Thursday, October 25, 2012

Don’t ignore poll “house effects”


House effects refer to the internal dynamics of how polls are constructed and conducted by major pollsters that can cause polls from a particular organization to favor one candidate over another.  House effects aren’t necessarily intentional.  Most legitimate polling organizations depend for their existence on getting their poll numbers right.  Never-the-less, some polls tend to favor Democrats and others tend to favor Republicans, if ever so slightly.  That can make a big difference in the results.  When you see poll results you need to take into account the House Effects.  Here is a chart that will help you.  Print it and keep it with you.  If you can’t see the chart, go here:  http://www.huffingtonpost.com/simon-jackman/house-effects-by-back-by-_b_2007907.html

As you can see from the chart, Zogby Internet polls tend to favor Obama by just a little less than 2 percent.  Gallup polls tend to favor Romney by a similar percent.  So, you should subtract 2 percent from Obama’s rating in Zogby polls and add 2% to Obama’s rating in Gallup polls.

No comments: