Question:
And in the race for Senator from Maine, if Election Day were today, who would you vote for:
Tom Allen, Democrat or Susan Collins, Republican? [Options were read and rotated; if
undecided, respondent was asked which way he/she is leaning at this time.] (Q4)
More than half (56.8%) of those polled said that, if today were Election Day, they would “vote / lean”
Susan Collins for Senator. Thirty-six percent (36.0%) of respondents said that they would “vote /
lean” Tom Allen, and 7.3% are undecided. Thus, respondents support Susan Collins by a margin
of 21 percentage points at this point.
“Vote / Lean” Susan Collins for Senator from Maine
Higher Lower
 Republican (92.5%)  Democrat (30.3%)
 < $35K income level (67.3%)  $65K+ income level (50.8%); $35K < $65K
income level (53.5%)
June
2008
(N=400)
October
2008
(N=400)
Vote Tom Allen 28.8% 27.8%
Lean Tom Allen 2.3% 8.3%
Vote Susan Collins 47.5% 49.0%
Lean Susan Collins 8.3% 7.8%
Still undecided 13.3% 7.3%
Vote / Lean Tom Allen 31.1% 36.0%
Vote / Lean Susan Collins 55.8% 56.8%

Poll Demographics

About this Poll
The most recent Pan Atlantic SMS Group Omnibus Poll was conducted between October 13th and
October 16th, 2008. This independent survey data is being released to the Maine media in the public
interest.
All interviews were completed on a CATI (Computer Aided Telephone Interviewing) system at the Pan
Atlantic SMS Group Interview Center by our in-house staff. This omnibus survey is the 40th in a series
of omnibus surveys conducted by Pan Atlantic SMS Group. Because we have conducted this poll
on a frequent basis over a long time period (since 1995), we are in a unique position to provide
reliable benchmarking on a range of important issues.
A randomly selected, stratified statewide sample of 400 Maine adults was interviewed. Since the poll
contains questions relating to issues which will be voted on in the November 2008 elections, the
survey was administered only to people who are registered voters and who identified themselves as
being “likely” to vote or have already voted via absentee ballot in the November 2008 elections. All
others were excluded from participation. Although the survey instrument was administered only
to registered voters who said that they are likely to vote or have already voted via absentee
ballot on November 4th, the results contained herein represent a snapshot of voters’ opinions
at a point in time. They do not purport to predict final poll results. As Election Day was
approximately three weeks away at the time in which the poll was conducted, changes in final
voter behavior may occur in the interim. This could result from increased voter scrutiny of the
issues, media coverage, editorials, advertising, etc.
The sample was stratified based on U.S. Census of Population and Housing data. The sample size
has statistical significance of  4.9 percent at the 95 percent confidence level. This means that if the
survey were to be repeated, 95 times out of 100 the results would reflect the results of this survey
within the ± 4.9 percent margin of error. The results are broken out by various demographic subsamples,
including Congressional District, political party affiliation, age, household income, and
gender. The margins of error for specific sub-samples are significantly higher than the ± 4.9 percent
margin of error for the entire sample. Please note: The margin of error for the District One
Congressional Race question is ± 6.81 percent and for the District Two Congressional Race
question is ± 7.05 percent at the 95% confidence level.