By Jim Ellis — Monday, March 16, 2026
Polling
Polling is an important element in campaigning for office and also an invaluable tool in helping to forecast elections, but one that can also mislead. That’s why it’s so important to understand the polling methodology in order to best comprehend what the data tells us.
A good example of not taking a ballot test at first glance comes in Maine’s 2nd Congressional District Democratic primary where two polls taken within the same relative time frame are producing very different results. Delving into the methodology gives us a better idea of which pollster has the better information.
The 2nd District is Maine’s northern congressional seat that encompasses most of the Pine Tree State’s vast land area and stretches to the Canadian border. The major population centers are the cities of Bangor, Augusta, and the Lewiston-Auburn metro area. Four-term Rep. Jared Golden (D-Lewiston) is not seeking re-election.
The ME-2 seat is also the most Republican congressional district in the country that sends a Democrat to the House of Representatives. President Trump has carried the 2nd District in all three of his national general election campaigns. The Dave’s Redistricting App partisan lean calculation shows a 52.9R – 41.1D split. The 2026 open race here may become the Republicans’ best conversion opportunity in the nation.
The 2026 campaign features former two-term Gov. Paul LePage, who faces only minor opposition for the Republican nomination, and two major Democrats vying to become their party’s standard bearer: state Sen. Joe Baldacci (D-Bangor), brother of former Governor, ex-Congressman, and previous state Sen. John Baldacci, and State Auditor and former Secretary of State Matt Dunlap.
The conflicting polls in question come from the Maine-anchored Pan Atlantic Research firm conducting its regular “Omnibus Poll,” and Tulchin Research for the Dunlap campaign.
The methodology disclosure for the Pan Atlantic survey reveals that 810 likely voters were questioned online during the period of Feb. 13 through March 2nd. Tulchin conducted their survey soon after Pan Atlantic finished, from March 5-8 of 400 likely ME-2 voters, via multiple sampling techniques.
When looking at the 2nd District Democratic primary, Pan Atlantic projects Sen. Baldacci as holding a significant 36-14-12 percent advantage over Dunlap and former congressional aide Jordan Wood, respectively. Tulchin, however, found a very different result. According to the TR ballot test, it is Dunlap who holds a 36-29-11 percent lead over Sen. Baldacci and Wood.
Obviously, the two pollsters come to contrary conclusions as to what may happen in this particular political campaign. Digging deeper, we can find which of these two surveys is likely the more accurate predictor.
Looking at the Pan Atlantic methodology, we see, as previously stated, that 810 likely voters have been surveyed. This, however, is the number for the entire state and contains many more than just 2nd Congressional District likely Democratic primary voters. The Tulchin Research data focuses only on the 2nd District, and though their sample size of 400 respondents is approximately only half the size of the Pan Atlantic aggregate, in actuality it is the better sample.
Looking closely at the Pan Atlantic report, we find that only 144 individuals comprise the 2nd District Democratic primary subset. This is barely large enough to form a segment cell let alone a sample for an entire congressional district. Therefore, while this Pan Atlantic survey is viable to forecast the statewide general election races, its candidate preference data for the 2nd District Democrats that shows Sen. Baldacci holding a large lead should be discarded.
The better picture comes from the Tulchin Research data. Their polling sample of 400 likely partisan primary voters is sound, thus the projection of Dunlap leading the Democratic primary beyond the polling margin of error is the more believable study.
Expect this Democratic primary to be hard fought through the June 9 primary election. The winner will then face former Gov. LePage in what promises to be a highly competitive general election congressional contest in a race that carries national implications relating to which party will control the House of Representatives in the next Congress.