Professor Uses Engineering Math Model to Forecast Election
（CSUF News Service，Nov. 3, 2016）
（Chandrasekhar Putcha, Cal State Fullerton professor emeritus
of civil and environmental engineering）
In the tight race for the White House, Cal State Fullerton
professor Chandrasekhar Putcha predicts that Democratic nominee Hillary Clinton
will become the next U.S. president.
Using an engineering mathematical model he created, Putcha
pegs Clinton with 52.78 percent and 284 electoral votes and Republican
candidate Donald Trump with 47.21 percent and 254 electoral votes — a 1.9
percent margin of error. A candidate needs at least 270 electoral votes to win
the presidency.
Putcha predicts Clinton will take the lead in winning 43.63
percent of the popular vote, with Trump capturing 42.52 percent of the popular
vote — a probability with only a .0001 percent margin of error, according to
Putcha, who announced his prediction today (Nov. 3).
With presidential election polls in recent days predicting
Trump in the lead in some states and Clinton ahead in others, Putcha
recalculated the numbers one more time on Monday. While Clinton's lead dipped
from earlier calculations, the probability was the same — Clinton will clinch
victory in both the Electoral College and popular vote.
This isn't the first time the professor has put his math
skills to the test to predict the outcome of a presidential election. Putcha is
two-for-two: He correctly predicted that Barack Obama would be elected in 2008
and 2012. How? By focusing on the statistics (polling data) and the
probability.
Putcha, professor emeritus of civil and environmental
engineering, applied probability and statistics — the same principles he uses
in engineering analysis. His primary research over his 35-year career focuses
on reliability and risk analysis.
For the election prediction, Putcha collaborated with one
his former students, 2016 CSUF graduate Vineet Penumarthy, who earned his
bachelor's degree in mechanical engineering in May, and Brian Sloboda,
associate chair of the Center for Management and Entrepreneurship at the
University of Phoenix. Putcha has worked previously with both on joint research
efforts.
Putcha and his team used state-by-state polling data from
various sources. For example, in California, polling data from KABC-TV Los
Angeles/SurveyUSA, USC Dornsife/Los Angeles Times Poll, Insights West poll and
the Field Poll were used to predict a 98.58 percent probability that Clinton
will win the state, with 57 percent of the popular vote and all 55 electoral
votes.
For North Carolina, the research team used such sources as
Public Policy Polling and New York Times Upshot/Sienna College. In that state,
Putcha is predicting Clinton to win with 48 percent of the popular vote,
compared to Trump's 44.50 percent.
In Florida, Putcha predicts Trump will prevail by a slim
margin of 46.33 percent, compared to Clinton's 46 percent of the popular vote,
capturing 29 electoral votes. The prediction also indicates that Trump will win
states such as Arizona, Ohio, South Carolina and Utah.
"The assumption is based on the premise that the polls
will capture the pulse of the people as to how they are planning to vote in the
actual presidential election," Putcha said. "If the polling data is
correct, the mathematical model that uses this information should predict
correctly the outcome of the election."
CSUF alumnus Penumarthy collected polling data and calculated
the probabilities of the winner from each state based on the polling data.
"Polls can be thought of as reactions of the people to
any particular event. Using these polls at the state level meant that we could
individually look at the reactions of the people in each state and then predict
the winner based on these responses," Penumarthy explained.
Now a graduate student in mechanical engineering at the
Illinois Institute of Technology, Penumarthy wanted to be involved in Putcha's
research because of the "sheer bizarreness of the election."
"I got on board partially due to my interest in
statistical analysis, but mostly because I wanted to help predict the winner of
the election by just focusing on the mathematics."
For Putcha, CSUF's 2007 Outstanding Professor, it has become
a challenge to forecast the next U.S. president, and he is looking forward to a
three-peat.
"Anybody can always predict, even by simply tossing a
coin, and the person may be correct. But there is no mathematical basis for the
outcome. I'm confident that our mathematical analysis is accurate in showing
that Clinton will win."
