Your SlideShare is downloading. ×
SurveyMonkey 2012 Presidential Election Poll
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Saving this for later?

Get the SlideShare app to save on your phone or tablet. Read anywhere, anytime - even offline.

Text the download link to your phone

Standard text messaging rates apply

SurveyMonkey 2012 Presidential Election Poll

26,059
views

Published on

Below is our newest wave of data from our presidential election poll. These results will be displayed here in two different ways: first, as popular vote percentages and second as Electoral College …

Below is our newest wave of data from our presidential election poll. These results will be displayed here in two different ways: first, as popular vote percentages and second as Electoral College distributions. SurveyMonkey has surveyed more than a million people from August 17th to October 30th. The results below, however, will be based only on the roughly 500,000 people surveyed from October 3rd through October 30th.


1 Comment
0 Likes
Statistics
Notes
  • Very interesting presentation, and obviously a huge commercial opportunity for SM. Well done! (How do I invest in SM?)

    I would have liked to see one more thing: the Electoral College Impact (or 'ECI' - I love TLA's :-)) of the above polling results. We;re pretty sure that Al Gore lost his election despite winning a majority the popular vote. Will the same happen to Obama? (We think not, but stranger things have happened. Will a chad hang once more?)

    So - how about keeping your interested northern neighbours posted by running a live-feed exit poll?

    And thanks for everything!
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Be the first to like this

No Downloads
Views
Total Views
26,059
On Slideshare
0
From Embeds
0
Number of Embeds
6
Actions
Shares
0
Downloads
20
Comments
1
Likes
0
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide

Transcript

  • 1. ! ! ! "#$"!%&()*+,)-.!/.0,)1+!%1..!! ! ! "#$%$&!()%(*+,!"#-.-,!/"!012#3+3%345!! ! ! .(61!3%+71)4,!89:! ;$(*(!9&<21$*,!"#-.-,!=1*$3)!012#3+3%34$<2
  • 2. ! "! 2345/6789:/6!! "#$"!%4/2;</9=;>?!/?/@=;89!%8??!! ! 9<2$>(2$*4!2#1!"3&?%(2$3*!SurveyMonkey (SM) facilitates survey deployment to 60 million respondents each month for 13million researchers worldwide. Roughly half of this activity occurs in the United States. Due to thiswide usage of our products and services, our monthly traffic broken down geographically provides agood approximation of the actual population density of the US. (See Figure 1.)A;B34/!$C SM Traffic and US Population Densities 32!%1DE.-,)1+!<+(),F! ! ! ! ! ! ! =3?)@1A!! B)11#?441)-@3>,!CDDE! ! ! 2E&GF71+HF!71+,I.F!=&-JJ)0!<+(),F!!
  • 3. ! #! 01(<?)$*4!"?7%$@!:&$*$3*!SM realized a unique opportunity to utilize its high traffic volume to measure public opinion in theUnited States. From August 16, 2012 to present, SM has directed varying portions of completedsurvey traffic to an appended SM home landing page. (See Figure 2.) This append offered theopportunity to answer a series of political questions with the aim of gauging presidential votingpreferences.A;B34/!"C SM Home Landing Page AppendDue to its high traffic volume, SM can easily generate daily tracking data. Since mid August, SM hascollected vote preferences from 810,477 Americans in 2,946 counties through a 30% averageresponse rate over the period. A preliminary geographical depiction of vote choice suggests facevalidity of this data collection method. (See Figure 3.)A;B34/!KC A Comparison of U.S. and SM Vote Preference by Geography "##L!>0,E-.!51,(! =3?)@1A!! F1G<>(H-@3>,!CDDI!!
  • 4. ! $! 2E&GF71+HF!"#$"!/.0,)1+!2,E*F!Respondents provided four pieces of information: their zip code, voting registration status, votinglikelihood, and candidate preference. (See Appendix for the exact question wordings.) To gaugevoting likelihood respondents were asked one of eleven different likely voter questions that arecurrently used by other polling firms. =0!"3%%$*4!T1<?%2<!The SM polling results below include only registered, likely voters. Of the 248,786 responses SM hasrecorded from October, 3 to October, 18, 91% were registered voters and 93% of those registeredwere likely voters. These remaining 191,387 registered, likely voters are summarized in the tablesbelow.The data in Table 1 and Figure 4 show SM’s recent estimates of the national popular vote for BarackObama and Mitt Romney using three and seven day trailing averages among likely voters.=>M?/!$C National Popular Vote Trailing Averages Among Likely Voters in October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
  • 5. ! %!Different telephone polling firms use slightly different methodologies, but they all face the commonobstacles of hand-dialing mobile numbers and randomizing the dialing of landline numbers withcomputer aided telephone interviewing (CATI) software. Moreover, all studies require extensive“call-backs” to achieve adequate sample sizes. Once this arduous data collection is finished, datacleaning, collation, and analyses add additional time to the project before polling firms are able torelease the data.The fact that SM can collect meaningful sample sizes to track trends on a daily basis allows for a highdegree of sensitivity to media and political events. This method of tracking public opinion drasticallyreduces the lag-time between any given event and the publication of the impact of that even on publicopinion. The large number of daily respondents allows SM to generate sensitive, daily estimates of thenational popular vote. (See Figure 4 for the daily tracking history.)The figures below summarize the voting preferences of 753,685 likely voters (taken from a total poolof 852,674 respondents) surveyed by SM from 8/17/12 – 10/18/12. To compare the SM data to thatwhich is publicly available, historical poll averages from 8/17/12 – 10/18/12 were downloaded fromReal Clear Politics (RCP). (For more information about the polls that are included in these averages,see: www.RealClearPolitics.com.)A;B34/!QC Daily National Popular Vote Trailing Averages Among Likely Voters <2 *+ )+ T3>*15!!TF8! :7(>(!.F8! ;$75(! JNM!6$+13! ;$75(! L !+17(21! C !+17(21! O !+17(21! <&11@#! <&11@#! (22(@U! %1(U1+! 7(@U%(<#! V")1<-W! V/"W! V")1<-W!The daily SM data is only weakly aligned with the RCP curve, with an r2 of .20 for Obama and an r2of .21 for Romney. This daily tracker, however, is quite different from how poll data is typicallyreported. The RCP daily average is essentially a "poll of polls". It is a rolling, historical average ofpast poll data; the polls included in these RCP numbers span anywhere from a two to twelve dayrange. This time-lapse makes the numbers that are reported less volatile. Thus, a 3-day trailingaverage of SM data doubles the fit with the RCP curve, increasing r2 to .41 for Obama and .40 forRomney. Our 7-day average aligns even more closely to the RCP data, with r2 further increasing to.56 for Obama and.57 for Romney. (See Figure 5.)!
  • 6. ! &!A;B34/!RC National Popular Vote Trailing Averages (SurveyMonkey 7-day average vs. RCP)As Figure 5 shows, using a 7-day average smoothes out the SM data curve, making the numbers lessprone to the effect of any given day. This increases the fit of SM data to RCP data; however, it masksthe impact of topical political events and obscures daily shifts in public opinion. SM numbers, overall,are more volatile than RCP data, but this is likely reflective of real-time shifts in public opinion.!
  • 7. X&&1*+$H!^!_?1<2$3**($)1! ! !Voting Registration. E- ]3G!@1)2($*!()1!53?!2#(2!53?!G$%%!6321Y! • X)1!53?!@?))1*2%5!(!)14$<21)1+!(*+!1%$4$7%1! GM3.>;/0>B&N0H/=<6& 6321),!3)!*32Y! O=<H>B&N0H/=<6& 203& -./&N0H/=<6& -.& C.6D/&E6.F& ! ! N- ]3G!%$U1%5!()1!53?!23!6321!$*!F361>71)<!Zip Code. &)1<$+1*2$(%!1%1@2$3*Y! • Z#(2!$<!2#1!S$61[+$4$2!R$&!@3+1!S3)!2#1!(++)1<<! PQ/H0@0>B&><E0>B& 53?!)14$<21)1+!23!6321!S)3>,!3)!$S!53?)1!*32! R0HB&><E0>B& )14$<21)1+!23!6321,!G#(2!$<!2#1!R$&!@3+1!53?! ?.@0FK=/&><E0>B& G3?%+!?<1Y! ?><SK/>B&><E0>B& 4.5067068089& -./&=/&=>>&><E0>B& ! I- B#$*U$*4!7(@U!23!2#1!1%1@2$3*<!#1%+!S3)! 83*4)1<<!$*!F361>71)!CDLD,!+$+!2#$*4<!@3>1!Voting Likelihood. ?&!2#(2!U1&2!53?!S)3>!632$*4,!3)!+$+!53?! L- ]3G!>?@#!2#3?4#2!#(61!53?!4$61*!23!2#1! #(&&1*!23!6321Y! ?&@3>$*4!1%1@2$3*!S3)!&)1<$+1*2Y! 203#&L./08& :;</0&=&>./& -.#&8<8&6./&L./0& ?.@0& C.6D/&E6.F& A6>B&=&><//>0& ! -.60& P- ]3G!$>&3)2(*2!$<!2#1!&)1<$+1*2$(%!1%1@2$3*!23! C.6D/&E6.F& 53?Y! ! PQ/H0@0>B&<@5.H/=6/& C- .3!53?!#(&&1*!23!U*3G!G#1)1!&13&%1!G#3!%$61! R0HB&<@5.H/=6/& $*!53?)!*1$4#73)#33+!43!23!6321Y! ?.@0FK=/&<@5.H/=6/& 203& ?><SK/>B&<@5.H/=6/& -.& -./&=/&=>>&<@5.H/=6/& C.6D/&E6.F& ! ! LD- QS!2#1!1%1@2$3*!G1)1!#1%+!23>3))3G,!G3?%+!53?! O- ](61!53?!161)!6321+!$*!53?)!&)1@$*@2!3)! U*3G!G#1)1!23!43!6321Y! 1%1@2$3*!+$<2)$@2Y! 203& 203& -.& -.& ! C.6D/&E6.F& LL- B#$*U$*4!7(@U!23!2#1!1%1@2$3*<!#1%+!S3)! ! 83*4)1<<!$*!F361>71)!CDLD,!+$+!53?!6321Y! J- ]3G!3S21*!G3?%+!53?!<(5!53?!6321!^!(%G(5<,! 203#&L./08& *1()%5!(%G(5<,!&()2!3S!2#1!2$>1,!3)!<1%+3>Y! -.#&8<8&6./&L./0& G>F=B3& ! -0=H>B&=>F=B3& I=H/&.J&/K0&/<@0& Voting Preference. ?0>8.@& • =?&&3<1!2#1!&)1<$+1*2$(%!1%1@2$3*!G1)1!#1%+! -0L0H& 23+(5-!Z#3!G3?%+!53?!71!%$U1%5!23!6321!S3)Y! C.6D/&E6.F& T=H=NE&AM=@=& ! U<//&V.@60B& K- .3!53?,!53?)<1%S,!&%(*!23!6321!$*!2#1!1%1@2$3*! C.6D/&E6.F&W&A/K0H& 2#$<!F361>71),!3)!*32Y! 203& -.& C.6D/&E6.F&