The Poll Sample Rules Work Both Ways WBUR vs Boston Herald UPDATE: Et Tu Fox?

Readability

The Poll Sample Rules Work Both Ways WBUR vs Boston Herald UPDATE: Et Tu Fox?

As every­one knows I’ve been hit­ting the polls lately for their sim­ply out­ra­geous attempts to cre­ate a false nar­ra­tive by play­ing games with Demo­graphic results.

In the last 48 hours two polls have come out con­cern­ing the Scott Brown Race in Mass­a­chu­setts with dia­met­ri­cally dif­fer­ent results, one by the lib­eral lean­ing WBUR a pub­lic radio Sta­tion out of Boston and once by the Boston Her­ald the right lean­ing News­pa­per of the city.

WBUR puts For­mer Obama Admin­is­tra­tion Offi­cial Eliz­a­beth War­ren up by 5, The Her­ald puts Sen­a­tor Scott Brown up by 6. That is fully a 11 point dif­fer­ent in two polls released a day apart. Obvi­ously they BOTH can’t be reflec­tive of the State as a whole so lets look at them one at a time:

WBUR: This is a real dif­fer­ence from the last poll I looked at that showed a dead heat so when this poll came out I was ready for the worst and went straight for the party splits.

The Good news is finally FINALLY I’m see­ing a poll where the splits match the demo­graph­ics of the state. The unweighed poll num­bers were as follows:

Total 507 voters

Democ­rats polled 185 37%
GOP 60 12%
Unen­rolled or other 262 51%

That’s a D+25 split which if this was a national poll would be a com­plete joke (and not quite unex­pected con­sid­er­ing the Whop­pers we’ve seen lately) but in Mass­a­chu­setts a state with under 5 GOP mem­bers in it’s sen­ate and where get­ting to 25% rep­re­sen­ta­tion in the House was an ACHIEVE­MENT that’s not a joke, that’s the real­ity It matches up with the pop­u­la­tion based on the offi­cial state num­bers from ear­lier this year.

WBUR’s poll gets War­ren up 5 by the fol­low­ing results:

Sen Scott Brown:

With Unen­rolleds (Inde­pen­dents): 51% elec­torate Brown up 4935 (ind) Brown gets 24.99 vote

With the GOP: 12% elec­torate Brown up 859 ® 10.2 vote

With Democ­rats: 37% elec­torate War­ren up 7115 (D) 5.55 vote

That adds up to 40.5 % of the vote for Scott Brown

On the other end For­mer Obama Admin­is­tra­tion Offi­cial War­ren break­down adds up to 17.85 (Ind) + 1.08 (GOP) + 26.27 (DEM)

That adds up to 45.2%

These num­bers match the polling and demo­graph­ics so I find this poll: Cred­i­ble

— — — — — — — -

Boston Her­ald:

Now lets look at the Her­ald poll that has Brown up by 6 num­bers

This poll shows Sen Brown up Buy 6 points among reg­is­tered vot­ers & 4 points among likely vot­ers. It’s the first poll I’ve ever seen show­ing a Repub­li­can doing WORSE among likely vot­ers than reg­is­tered ones.

Sen Scott Brown takes: 93% of GOP & Lean­ers 58% of Inde­pen­dents & 20% of Democ­rats.
For­mer Obama Admin­is­tra­tion Offi­cial Eliz­a­beth War­ren takes: 7% of Repub­li­cans & Lean­ers, 35% of Inde­pen­dents & 73% of Demo­c­rats

The results are dra­mat­i­cally dif­fer­ent vs the WBUR poll Brown’s share of the GOP vote is 8pts higher Warren’s 2 pts lower, With inde­pen­dents Brown is +9 in the Her­ald with War­ren flat & with Dems Brown’s Share jumps +5 pts while Warren’s is +2

That accounts for part of the dif­fer­ence but lets look at the all impor­tant demo­graphic split

lets focus on that high­lighted area

This rep­re­sents a split of +17 Dems with­out lean­ers and +20 with peo­ple who lean one way or the other. This how­ever doesn’t match the lat­est demo­graph­ics of the state. While the GOP fig­ure is about right the Demo­c­rat fig­ure is of 28% among likely vot­ers is a full 9 points lower than it should be.

The Her­ald poll indi­cates Sen Brown win­ning one in five democ­rats so that under-​sampling has a lesser effect but even so under-​sampling by 9 points? It just won’t do. If you adjust based on that under-​sampling that Turns the Sen Brown 46 pt lead into a War­ren 13 pt lead.

Thus I must dub the Her­ald poll: Not Credible

I’d like to Believe the Her­ald poll, I really would and I’m aware of a shift of peo­ple away from the Demo­c­rat party to the ranks of the Unen­rolled but with­out solid data to back it up I have to believe the WBUR poll over the Her­ald one. More­over sev­eral anec­do­tal bits of evi­dence have cropped up lately that sug­gest For­mer Obama Admin­is­tra­tion Offi­cial Eliz­a­beth War­ren has improved her stand­ing which makes the WBUR poll more believ­able to me (more on that tomorrow).

But “DaT­e­chGuy” you Might say, “You’re a Repub­li­can, you want Brown to win. It’s your duty to back up the poll to help swing the momentum.”

That may be the MSM way but I say balder­dash! The way to win a cam­paign is to look at it hon­estly and base your actions accord­ingly. If I am win­ning I want to know it so I can keep exploit­ing my suc­cess, if I’m not I want to know it so I can cor­rect any prob­lems or mistakes.

Self delu­sion is no way to win an elec­tion, a les­son the MSM will learn nation­ally in November.

Update: Sharp eyed read­ers will notice the orig­i­nal name for this post was “WBUR shows how it’s done”. I started writ­ing it yes­ter­day but crashed on the couch exhausted before I fin­ished it. When I woke up at 2:30 AM the Her­ald poll was out and decided to shift the focus to a com­par­i­son and sim­ply for­got to rename the inter­nal title.

Update 2: Today new fox poll that is sup­pos­edly “dev­as­tat­ing” to the Rom­ney Cam­paign Parisian splits Ohio D +6 Vir­ginia D +5 Florida D +5 Ho Hum, more of the same.

Update 3: it is spelled Pew but per­haps it should be spelled PU. Pew reports Obama +8 Sam­ple D +9 the biggest joke of the poll is this:

That’s a 15 point Obama edge in an elec­tion that Obama won by 7.2 but it’s even worse because those num­bers are based on 86. if you do the full per­cent­ages based on 100% of vot­ers it comes out Obama 54.6% (+1.7 over 2008) and McCain 37.2% (-8.6 under 2008) in this sample.

Now you might think Obama is ahead but does any­one ANY­ONE believe he will do 1.7 point bet­ter than 2008. Does any­one ANY­ONE believe Rom­ney will under­per­form McCain by 8.6?

Maybe that will fool Talk­ing points Memo, Fire­doglake and the like but it won’t fool any­one who can, you know do MATH.

As everyone knows I’ve been hitting the polls lately for their simply outrageous attempts to create a false narrative by playing games with Demographic results.

In the last 48 hours two polls have come out concerning the Scott Brown Race in Massachusetts with diametrically different results, one by the liberal leaning WBUR a public radio Station out of Boston and once by the Boston Herald the right leaning Newspaper of the city.

WBUR puts Former Obama Administration Official Elizabeth Warren up by 5, The Herald puts Senator Scott Brown up by 6. That is fully a 11 point different in two polls released a day apart. Obviously they BOTH can’t be reflective of the State as a whole so lets look at them one at a time:

WBUR: This is a real difference from the last poll I looked at that showed a dead heat so when this poll came out I was ready for the worst and went straight for the party splits.

The Good news is finally FINALLY I’m seeing a poll where the splits match the demographics of the state.  The unweighed poll numbers were as follows:

Total 507 voters

Democrats polled 185 37%
GOP  60 12%
Unenrolled or other 262 51%

That’s a D+25 split which if this was a national poll would be a complete joke (and not quite unexpected considering the Whoppers we’ve seen lately) but in Massachusetts a state with under 5 GOP members in it’s senate and where getting to 25% representation in the House was an ACHIEVEMENT that’s not a joke, that’s the reality It matches up with the population based on the official state numbers from earlier this year.

WBUR’s poll gets Warren up 5 by the following results:

Sen Scott Brown:

With Unenrolleds (Independents):   51% electorate Brown up 49-35 (ind) Brown gets 24.99 vote

With the GOP:   12% electorate Brown up 85-9 (r)  10.2 vote

With Democrats:  37% electorate Warren up 71-15 (D) 5.55 vote

That adds up to 40.5 % of the vote for Scott Brown

On the other end Former Obama Administration Official Warren breakdown adds up to 17.85 (Ind) + 1.08 (GOP) + 26.27 (DEM)

That adds up to 45.2%

These numbers match the polling and demographics so I find this poll:   Credible

———————-

Boston Herald:

Now lets look at the Herald poll that has Brown up by 6 numbers

This poll shows Sen Brown up Buy 6 points among registered voters & 4 points among likely voters. It’s the first poll I’ve ever seen showing a Republican doing WORSE among likely voters than registered ones.

Sen Scott Brown takes:   93% of GOP & Leaners 58% of Independents & 20% of Democrats.
Former Obama Administration Official Elizabeth Warren takes:   7% of Republicans & Leaners, 35% of Independents & 73% of Democrats

The results are dramatically different vs the WBUR poll Brown’s share of the GOP vote is 8pts higher Warren’s 2 pts lower, With independents Brown is +9 in the Herald with Warren flat & with Dems Brown’s Share jumps +5 pts while Warren’s is +2

That accounts for part of the difference but lets look at the all important demographic split

lets focus on that highlighted area

This represents a split of +17 Dems without leaners and +20 with people who lean one way or the other. This however doesn’t match the latest demographics of the state. While the GOP figure is about right the Democrat figure is of 28% among likely voters is a full 9 points lower than it should be.

The Herald poll indicates Sen Brown winning one in five democrats so that under-sampling has a lesser effect but even so under-sampling by 9 points?   It just won’t do. If you adjust based on that under-sampling that Turns the Sen Brown 4-6 pt lead into a Warren 1-3 pt lead.

Thus I must dub the Herald poll:  Not Credible

I’d like to Believe the Herald poll, I really would and I’m aware of a shift of people away from the Democrat party to the ranks of the Unenrolled but without solid data to back it up I have to believe the WBUR poll over the Herald one. Moreover several anecdotal bits of evidence have cropped up lately that suggest Former Obama Administration Official Elizabeth Warren has improved her standing which makes the WBUR poll more believable to me (more on that tomorrow).

But “DaTechGuy” you Might say, “You’re a Republican, you want Brown to win.  It’s your duty to back up the poll to help swing the momentum.”

That may be the MSM way but I say balderdash! The way to win a campaign is to look at it honestly and base your actions accordingly.  If I am winning I want to know it so I can keep exploiting my success, if I’m not I want to know it so I can correct any problems or mistakes.

Self delusion is no way to win an election, a lesson the MSM will learn nationally in November.

Update:  Sharp eyed readers will notice the original name for this post was “WBUR shows how it’s done”.  I started writing it yesterday but crashed on the couch exhausted before I finished it.  When I woke up at 2:30 AM the Herald poll was out and decided to shift the focus to a comparison and simply forgot to rename the internal title.

Update 2:  Today new fox poll that is supposedly “devastating” to the Romney Campaign Parisian splits Ohio D +6  Virginia D +5 Florida D +5 Ho Hum, more of the same.

Update 3:  it is spelled Pew but perhaps it should be spelled PU.  Pew reports Obama +8 Sample D +9 the biggest joke of the poll is this:

That’s a 15 point Obama edge in an election that Obama won by 7.2 but it’s even worse because those numbers are based on 86.  if you do the full percentages based on 100% of voters it comes out Obama 54.6% (+1.7 over 2008) and McCain 37.2% (-8.6 under 2008) in this sample.

Now you might think Obama is ahead but does anyone ANYONE believe he will do 1.7 point better than 2008.  Does anyone ANYONE believe Romney will underperform McCain by 8.6?

Maybe that will fool Talking points Memo, Firedoglake and the like but it won’t fool anyone who can, you know do MATH.