Several polls say Trump is winning, others say Clinton is ahead.

An analysis of states across the country is also projecting one candidate over the other, but it's taking a different approach. A model called the ‘Moody’s Analytics Model’ has successfully predicted the winner of every presidential race since Ronald Reagan.

It’s not a poll, it’s more an analysis of certain patterns that economists plug into a model to determine the chances of a candidate from the incumbent party being re-elected.

The economists behind the analysis use six factors, three political and three economic, determined for each state. Dan White, a senior economist for Moody’s says factors include: the percentage of a state that voted for each party in the previous election, which party is currently in office and how many terms they’ve served, and the incumbent president’s approval rating.

Economic factors including household incomes, house prices, and gas prices are also factored in. Each factor is plugged into a model and that model forecasts the likelihood of an incumbent party being re-elected.

When all the data was plugged into the model this year, all factors pointed to a landslide win for Hillary Clinton. Moody's analysis predicts she will win with 332 electoral votes and Donald Trump will lose with 206 electoral votes.

White says though the model has been proven to forecast the winner for several years, this year could be different because of the uniqueness of the election.

“Our model only forecasts whether or not an incumbent party will be re-elected, it doesn’t say anything about the specific personalities running for president,” White says “in a normal election year our model says a generic democratic candidate should beat a generic republican candidate, but this is certainly not a year where we have a generic republican or a generic democratic candidate.” He says the different personalities of each candidate are the biggest risk factor with the analysis during this election cycle.

You can find White’s analysis here.