get alarmed over, but not easy to understand (as we project engineers would understand it). Now, no one expects to see or hear detailed information on how the data was collected, or how thesample size was calculated.Buthow many people read the full story in depth? At least, we should train ourselves (and our kids) to realize when data is being presented as a teaser for the story. As I put it in my Lean Six Sigma class, “What questions should you ask about how this data was collected?”

Handpicked Content:   Hands of the Customer

The other reason that I worry is that the math that my kids were taught, in their suburban-Detroit high school, had very little to do with real life; they could figure cosines andvectors and the slope of a line, but not how to figure whether a drop of 12,600 to 12,500 was cataclysmic. I for one would eliminate geometry in favor of astatistics class – including statistical process control, presented with real-life scenarios. Then readers and viewers and listeners could have an idea about whether data was being presented in a rational way by the news media.

What do you think? Is data presented in the news in an ALARMING fashion???

Comments 6

  1. Gary

    My favorite of this decade is "Bush sees victory as a mandate"

    I would not eliminate geometry, but I would make basic stats mandatory. I made it mandatory for my two kids when they went through high school – they had the opportunity to take an AP Stats course, not something every kid in the Michigan, underfunded school systems, get to do.

    0
  2. Stephen C. Crate

    I remember from my graduate statististics classes and work as an analyst that you can present data in many ways depending on what your goal is. I am sure that fact carrys some truth when it comes to the media. The problem is that the public believes what ever presentation is given as "the truth", when in fact there can be many truths from the same set of data. There is a great book I still have from my graduate work by Edward R. Tufte called The Visual Display of Quantitative Information. The author talks about Graphic Integrity which speaks to your point. Can we hold the media to those standards? I think we should.

    0
  3. Cathy Johnson

    I am trying to contact Ms. Kozlowski. Sue, please contact me at cjohnson@cola.org about a speaking engagement at the Symposium for Clinical Laboratories: An Interactive Experience in Quality.

    Cathy Johnson
    800-981-9883, ext 568

    0
  4. AMD

    Another good one I remember from the 2004 election was Bush supporters saying that in the last election more people voted for President Bush than in any election in the history of the United States of America. Of course, that was a reflection on the unprecedented large turnout in that election, not a reflection on President Bush’s "mandate". After all, John Kerry had the second most number of votes in all of American History.

    0
  5. Ian Furst

    I notice your blog post is labelled Numb3rs not "a deeper look at statistical variation in quality measurements". If there is no sizzle no one will eat the steak and you know it Sue. I think most people are suffiently jaded to the marketing aspects of mass media but lacking the statistical training to understand the basis of numb3rs. During undergrad training we all had to read "A User’s guide to the medical literature". It was an easy to read guide to understanding clinical data with statistical analysis. I’m not sure if six sigma has an equivalent.
    http://www.waittimes.blogspot.com

    0
  6. Sue Kozlowski

    Thanks for your comments, Ian. For anyone trained in medicine or research, I would agree with you. But think about everyone in the US – and estimate the percentage that have had even ONE statistics class. How many people watching the "Nightly News" believe it when the announcer says, "Good news, folks, breast cancer deaths are on a downward trend" while the data shows that the numbers were down last year after increasing six years in in row?

    You may want to take a look at this e-newletter that I just received today, with a similar theme… coincidence? correlation? causality? Hmmmm….
    –Sue K.

    Crucial Skills, Vol. 6, Issue 25

    0

Leave a Reply