Welcome to poll season.

Whether it is traditional polling entities such as Marist, Quinnipiac, Franklin & Marshall or Muhlenberg (pictured) or those funded by campaigns, expect to see numbers, numbers and more numbers released during the next 40 days.

There is unprecedented interest in the Pennsylvania primaries for U.S. Senate, Governor and Lieutenant Governor – not only among the larger than usual number of candidates – but also those of us who report on the campaigns trying to make sense of what it all means.

One day it’s an internal memo telling us that John Fetterman has a 30-point advantage over Conor Lamb. The next day, it’s dueling polls issued by the Mehmet Oz and David McCormick campaigns – each stating that their candidate has a lead. The one constant in every poll released to date – the number of undecided or unsure voters.

There are three reasons behind why primary surveys are less accurate and more difficult to conduct for pollsters than general election surveys.

1. Voters do not pay much attention to primary elections

2. Uncertainty about candidate preferences happens in closed primary states such as ours because people are voting for candidates who represent their chosen party.

3. Primary voters are less certain about whether they will actually vote.

According to the Franklin & Marshall Center for Opinion Research, primary turnout since 2000 in the Commonwealth has exceeded 40 percent only once among Republicans (2016) and three times among Democrats (2016-18-20). During the same period, turnout has never been less than 40 percent in general elections.

The bottom line – primary voters take much longer to make up their mind and may not even do so until Election Day.

With this in mind, how do we read primary polls? 

When looking at past trends, there are three important patterns.

1. Primary polls tend to be more accurate the closer to Election Day

2. There is a lot of variability in primary polling estimates.

3. Primary polls most often tend to underestimate the lead of the winning candidate.

A great example would be the presidential campaign of 2016 in Pennsylvania. Donald Trump and Hillary Clinton were both expected to easily win their respective primaries. Polls in the last month showed a Trump lead running from eight to 26 points and the Clinton advantage running between 6-28 points. Trump won by 35 and Clinton by 12.

The takeaway – polls are a snapshot in time. They are relevant in the moment and represent the opinions of a population. As Courtney Kennedy wrote in 2020 for the Pew Research Center, “a robust public polling industry is a maker of a free society.” Different polling organizations conduct their surveys in different ways. While some consider themselves “nationally representative,” it is not a guarantee that the methodology is solid.

As the noted political scientist Sidney Verba explained, “Surveys produce just what democracy is supposed to produce – equal representation of all citizens.”