At first glance, a new Washington Post-ABC News poll appears to give President Barack Obama cause to celebrate. A random national sample of 1, 103 adults found Obama would beat former Gov. Mitt Romney (R-Mass.) 51 to 43 percent in the race for the White House.
However, Rasmussen has Obama and Romney tied at 45 percent each, based on a poll of likely voters taken the same week the Post-ABC poll was conducted.
In 2008 polls often led news cycles for days, raising questions about whether candidates gain an advantage in the court of public opinion simply by pushing favorable poll results.
Considering the abundance of polling companies, what polls can a voter really trust? Furthermore, what should a person consider when hearing results of a poll?
The Post-ABC poll suggests it’s a good idea to take a look at the crosstabs. What percentages of the total sample do Democrats, Independents and Republicans make up? The Washington Examiner picked the Post-ABC poll apart and put things “in context”:
“34 percent of those polled identified themselves as Democrats, 23 percent identified themselves as Republicans, and 34 percent identified themselves as Independents. That works out to an 11 point advantage for the Democrats.”
The Post did the same thing in 2008, less than one month before the General Election, with a poll that found Obama 10 points ahead of Arizona Sen. John McCain (R). In that poll of likely voters, 39 percent leaned Democrat, 30 percent leaned Republican, 27 leaned Independent and 3 percent leaned “other.”
Obama ended up beating McCain by 7 points.
In 2008 there was even a poll meltdown of sorts. Blogger Markos Moulitsas, an influential Democrat who founded the Daily Kos, announced that polls conducted by a company he hired were “likely bunk.” Numerous conservatives had questioned the polls repeatedly.
The website Real Clear Politics features an updated average of major polls for the 2012 election. There’s a variation in the size of Obama’s lead averaged among six polls, and the margin from the Post-ABC poll gave Obama’s numbers a bump.
Polls created so much debate among pundits in 2008 that the Department of Political Science at Fordham University compiled a report on accuracy. The report found Rasmussen and the Pew Research Center to be the most accurate when poll results were compared to the actual election results.
The Post-ABC polls were ranked in 14th place—7 spots below the polls Moulitsas claimed were “bunk.”
Of 23 organizations rated, the Fordham report also found 4 polls overestimated McCain’s strength while a whopping 17 overestimated Obama’s strength.
Do inflated poll results help sway undecided voters?
Considering a 10-point spread between the polls conducted by Rasmussen and The Post-ABC, it may be a good idea for pundits and journalists to put polls in the proper context by giving the specifics along with the hype.