Making Sense of Poll Data from Massachusetts' Senate Race

commentary on data from Public Policy Polling ^ | 1-10-10 | Dangus

[dangus]

Washington was rather shocked by a Public Policy Polling (PPP) which showed Republican Scott Brown with a one point lead over Democrat Martha Coakley in the race for Ted Kennedy's former senate seat. If the poll is accurate, and a subsequent poll by the Boston Herald suggests it is, the internals of the polls ought to send every elected Democrat office holder screaming in terror.

PPP shows Obama's approval rating in Massachusetts, one of the most Democrat-dominated states in the union at 44%. That would seem to suggest a 50-state sweep in 2012. But their home page is more in line with the sort of polls coming out of the mainstream press: Obama at 52% nationally. Could Obama's approval rating in Massachusetts really be eight percentage points lower than in the rest of the country? Or are polling firms only giving us honest results when there is an election about to demonstrate the electorate's true leanings?

The PPP press release implies that one reason for their surprising find is that they are surveying likely voters, and that Brown's supporters are more likely to be counted that Coakley's. But their partisan breakdown doesn't bear that out. According to their likely voter model, 44% of those who turn out will be Democrats, whereas only 37% of voters in Massachusetts are registered as Democrats. And their poll asked registration, not how they voted; fully 39% indicated that they were Independents. Only 17% said they were Republicans.

PPP points out that in their sample, Obama voters outnumbered McCain voters by 16%, as opposed to the 26% that Obama won by. But 7% couldn't remember... or didn' want to remember. Usually most of those respondents are those who regret their decision.

Demographics internals don't make it seem any more likely that PPP's "likely voters" are more Republican than the general population, either. 40-65 year-olds seem very heavily represented in PPP's results, but Coakley actually won that group. And blacks, made up 9% of PPP's voters, versus just 7% of Massachusetts residents, but 20% of Democrats.

The only way PPP's polls make any sense (without any duplicity) is if Obama for whatever reason is vastly more unpopular wherever elections are held, like Massachusetts, New Jersey, Virginia, and Pennsylvania. This is possible... Perhaps as campaigns focus voters onto the issues, Democrats simply lack credibility. That doesn't augur well for 2010, when most statehouses, all house seats and 37 senate seats are up for elections.

Polling data is available at http://www.publicpolicypolling.com/pdf/PPP_Release_MA_45398436.pdf

[okie01]

Could Obama's approval rating in Massachusetts really be eight percentage points lower than in the rest of the country?

That is a very, very good question.

[STARWISE]

Scott and Gail Brown, daughters: Ayla and Arianna, both college students (what a lovely family).

Thanks to Daisyjane69, I found out Texas Darlin DID reopen her blog .. just for THIS race.

"After a 5-month hiatus, I have temporarily re-activated the blog for one reason: to be part of the nationwide grassroots revolution to elect Scott Brown to the U.S. Senate on Jan. 19. His win would end the 60-seat fillibuster-proof majority currently being abused by Democrats. His vote could kill ObamaScare."

[DontTreadOnMe2009]

the internals of the polls ought to send every elected Democrat office holder screaming in terror...

Man, I hope they scream in terror

and scream

and scream

and keep on screaming.