FILENAME: 20071010 SES Ontario Election Memo - Group Wisdom
Page 1 -- 10-Oct-07
Ontario Election
From: Nik Nanos, SES Research
Date:
October 10, 2007
Re:
SES Special – Online Group Wisdom Project – Ontario Election Projection
As some of you know, at SES we’ve taken the opportunity presented by the Ontario
election to test some new math regarding projection techniques using an online
approach.
Yesterday we conducted our first test, which was among a group of self-selected but
demographically screened individuals who participated in the online study. We
recruited study participants through our “You be the Pollster Contest”. In the contest a
$500 prize will be given to the study participant(s) who submit the most accurate election
call. The group was not randomly selected nor is it representative of the populace.
We’ve decided to test an unrepresentative self-selected group to see what happens.
The study participants were asked a series of questions by means of an online survey
including a percentage vote projection question. We have run some quite promising
pilot tests but need an election to see if there is any merit in this approach.
Of note, of the 1,011 potential study participants, 268 completed the online questionnaire
yesterday. The table below details a comparison of the traditional random representative
telephone survey and the “group wisdom” of 268 screened, self selected,
unrepresentative online study participants. After the election we will be crunching the
numbers to see whether this approach merits further exploration and fine-tuning.
Regardless, it does provide some food for thought (see below).
Traditional Telephone
Non-traditional Group
Wisdom
Ontario Ballot Traditional
Random Representative
SES Telephone Survey
(N=441 Decided Voters)
Ontario Ballot Projection
Non-representative, Self
Selected Online Survey
(N=268 Test Subjects)
Liberal
42.6%
42.7%
Progressive Conservative
30.5%
32.9%
New Democratic Party
17.5%
16.7%