The Complete Idiot's Guide To Reading Political Polls

Two pollsters and one reporter demystify polling.

Slideshow
<< Previous
0 of 1 Images
Next >>
Jennifer Dineen
Jennifer Dineen said it's important to note how much information the pollster is willing to disclose about poll methodology. Photo:Chion Wolf
John Dankosky
WNPR News Director John Dankosky sports a fresh band aid while recovering from an incident with a sharp can lid. Photo:Chion Wolf
The Complete Idiot's Guide To Reading Political Polls
Download Audio
Audio Playlist
The Complete Idiot's Guide To Reading Political Polls

A poll conducted in 1997 showed Congresswoman Barbara Kennelly leading incumbent John Rowland by four points. 

On the strength of this, Kennelly was persuaded by party leaders to give up her seat and take on the Governor. Other possible challengers were persaded to move aside. Kennelly embarked on a race that had, to use the New York Times's phrase, "a devastating trajectory."
 
She lost by a 63 to 35 ratio, a terrible, terrible blowout for a popular and accomplished politician. The lesson?  We use polls not only to predict elections but to make plans,  and polls are imperfect. Nobody really knows what happened in that 1997 poll. One theory was that Kennelly's campaign had the curious effect of draining support from her. Or maybe the poll was wrong. 
 
Leave your comments below, e-mail colin@wnpr.org or Tweet us @wnprcolin.

  

Comments

E-mail from Joan

What about polls that have a clear political bent? For example, you'll answer questions along the line of who do you plan on voting for, how likely are you to vote, and then it will veer into "questions" that make the same claims that are in negative ads.

Seems shady.

E-mail from Jim

My approach to pollsters is to lie. Does that help screw things up?

E-mail from Sheri

How is the margin of error determined for each poll?

E-mail from Peter

I would be interested in hearing comments about the rasmussen poll.