Below is an analysis by Pollster.com that lays out the “House Effect” — more pejoratively known as “polling bias” — of the many survey organizations that report Presidential Approval Ratings.
Pollsters at the top tend to be more favorable to Pres. Obama (some very favorable); those at the bottom tend to be less favorable.
[See Ken’s Take and an important UPDATE at the bottom of this post]
http://www.pollster.com/blogs/why_is_rasmussen_so_different.php
* * * * *
Ken’s Take:
(1) Note that the mainstream media outlets (ABC, WP, CNN, CBS, NYT, AP) hold the top slots. Hmmm.
(2) Note that FOX is smack dab on the median. Pollster emphasizes that the line that corresponds with the zero value is NOT a measure of “truth” or an indicator of accuracy. It’s simply a “normalizing” measure — in effect, a median value. TakeAway: sure seems fair & balanced.
(3) Quant jocks generally attribute the differences to methodology or samples. For example, Rasmussen is an automated phone survey — and people tend to be less hesitant with negative responses when dealing with an automaton than they are when answering to a humanoid.
Regarding samples, less favorable surveys tend to sample likely voters, not all adults. Critics argue that minorities and young adults are under-represented when the cut is likely to vote.
(4) A bigger deal, in my opinion, is that sampling tries to get a representative number of Dems and GOPs. My bet: surveys at the top over-sample Dems and the ones at the bottom over-sample GOPs.
(5) Regardless of the specific poll, the conclusion: country is divided down the middle … plus or minus some random noise.
* * * * *
UPDATE
CNN — #2 in favorable leaning to the President — released new poll results on Fri. Dec. 4.
During November, President Obama’s approval dropped from 55% to 48%.
* * * * *
Question: Do you approve or disapprove of the way Barack Obama is handling his job as president?
For the full CNN survey results:
http://i2.cdn.turner.com/cnn/2009/images/12/04/rel18a.pdf
The Homa Files 
Leave a comment