Several national pollsters have concluded that Donald Trump is leading in Pennsylvania over Joe Biden. The RealClearPolitics polling average as of April 3 shows Trump with a 0.6% edge in the Keystone State.
Some pollsters even show Trump at 52% support. Susquehanna Polling and Research takes issue with some of these polls and questions their understanding of the geographic factors that make Pennsylvania a difficult state to poll.{/div}
In SP&R’s recent poll, Trump trails President Biden by a 50–45 margin. The poll’s margin of error is +/-4.6%, so Biden’s support can be as high as 54% or as low as 46%. The race could really be a tossup at best for Trump. SP&R polls have never shown Trump in the lead. For instance, in our January poll, Trump trailed Biden by 7 points. Now it’s 5 points. According to RealClearPolitics, we’re one of the only firms to show Biden with a lead.
We believe that polls may be overstating Trump’s support for several reasons.
First, our polling shows Biden leading handily with Independent and “swing” voters. These are critical cohorts of the Pennsylvania population that exit polls showed Trump lost to Biden in 2020. For instance, Independent voters — which include both registered Independents but also self-described Independents, some of whom are registered Republicans and Democrats — are leaning toward Biden by 2–1 or better margins. Trump can’t carry the Keystone State without this critical bloc.
In our recent poll, about 42% of registered voters of both parties say that they are “straight ticket” voters. This leaves the remaining 16% who say they are “ticket splitters” or “swing” voters who split their tickets in most elections. Among them, Biden leads Trump handily. The straight-ticket vote can’t get you to 51% in a Pennsylvania statewide election.
Independents, ticket-splitters, and swing voters are vastly different demographically from Eastern to Western Pennsylvania. For instance, in Southeastern Pennsylvania, swing voters are centrist Republicans who are socially moderate and/or progressive on social and cultural issues, but fiscally conservative on fiscal and taxing issues. These voters despise Trump’s harsh immigration rhetoric, are unabashedly pro-gun control, and are liberal on transgender and abortion rights. They don’t support Trump, and they’ll be a key factor in determining how big he loses the vote-rich, Southeast collar counties around Philadelphia.
By contrast, in the Southwest/Pittsburgh market, swing voters are culturally conservative Democrats who vote Republican in statewide and local elections. These are Western Pennsylvanians who grew up in Democratic families, registered Democratic because of ties to the steel or coal or labor markets (or because they had to register as Democrats to get jobs), but grew up believing in God, country, and guns. They are leaning toward Trump but make up a small share of the overall vote in statewide elections. Trump will win these swing voters, but are there enough of them to make the difference? According to our review of publicly available polling, few national pollsters are tracking support with critical swing voters.
The second reason why Trump may not be winning Pennsylvania has to do with who is answering polls. We suspect because Trump is the only candidate with “enthusiastic” voters, it’s Trump voters in particular who are disproportionately talking to pollsters. It’s the reverse of what happened in 2016, when the phenomenon of “shy” Trump voters meant that many pollsters undercounted Trump’s base of support. Many voters were afraid to admit they were Trumpers back then. Today, we suspect many pollsters are not adjusting their samples to account for this “non-response” bias, as it’s typically called. But SP&R is doing so.
The last reason why Trump is probably not winning Pennsylvania has to do with geography. To win, Trump will need to overperform or at least match his 2016 margins in the conservative “T” — northeast and south central Pennsylvania — but not get murdered in the vote-rich southeast collar counties around Philadelphia. This was how Trump carried Pennsylvania in 2016.
But according to our latest poll, Trump is underperforming in Trump areas while losing even bigger to Biden in the Philadelphia suburbs which account for one in three votes cast (including the city) in statewide elections. Trump can’t thread the Pennsylvania needle if he doesn’t overperform or match his 2016 margins. He isn’t doing it so far. He’s instead underperforming in the conservative “T,” and also not matching his 2016 numbers in the south central/Harrisburg media market, which is trending less Republican in recent years.
If you question this, just look at Trump’s margins in 2020 versus Biden and compare his drop-off from his 2016 margins against then-candidate Hillary Clinton in these same areas.
It’s our belief that national pollsters polling Pennsylvania need to create hard and fast “area breaks” — groupings of interviews by county and media market — to measure Trump’s authentic support in the collar counties around Philadelphia and the other vastly different media markets of the Keystone State. There are six distinct media markets in Pennsylvania: Erie, Johnstown/Altoona, Pittsburgh, Harrisburg/Lancaster/York/Lebanon, Northeast/Wilkes-Barre/Scranton, and Philadelphia. Each has its own nuanced voting electorate and demographic makeup.
Many pollsters are simply not doing this in their rush to track all battleground states in a condensed timetable. Area breaks help control the flow of interviews in these regions and can be invaluable in tracking who is winning and losing, especially in a state like Pennsylvania.
The Biden campaign understands that to win Pennsylvania you need a surgical approach to both polling and getting out of the vote. Biden won the state in 2020 by 82,000 votes; he beat Trump in just 13 of the state’s 67 counties. These include Bucks, Philadelphia, Chester, Delaware, and Montgomery in the Southeast. Add to this Lehigh, Northampton, Lackawanna, and Monroe in the Northeast. Then add Allegheny, Erie, Dauphin, and Centre, and you have the magic formula.
Are national pollsters accurately tracking the sample of interviews in these vastly different regions of Pennsylvania? If they were, they’d be showing Trump losing to Biden in the southeast, and Trump underperforming in the areas Trump bested Clinton in 2016. These polls would most likely produce either a statistical tie (at best) or a narrow Biden lead at worst.
In short, we think Biden has the edge based on our current polling.
(Jim Lee is pollster and CEO at Susquehanna Polling & Research, Inc. This article was made available via RealClearWire.)
