Skip to main content

View Diary: How to win re-election on a job approval of under 50 percent (171 comments)

Comment Preferences

  •  Cherry picking (0+ / 0-)

    I don't have muchnto add to the excellent post Kraushaar is clearly being contrarian and or determined to claim that there will be an interesting horse race.  The only point to add is that his argument about Clinton's 92 fav/unfav is based on outrageous cherry picking.  He notes one and only one poll.  A serious approach would be to average available polls.  This isn't trivially easy given the mass of more recent Clinton data (that is pollingreport only gives Clinton fav/unfav back to 1997) but even just Gallup  by Kraushaar has other relevant numbers.  

    Kraushaar's claim is, in fact technically false.  He cites the Gallup poll taken April 9-12 1992.  He doesntp't give the full cite (no "Gallup" no 9-12) but the numbers match.  To be "April" not "April 9-12" Kraushaar would have to average in the April 20-22 42-49 fav/unfav result.  But looking at more than one single solitary poll would have reminded readers that even just sampling error in one poll is important and also that views of newcomer Clinton were fluid in 1992.

    The poll closest to exactly 16 years ago is March 26-29 with 38 fav 39 unfav.  Noting that poll would totally undermine Kraushaar's attempt to mislead with a technically true statement.  

    Failing to makr the required qualifications Including "9-12" Kraushaar failed to make his misleading statement technically true.  I think the proper level of toleration when judging if cherries have been successfully picked is zero.  The intent is dishonest - a minor slip which would be ok if there wasn't the clear intent to deceive makes it a lie.

    More generally, I see no reason to continue to tolerate cherry picking of polling data (which was standard journalistic practice until recently). Kraushaar selected one winning candidate then quoted only the one single solitary Gallup poll which most supported his case.  He suppressed the necessary qualifiers which would have made it clear how carefully he selected the data (the really key one as you note is that Clinton was new to the scene - the oldest Gallup poll is a 21/10 fav unfav "Bill who?" poll from January 3-6
     2012.

    Your discussion of what polls tell us about Navember is perfect, but you are very kind to Kraushaar.  His approch is still standard practice, but it is dishonest.  I think it is time to denounce the cherry picking of polls.

Subscribe or Donate to support Daily Kos.

  • Recommended (128)
  • Community (60)
  • Media (32)
  • Elections (32)
  • Trans-Pacific Partnership (30)
  • 2016 (29)
  • Law (28)
  • Environment (28)
  • Civil Rights (26)
  • Culture (24)
  • Barack Obama (24)
  • Hillary Clinton (23)
  • Republicans (22)
  • Science (21)
  • Climate Change (21)
  • Labor (19)
  • Economy (19)
  • Josh Duggar (18)
  • Jeb Bush (18)
  • Bernie Sanders (17)
  • Click here for the mobile view of the site