Coakley And Brown In Statistical Tie

By Fred Thys
Jan 5, 2010, 1:59 PM
24 Comments

Correction appended — With just two weeks until the U.S. Senate special election, a new poll finds Republican state Sen. Scott Brown and Democrat Attorney General Martha Coakley in a statistical tie.

Coakley leads Brown 50 percent to 41 percent. The survey by Rasmussen Reports of 500 voters who say they are likely to cast a ballot on Jan. 19 has a margin of error of plus or minus 4.5 percent.

“Special elections typically feature low turnout,” the company says on its Web site. “That’s one reason the race appears to be a bit closer than might typically be expected for a Senate race in Massachusetts. Kennedy carried 69% of the vote when he was re-elected in 2006.”

One percent of those polled say they would vote for “some other candidate.” Joseph Kennedy, who is running as an independent and is of no relation to the late senator, is the only other candidate on the ballot.

Seven percent said they were not sure who they would vote for.

Correction: Coakley and Brown are not in a statistical tie.

Please follow our community rules when engaging in comment discussion on wbur.org.

Hilary

Republicans don’t think we’re worth the money. They don’t think we’re worth the cost of health-care or education or civilization. They don’t think we’re worth giving a s#*t about. Why would anyone in their right mind vote for those guys?
Luke

The dems burdened Americans with mandate to buy insurance without providing enough subsidy to make it economical then they slam a tax if you don’t buy it and call it fee. They squander taxpayers money to bail out wall street and the auto industry and miserably fail in creating jobs… who in his right frame of mind would give them a vote????
http://aol Christopher Rinaldi

This state goes against and Liberal Democrats go against everything our country was founded on. The Democrats including Coakley are reducing our freedoms and rights. They want government to control every aspect of our life, no thanks. Millions of Americans have died so I can be free, even though I live in one of the three most restrictive states on personal freedom.

It is proven that firearm safe storage laws are causing thousands of Americans their life. These laws have not reduced accidents or suicide but they are responsible for many, many deaths. Read John Lott’s book The Bias Against Guns for the facts/truth.

I have met Senator Scott Brown and he is one of the most sincere caring people I have ever met. He unlike the Democrats in this state listens to his constituents. Help turn this country around by voting in someone who is a true Patriot, strong leader and someone who actually has experience. Please vote for Scott Brown
BILL

HILLARY ARE YOU SERIOUS – OR JUST UNIFORMED- THE GOP HAS BEEN CHAMPIONING THE IDEA OF REAL HEALTH REFORM INCLUDING TORT REFORM – BUT THE DONKEYS DON’T WANT IT BECAUSE TRIAL LAWYERS SUPPORT THE DNC…H DEAN JUST SAY IT ANINT SO- YES HE DID- Rep’s believe that people need to be able to afford there – own insurance – and by going the way of the Dem’s Everyone will have ins. even those who don’t need it- those who can’t afford and the working middle class will bear the burden of paying for there own- plus the cost of those who DON’T WORK- CAN’T WORK- DON’T CARE TO WORK – yeah thats the SPREAD THE WEALTH MANTRA THAT CALIFORNIA HAD – AND LOOK WHERE THEY ARE – BROKE- WRITING IOU’S and that will be how the Nation will look- torn down- broke- oh- but everyone will have Insurance And we get to pay for it for 4 flippin years before it even becomes a Force – Nice- now if that isn’t UTOPIA I DON’T KOW WHAT IS- what a flippin tub of PORK FAT- THE MONEY THEY ARE BUYING THE VOTES WITH …….SHOULD HAVE BEEN PUT IN A “POOL” FOR THE UN AND UNDERINSURED- AND THEY WOULD HAVE PLENTY TO INSURE THE 20-30 MIL – LEGAL OR ILLEGAL PEOPLE WHO NEED HELP- AND THIS FIASCO COULD BE OVER ..Instead they are slammin this Takeover – Euro style Failure Bound Healthcare takeover to Doom America’s Mayo Center HAS LOUDLY SAID THIS IS NO GOOD- AND NOW- ARE NOT TAKING MEDICARE- BECAUSE THEY SAID THE PAYMENTS ARE TOO LOW- =WHAT DOES THAT TELL YOU Better think about what that is you wish for- Unions Taking over our Healthcare ?? Not for me ! the PREZ. clearly works for SEIU- AND NOT THE AMERICAN CITIZENS
Jesse

Hilary is a great example of how most of the Commonwealth have fallen victim to the indoctrination by the left that begins early on in our school system. Hilary, most politicians, regardless of political affiliation, do not care about us. Most people are only concerned with their own; you see this everyday. But to target one political party as being ‘evil’ is very naive on your behalf. Both Democrats and Republicans have a positive message based on a different ideology; this is good, we need to have checks and balances so things remain fair for all. Right now, things are not fair for all. The politics in Washington right now are FAR left leaning and that is not good for most of those in our country, particularly the tax payers. Hopefully, with the upcoming elections we get some new blood that can hopefully reverse most of the damage that these current politicians are doing to our country. I work in industry and visit many manufacturers in New England and the general mood I’m seeing is that most people are not happy with the politics in Washington now.
Tom

The GOP had 8 years to implement healthcare reform. What did we get? Zilch! Meanwhile, the insurance companies continued to gouge the public with obscene rate increases.
krisd

Just a simple question – the title to this article references a “Statistical Tie” when Coakley is reported to be in the lead by 9%. Even if you use the standard that the lead has to be smaller than twice the margin of error, in this case 4.5% with a 95% level of confidence, that should still leave you statistically a bit shy of a tie. I’m not a mathematician or a pollster, just curious why this is the only article that I’ve seen calling it a tie.
http://www.wbur.org Fred Thys

This is WBUR political reporter Fred Thys. Coakley and Brown are in a statistical tie because Coakley only leads by 9 percent, and the margin of error is 4.5 percent. This means that instead of being ahead 50 percent to 41 percent, Coakley and Brown could be tied at 45.5 percent.
David

I’m very torn. I’m a Dem and I should favor Coakley. But she’s been so bad on issues of illegal immigration–which is a major environmental issue since mass immigration is projected by the Pew Center (2008) to add 110 million to the population by 2050) as well as a major labor issue (google “compassion that hurts” or go here for NYT columnist Nicholas Kristof’s excellent analysis select.nytimes.com/2006/04/09/opinion/09kristof.html/) that I am tempted to vote for Brown. I am disgusted with the Democratic establishment for supporting amnesty, and for failing to stop the granting of 125,000 work visas every month during this period of high unemployment. Mass immigration is welfare for big business, which is why the US Chamber of Commerce and big business republicans such as McCain, Arlen Specter (yeah, I know he switched parties), et al. are so much in support of amnesty and mass immigration. Mass immigration is why meat packers today earn $8/hr in horrible conditions instead of the $25 (inflation adjusted) they earned a generation ago, and why Walmart can pay minimum wage for grocery checkers in S. CA who used to make around $17/hr.
Leo Bloom

Whatever the mathematical reasoning, the headline remains deeply misleading. Statistically speaking, Coakley is as likely to have an 18 point lead as to be tied with Brown, according to the spread and the margin or error. But of course, “Coakley has near-20-point Lead” probably wouldn’t get a lot of clicks. I would expect misleading writing like this from the Boston Herald, not from WBUR. Very disappointing.
Jonathan

I’m sorry but your headline simply isn’t true. A 9% difference with a 4.5% margin of error at a 95% confidence level means that 2.5% of the time Brown would win while 97.5% of the time Coakley would win, and furthermore that the same 2.5% of the time Coakley would win by more than 18%. Furthermore these polls typically assume nice bell shaped distributions meaning that a 9% lead is probably within a few percent of the true number the majority of the time.

If for instance there was a 4% difference between the numbers and a 4.5% margin of error, the error would be greater than the difference and so statistically the leader in the poll would have more than a 5% chance of not winning, though still a much better than even.
A statistical tie would be if both candidates polled evenly. In any poll there is always a statistically non-zero chance that in fact the two things being polled are exactly even.

In this case there is definitely not a statistical tie.
Of course all this assumes that you have confidence in Rasmussen’s methodology. Nate Silver has an interesting discussion on the details at 538.com.

The web has a lot of very accessible discussions on statistics that have many more examples on the meaning of confidence intervals and margins of error without going into the admittedly boring details of statistics.
Kenny

There is a sliver or hope. Now that Margaret Coakley has come out of hiding her arrogance and extremely leftist leanings are proving to be problematic. I smell an upset.
KM Brown

Dear Christopher Rinaldi–
This state embodies the very best of American values. It values education, tolerance (a christian value, supposedly right?), it cares for its poor and it values life b/c it provides access to healthcare for all. Further, unlike so-called “red states”, it has the lowest divorce rate, one of the highest education rates and its citizens are among the healthiest. Combine this with a vibrant advanced industrial economy (tech, healthcare education) where small businesses grow into market leaders–think anywhere from Sam Adams beer to EMC Corp–and I think there is a great deal of economic freedom as well as personal freedom. But, I don’t think your gun loving trumps my right to avoid being in the line of fire of one of your fellow gun-toting nutjobs ala Virginia Tech where the killer bought his weapons at a gun show where background checks aren’t required. Having my kids avoid a similar fate as those defenseless victims is a truer sense of personal freedom than your love of guns. If you dislike it here so much you are welcome to join your fellow chuckleheads in a dead-end, unhealthy and under-educated red-state. Don’t forget to take your guns with you and please enjoy all the fried catfish you can eat.
Boston Bozo

Fred,

It really is a shame you are so lazy. Citing the Rass poll the way you did shows you really aren’t actually interested in commenting on the race in a way that would inform WBUR listeners. This concept of a statistical tie you use as an interpretation is questionable at best and unhelpful in any circumstances. In fact, polls in special elections are highly speculative as a general matter due to high sensitivity to very small changes in the likely voter screen and, ultimately, the low turn out in such elections. In addition there is, as we all know, a lot of political noise out there so what the poll might actually show is high volatility in the electorate, more than Democrats might expect all other things being equal. You could have made more or less the same point, that Coakley could be vulnerable to Brown and explained something about it, without looking like someone trying to advance a narrative by choosing to use questionable statistical constructs as the story. I must say I am really disappointed in this kind of mediorce journalism.
tara

KM Brown, don’t break your arm patting yourself on the back. Getting taxed on every move you make in life does not create a “great deal of economic freedom”.

Have you missed all the criminals from the state house lately? This state is a political cesspool.
dh

Another words Boston Bozo, Jonathan and Leo Bloom you are a little nervous that Coakley might lose! Getting a little defensive aren’t we?
TheTownCrier

Brown seems to ‘get it’ on immigration, people. This isn’t a left or right issue.

http://www.progressivesforimmigrationreform.org/

State Sen. Scott Brown, R-Wrentham, today filed legislation that would require the state Attorney General to receive and document proof of citizenship or right to work in the United States for wage enforcement cases. Historically, the Attorney General’s office has adopted a policy of not asking about immigration status of employees when proceeding with a wage enforcement case against an employer. Under the bill filed by Brown, who is also a Republican candidate for U.S. Senate, wage enforcement cases would require valid proof of citizenship or immigration status before proceeding. “This is a responsible and effective immigration reform…

http://www.milforddailynews.com/homepage/x1767504421/Brown-wants-to-exclude-illegal-immigrants-from-wage-abuse-cases
Phil

GO TEAM COAKLEY. I will be out early and will bring friends!!!
MR

Hi, I’m a statistician. I work at one of the top schools in the area.

The title of this article is wrong. The margin of error is the 95% confidence interval. That means there is a 2.5% chance that Brown’s support is 45.5% or higher. (There is also a 2.5% chance that Brown’s support is 36.5% or lower.)

Similarly, there is just a 2.5% chance that Coakley’s support is at 45.5% or lower.

So, if we assume that the poll’s methodology is sound, and the margins of error for each candidate are independent of each other, then the chance of an upset or a tie is actually (1/40)*(1/40) = 0.0625% . That’s very low.

It is actually a stretch to assume that the poll’s methodology is sound. This poll was run by Rasmussen, a firm that consistently has a pro-Republican bias. For more information about why the assumptions behind this poll are not very sound, see Nate Silver’s superb polling stats and politics blog, fivethirtyeight.com (http://www.fivethirtyeight.com/2010/01/wicked-awesome-thoughts-on-massachsetts.html)

I will be using this article as an example of innumeracy in any future HS Statistics classes I teach.

No matter what any polls say, I’m going to be out stumping for Coakley on election day. A 0.0625% chance of a Brown victory is still too high! Get out the vote, Democrats!
ted

“It is actually a stretch to assume that the poll’s methodology is sound. This poll was run by Rasmussen, a firm that consistently has a pro-Republican bias.”

Give everyone a break will you?
After elections, there have been acknowledgments where Rasmussen polling was shown to be the most accurate. What are the poll numbers from the fivethirtyeight whom admits he votes democrat? It seems whenever democrats don’t like results there is always some conspiracy.

Just because you’re bias does not make everyone else bias.
Fred Thys

I interviewed Scott Rasmussen about his poll, and just posted his comments separately in this blog.
http://maryellenburke2005@comcast.net Mary

Hey you Coakley supporters -why not just hand her your wallet-save time! Give over your job to an illegal invader who will work for less! Yeah that makes a lot of sense-she will continue the stupid Kennedy policies that have endangered this country and brought us to the breaking point….Bankruptcy anyone???
Oh and that terrific health care plan that without knowing whats in it doesn’t bother her-cuz she will get a better one than you…Ole Lockjaw won’t even have to speak to gain a senate seat-says a lot about her dumbo supporters!
David

I am happy to live in the most liberal state in the union. I get nervous seeing all of the right wing posts on this site –a clearly vocal minority who could destroy our bastion of reason if we are not careful. Fellow liberals take heed – our activism in 08 rescued us from 4 more years of republican greed and incompetece; complacency now in this special election could add another
Visigoth into the fragile mix. Remember the disaster that was Bush, and get out the vote for Coakley!!!
kenny c

David: do you really think that massachusetts is a
“bastion of reason”. that is either a joke or one of the
most naive comments that i have ever heard of.
you should read “the bluest state: the massachusetts recipe for political disaster”. it is time to end the
one party state that is massachusetts. it never worked in
the soviet union and it hasn’t worked here.

Vt. Ski Resorts Await More Weather After Irene

Coakley And Brown In Statistical Tie