23 June 2013

2013 Massachusetts Special Senatorial (Markey vs. Gomez) - the Polling Wars

In two days, on June 25th, 2013, Bay State voters will go to the polls to select a US Senator to fill the vacancy left by now Secretary of State John Kerry.

Photo montage: Politico

In presidential politics, Massachusetts is a deep blue state. In local politics, it is also blue, but somewhat less so. Scott Brown (R) proved in the 2010 special Senatorial election that it is indeed possible for a Republican to be elected to the US Senate from this state, but he was the first Republican to hold a Senate seat from the "Old Colony State" since Edward Brooke from 1967-1979, making him the first GOP Senator from MA in 31 years. Here is a complete listing of US Senators from Massachusetts.

One thing is for sure: with the election of Scott Brown (R) in 2010 and his subsequent defeat to Elisabeth Warren (D) in 2012, to John Kerry's selection to become Secretary of State, Massachusetts will now have a completely new and "young" senatorial team after Tuesday's election. Ted Kennedy served as Senior Senator from Massachusetts from November 7th, 1962 until his death on August 25th, 2009 (close to 47 years). John Kerry served as Junior Senator from Massachusetts from January 2, 1985 until February 1, 2013 (a little more than 26 years) which means that for 24 years in a row, Massachusetts had the same two Senators. So, we are looking at a major turnover in this state. And it also means that Elisabeth Warren will become the Senior Senator from Massachusetts, after have served only one-half year thus far.

Ed Markey has served as a US Representative from Massachusetts for 35 and 1/2 years, since November 2, 1976. He was first Rep from MA-07 from 1976 until 2013 and since 2013, from MA-05 (redistricting). A Roman Catholic with Irish roots, Markey also appeared in a cameo role in the 2003 film "Gods and Generals". Rep. Markey will turn 67 on July 11th.

Gabriel Gomez (R) is to my knowledge the first Latino candidate for a high office in Massachusetts. His resume is also unique in that he is a former US Navy Seal, which alone deserves our respect and thankfulness. Mr. Gomez will turn 48 on August 27th.

So, the race is not just between two parties or ideologies, it is also between two generations.

Polling is showing Markey firmly in the lead, but more importantly, we are again seeing a form of polling war going on between established firms and Right-wing leaning firms that are putting out very disparate results, just as we saw in both the 2008 and 2012 general elections.  Now, it is likely that two or three more polls will come in between today and Tuesday and if that happens, I will update this blog entry, but to date, I have found data on 19 polls with direct Markey vs. Gomez head-to-head matchups:

Nr. 2013 MA Senatorial Date Sample MoE Markey Gomez Und/Oth Mar.

AVERAGE (one week): N/A N/A N/A 52,00 40,17 7,83 11,83

19 Emerson College 21.06.13 1422 RV +/-2.5 51 41 8 10
18 New England College 21.06.13 579 LV +/-4.07 56 36 8 20
17 McLaughlin (R) 20.06.13 1100 LV +/-3.0 47 44 9 3
16 UMass Lowell/Boston Herald 20.06.13 312 LV +/-5.6 56 36 8 20
15 On Message (Gomez - R) 18.06.13 600 LV +/-4.0 48 41 11 7
14 Boston Globe 15.06.13 508 LV +/-4.3 54 43 3 11
13 WBUR/MassINC 10.06.13 500 LV +/-4.4 46 39 15 7
12 Suffolk University 10.06.13 500 LV +/-4.4 48 41 11 7
11 McLaughlin (R) 07.06.13 400 LV +/-4.9 45 44 11 1
10 UMass/YouGov 05.06.12 357 LV +/-5.4 51 40 9 11
9 New England College 02.06.13 734 RV +/-3.62 52 40 8 12
8 Emerson College 23.05.13 867 RV +/-3.26 45 33 22 12
7 WBUR/MassINC 09.05.13 497 LV +/-4.4 46 38 16 8
6 Suffolk/7News 08.05.13 500 LV +/-4.4 52 35 13 17
5 PPP (D) 03.05.13 1539 LV +/-2.5 44 40 16 4
4 Emerson College 02.05.13 797 RV +/-3.4 42 36 22 6
3 Western NE University 20.04.13 480 LV +/-4.5 51 36 13 15
2 WBUR/MassINC 22.03.13 610 LV +/-4.1 44 25 31 19
1 UMass Lowell/Boston Herald 06.03.13 589 RV +/-4.0 47 28 25 19

16 of those 19 polls are from firms that have longstanding reputations, but not all of them necessarily positive. 3 of those 19 are right-wing leaning polls, one is even an Gomez internal poll (On Message). It is unusual for a candidate to leak his internal polling, but it is not unheard of. When it does happen, it is usually the underdog who does it.

I want to start with the Gomez internal poll, which shows him down 7 points to Markey, and point out a sentence that is at best very deceptive, but actually, just a huge lie:

"By the way, in 2010 Scott Brown was 9 points behind 10 days before the election."

This is not really true:

The Massachusetts special election that Scott Brown won was on January 19th, 2010. Here are the polls in the days that led up to his election:

Nr. 2010 MA Senatorial Date Sample MoE Brown Coakley Und/Oth Mar.

AVERAGE (one week): N/A N/A N/A 50,96 44,64 4,80 6,31

10 Rasmussen 19.01.10 1000 LV +/-4.5 51 47 2 4
19 Insider Advantage 18.01.10 804 LV +/-4.3 52 43 5 9
18 ARG 18.01.10 600 LV +/-4.0 52 45 3 7
17 Daily Kos / R 2000 (D) 18.01.10 600 LV
48 48 4 0
16 PPP (D) 17.01.10 1231 LV +/-2.38 51 46 3 5
15 Cross Target -Pajamas Media (R) 17.01.10 574 LV +/-4.09 51,9 42,3 5,8 9,6
14 Merriman Group (R) 17.01.10 565 LV
50,8 41,2 8 9,6
13 ARG 16.01.10 600 LV +/-4.0 48 45 7 3
12 Suffolk 15.01.10 500 LV +/-4.4 51 47 2 4
11 BMG / Research 2000 14.01.10 500 LV +/-4.0 41 49 10 -8
10 Rasmussen 12.01.10 1000 LV +/-4.5 47 49 4 -2
Only in the BMG R2000 poll was Coakley ahead, but all polls newer than that showed Brown clearly ahead. Not only that, R2000 went bust shortly thereafter due to polling fraud, so relying on their data now would be moot, anyway. Now, compare that with the 2013 polling above, and that should tell you something.  It appears that there has been strong and mathematically verifiable movement toward Rep. Markey since the debates in which Gomez did not do as well as he needed to.

And now, to the individual pollsters:

In 2012, Suffolk pretty much ruined it's reputation forever by proudly proclaiming almost ONE MONTH before the General election that it was not going to do any more polling for FL, NC and VA, that is has "already painted those states red. We're not polling any of those states again". As we all know, Obama retained both Florida and Virginia, worth 42 electoral votes, or 8% of the electoral "college".

Suffolk was also pretty atrocious in Massachusetts vis-a-vis the 2012 GE, but in the other direction:

State Pollster Pollster Margin Actual Margin Bias Bias (simplified)
Massachusetts Suffolk D +32 D +23,14 D +8,86 D +9
Massachusetts Exit Polling D +22,1 D +23,14 R +1,04 R +1
Massachusetts UN Mass / Lowell / Herald D +20 D +23,14 R +3,14 R +3
Massachusetts YouGov D +20 D +23,14 R +3,14 R +3
Massachusetts MassINC / WBUR D +20 D +23,14 R +3,14 R +3
Massachusetts Rasmussen D +19 D +23,14 R +4,14 R +4
Massachusetts WNEU / Mass Live D +18 D +23,14 R +5,14 R +5
Massachusetts PPP (D) – FINAL POLL D +15 D +23,14 R +8,14 R +8
Massachusetts Boston Globe / UNH D +14 D +23,14 R +9,14 R +9
Massachusetts Kimball Consulting ( R ) D +13 D +23,14 R +10,14 R +10
Those are the end polls for this state from November 2012. Then again, they were all off to some extent, EXCEPT the exit polls. Notice that all the pollsters that the Right loves to say are left-wing, like UN Mass / Lowell / Herald or PPP (D) or Boston Globe / UNH - all had a Republican mathematical bias in their polling for Massachusetts in 2012. Only Suffolk, which saw an Obama +32 race, was way off to the Left.

So, it is entirely possible that Ed Markey will do considerably better than the current +11.83% average he is holding.

McLaughlin is a pollster worth noting for it's inaccuracy:

In 2012, McLaughlin was off in the following states:

Colorado, Romney +4 (end result: Obama +5) - McLaughlin was off by 9. But it was not an end poll.
Florida, Romney +3 (end Result: Obama +1) - McLaughlin was off by 4. But it was not an end poll.
Indiana, Romney +19 (end result: Romney +10) - McLaughlin was off by 9. This was an end poll.
Virgina, Romney +7 (end result: Obama +4)  - McLaughlin was off by 11. This was not an end poll.

The one poll that McLaughlin got right in 2012 was it's October National poll:

National, Obama +4 (end result: Obama +4) - spot on. The poll was from October 4th, 2012.

It should be noted that the McLaughlin national poll and the Colorado poll were part of the same survey: the Colorado numbers were party of a deliberate oversampling for that state. This is not entirely uncommon and sometimes done to save money, and if the methodology is sound, there is really no reason why not, only it is interesting that the results were so disparate.

The McLaughlin poll is the only recent poll from Massachusetts to have the Democrat under 50%, which is highly unlikely when you look at the rest of this polling.

I am not saying that the McLaughlin poll from 20.06.2013, showing Markey +3, is wrong and that the the New England College AND UMass polls, both of which show Markey +20 and came out within 24 hours on one side or the other of the McLaughlin poll, are absolutely right. But I am saying that it is statistically impossible to have a 17 point disparity in polling with polls that were conducted and published concurrent to each other. When you look at the entire field of polls, McLaughlin is the obvious mathematical "outlier". If I take the four McLaughlin state polls from 2012 and average the number of points that McLaughlin was off, then the average would be +8.25 points to the Right. Assuming that this current McLaughlin poll could also be 8.25 points to the right and that Markey is actually up by +11.25 instead of by just three, please notice that this value is almost identical to Markey's actual polling average. In other words, McLaughlin's record of polling inaccuracy may indeed be showing in this race as well. And after Tuesday, we will all know for sure.

Now, with off-year elections, anything can happen. The weather can turn terrible and then turnout can be very low, which is usually good news for Republicans. But after the very complacent Martha Coakley campaign of 2010 against Scott Brown, I don't think that Democrats are taking anything for granted this time around. John Kerry won re-election for this seat in 2008 with an easy 66% of the vote, but remember, this is after his presidential candidacy of 2004 and virtually 100% name recognition. Neither Ed Markey nor Gabriel Gomez share this kind of name recognition - yet.

On a final but important note: for some reason, RasmussenReports has decided to not poll this important marquee race, at least until now.

After the election is called and the count is finalized, I will be running a comparison between the various pollsters predictions and the actual results.

No comments:

Post a Comment

Constructive comments and critique are always welcome. Please keep it polite and respectful.