Bad Data Examples Appeal

By San | June 19, 2007

Back in early April, I canvassed the IASSIST list for examples of articles, studies, or any publication that misuses data.  My original message:

“Do you keep news stories that mis-state, misinterpret or otherwise misuse numeric data?  I have a faculty member who would like to use such examples to demonstrate to her students the importance of numeracy in real life.  I have a few examples but suspect others of you may stash this kind of thing in a “Can you believe this?!” kind of folder.  If you would be willing to share citations with me, I would very much appreciate it.  This could be opening a floodgate, I realize, so let’s say English-language stories that have appeared in the last year.

Thanks!”

In fact, it wasn’t a floodgate and I’m hoping to collect more examples, but below is a summary of the responses I received.

=========================================

Book:  “The Obesity Myth” by Paul Campos might be worth looking at. The first few chapters are all about how questionable studies with relatively small samples have been used to shape health care policy regarding obesity, while larger studies whose methods are less controversial are ignored.

Sources for Such Stories

  1. One very nice resource for this sort of thing is the Statistical Assessment Service at George Mason University (STATS) at: http://www.stats.org/          (2 people suggested)

  2. (the following two are more about visualizations of data) http://junkcharts.typepad.com/junk_charts/      (2 people suggested) http://statisticalgraphics.blog.com/

  3. Many sessions at the annual conference of the American Association for Public Opinion Research (AAPOR) have touched on this topic over the years.  You might take a look at past conference programs for sessions dealing with “The Media and Public Opinion Research” (or similar titles) to identify specific panelists who could provide numerous examples.  The AAPOR web site is http://aapor.org/.

=========================================

The trouble is, these examples/sources mostly point out why they’re bad examples.  The faculty member I’m working with would like to have her students read the items (so, preferably articles) and then identify for themselves why each is a bad use of data.  Obviously, there could be a wide range of “badness” and what level of student would recognize the problem(s), but she is working with first-year undergraduates so things for that audience are of greatest interest.  I think it would be great to keep a bibliography of such examples.  Or is this a libelous venture?  Am I better off culling citations for bad examples from these sources?

Posted on behalf of Michele Hayslett