Pan Atlantic Research Omnibus Poll - How it Works

  • The poll of 800 randomly selected Maine citizens is a low-cost, shared method of assessing public opinion and market data in Maine.

  • The results of our Omnibus Poll are representative of the adult population of Maine.  Data is segmented by geographic region, gender, income level, age, educational attainment level, etc., and weighted to match US Census demographic estimates.

  • The Pan Atlantic Research Omnibus Poll offers regular benchmarking results on any issue.

  • Your organization may elect to include a proprietary question/s at any time, with no obligation to purchase future questions.

  • The specific results of your proprietary question or questions are only released to your organization

  • You will receive the results of your question or questions in both tabular and graphical report formats.

  • Free of charge, we will assist you in writing an unbiased question(s) so that your company or organization will have actionable information.

  • To include a question in any of our Omnibus Polls, we charge between $700 (cost of single question) and $900 (multiple-part question) per question, depending upon the scope and complexity of the question.

  • The proprietary results are the exclusive property of those clients who commissioned them.  These are available within 5 business days of the last day of interviewing. 


Client Questions and Publicly Released Questions Are Separate

Each Omnibus Poll contains two distinct kinds of questions:

  • Client questions are proprietary questions commissioned by businesses, nonprofits, campaigns, and other organizations. The client funds the question, and the results are confidential — released only to the organization that commissioned them, as noted above.

  • Publicly released questions are written and paid for by Pan Atlantic Research and shared with the media and the public free of charge, to provide an independent benchmark of Maine public opinion.

The two are kept strictly separate. No client can purchase placement in the publicly released questions, and no client writes, reviews, edits, or sees the results of those questions before they are published. Over the years we have declined every request from candidates and organizations seeking to buy their way into our public polling. That independence is what makes the public results worth publishing, and we protect it accordingly.


The Pan Atlantic Research Omnibus Poll Results Have Matched:

  • 5 out of 5 Citizens Initiative questions polled in 2024

  • 7 out of 7 Gubernatorial elections polled since 1998

  • 6 out of 6 Presidential elections polled since 2000

  • 7 out of 8 Senate elections polled since 2000

  • 7 out of 8 US House elections polled since 2018

Notes: PAR did not poll 2025 Citizen's Intiative questions. Based on final state- or district-wide first-round election results released in each election cycle. PAR did not release political polling in 2016. PAR did not consistently poll US House races prior to 2018.


Pan Atlantic Research 69th Omnibus Poll Results

Pan Atlantic Research is pleased to release our May 2026 Omnibus Poll. This independent poll, the latest in a series which has been running since 1996, reports on a representative stratified sample of 827 likely voters' intentions, preferences, and opinions.

Methodology Notes:

  • The 69th Pan Atlantic Research Omnibus Poll was conducted between May 8th and May 18th, 2026.

  • A randomly selected sample of 1097 Mainers age 18 or older was recruited online using a panel provided by our vendor partner Dynata, a highly respected national research panel provider. Respondents who indicated that they were "likely," "very likely", or "certain" to vote in the 2026 November election were asked all questions on the survey, while others were asked only for demographic information for weighting purposes.

  • Data were weighted using US Census Bureau targets to match the composition of Maine’s population based on Congressional District, age, education, and gender.

  • The result is a sample of 827 likely voters, with a margin of sampling error of ±3.7% at the 95% confidence level.