Aside from the polls, which can vary widely depending on the voter models and pollster weighting, we can also look at election models that create forecasts based on various data points. There are a handful of models, mostly produced by college professors or other academics which attempt to predict presidential elections based on a variety of factors.

Public polling is just one small part of the model. More important data points, such as the state of the economy, presidential approval, and strenth during primary voting, are weighted much heavier in these models compared to opinion polling.

One of the most lauded and interesting election models is that of Stony Brook University Professor Helmut Norpoth. The Primary Model produced by Norpoth in 1996 has been successfully back-tested against every election since 1912, with few exceptions. The model also predicted Donald Trump’s victory over Hillary Clinton during a time when polling data suggested Clinton would have an easy win.

Norpoth is now out with new numbers suggesting that his model gives President Trump a 91% chance of being re-elected in November:

The model has a laudable track record. Established in 1996 by Stony Brook University political science professor Helmut Norpoth, the Primary Model correctly picked the victors in multiple presidential elections — including the last one.

On March 7, 2016, it predicted then-candidate Donald Trump had an 87% chance of defeating Hillary Clinton. Will the president also vanquish Democratic nominee Joseph R. Biden?

“The Primary Model gives Trump a 91% chance of winning in November. This model has picked the winner in all but two elections since 1912, when primaries were introduced, including, of course, Trump’s victory in 2016,” Mr. Norpoth tells Inside the Beltway in a statement.

“As for polls showing Trump trailing Joe Biden right now, remember 2016. Polls and poll-based forecasts all handed Clinton a certain victory. But this is not the only failure,” he continues.

Models can be terribly wrong, of course, and Norpoth admits that issues such as the Covid-19 pandemic could corrupt his data and/or cause a situation where no amount of other factors, such as the economy, could produce historically expected election results as they have in the past.

Digging deeper into Norpoth’s model, which is reliant heavily on data rather than forecasts or predictions, Donald Trump’s performance in several uncontested and meaningless 2020 Republican primaries in the Spring offers a glimpse at voter enthusiasm:

“I don’t remember any uncontested primary campaign for a sitting president like this,” Stony Brook University professor Helmut Norpoth told The Post. “People don’t normally turn out in large numbers for a race without a challenger.”

But in 23 of the 27 states that held primaries both this year and in 2012, when President Obama ran for re-election, Trump has racked up higher raw vote totals than Obama did — often doubling or tripling his predecessor’s numbers. His primary vote totals also beat President George W. Bush’s in 2004.

The high Trump turnout showed up not only in reliably Republican states like Montana and Arkansas, but even in deep-blue strongholds and in purple states that Joe Biden hopes to flip come November.

In Michigan’s March primary, Trump’s 639,144 votes dwarfed the 174,054 that Obama notched in 2012. A few weeks later, Trump scored 713,546 votes in Ohio — more than twice Obama’s 2012 total of 285,990.

Trump’s primary numbers are impressive, especially given that some of the primaries took place in states and localities dealing with rising Coronavirus infections which may have deterred some voters. The desire for the President’s base to come out and support him, in what amounts to a beauty contest, is part of Norpoth’s modeling. Regardless of what opinion polls say, it’s the actual bodies in the voting booth that matter in November.

Back to Norpoth’s 2020 model, what does the professor say about Joe Biden’s sizable and almost insurmountable leads in some national and state polls?

As a historical exercise, the model also forecast past bouts.

“The terrain of presidential contests is littered with nominees who saw a poll lead in the spring turn to dust in the fall. The list is long and discouraging for early front-runners. Beginning with Thomas Dewey in 1948, it spans such notables as Richard Nixon in 1960, Jimmy Carter in 1980, Michael Dukakis in 1988, George H.W. Bush in 1992, and John Kerry in 2004 — to cite just the most spectacular cases,” Mr. Norpoth advises.

The rise and fall of opinion polling leads spans across both parties. Republican and Democratic candidates have seen eye-popping numbers in the Spring and Summer, only to watch them evaporate in the Fall as Election Day draws closer. None of this means that Joe Biden will suffer a similar fate, perhaps his numbers continue to increase as Trump’s numbers soften into the Summer months.

Looking at the map of opinion polls, there isn’t much good news for the Trump campaign virtually anywhere. States like Wisconsin, Michigan, and Pennsylvania, which Trump won in 2016, are trending for Biden in 2020.

If the election were held today, or this week, Trump could be in serious trouble. His base appears less enthused than 2016, moderates seem to have soured, and Democrats are energized to vote against him. Unfortunately for Joe Biden, there are still weeks and months to go, and Norpoth urges caution to voters on either side drawing conclusions at this point on the calendar:

The message from one political expert who predicted Donald Trump’s 2016 victory: it is far too early to make conclusions based on random polls. “At this time of the season, you have about as many times as not that somebody leading in the spring is going to lose in November,” says Dr. Helmut Norpoth, political science professor at Stony Brook University. “So (based on polls) it’s pretty much a 50-50 tossup right now.”

The bottom line? Leading all the polls during the Spring months does not always equate to winning big in November. This is a model, of course, not an exact science, so the model could be wrong and Trump could easily lose in November and buck the trend of Norpoth’s accuracy.

In the meantime, however, Norpoth’s model is another data point to examine in addition to the daily deluge of polls showing Biden leading around the country.