What matters in a presidential election, and what will matter in the 2016 election?

This simple question has produced a virtual cornucopia of theories, suppositions and hypothesis, all developed to explain why one party’s candidate is elected while the other party’s candidate is rejected.

Prominent among these include the state of the economy, foreign policy crises, job performance ratings, the relative strength of the two major parties, nomination battles, incumbency, and much more.

What has not been studied is the personal popularity of candidates.  Does popularity really matter?

This question looms acutely compelling this year since both candidates, measured by the polling standard “favorable or unfavorable,” are both more unpopular than popular.

Hillary Clinton and Donald Trump’s dual unpopularity is historically unprecedented. Voters in 2016 will be asked to make a choice between two candidates, both of which they dislike.

The poll numbers are unequivocal. The Real Clear Politics national average for Clinton shows a 15-point plus gap between her “favorables” and “unfavorables.” By any standard she is very unpopular with voters, an unpopularity almost certainly linked to deep voter skepticism about her honesty and trustfulness.

Trump is in even more historic territory. He is viewed negatively by more than six of every ten Americans. The yawning gap between his unpopularity and popularity is almost 30 percent. Trump is viewed unfavorably by seven of every ten women and almost nine of every ten African Americans.

So, we have a paradox here. For the first time in modern history both major candidates are far more unpopular than they are popular. Yet one of them is going to win the election.

How has candidate popularity impacted past presidential elections? We can answer that question using a combination of state (Pennsylvania) and national data stretching back a quarter century.

If popularity or “favorables” are important then an examination of favorable scores in recent presidential elections might offer a clue as to the likely winner in the competition for the Keystone state’s 20 electoral votes.

Looking at the first four presidential elections of the 21^st century in Pennsylvania back to 2000, what does candidate popularity show?  The final Franklin & Marshall poll taken before Election Day for all four elections reveals that the most popular candidate carried the state in each of these elections.

1. In 2000, Al Gore won the state with favorables of 45 percent over George W. Bush’s 43 percent, winning by 4 points.

2. In 2004, John Kerry won the state with favorables of 48 percent over Bush’s 42 percent, winning by 2.5 points.

3. In 2008, Barack Obama won the state with favorables of 53 percent over John McCain’s 45 percent, winning by 10 points.

4. In 2012, Obama won the state with favorables of 50 percent over Mitt Romney’s 43 percent, winning by 5 points.

In general, the higher a candidates favorables the wider the victory. And the closer in popularity candidates are to each other, the closer the election.

So, we see consistent outcomes in one state tied to popularity. The most popular candidate always won and the closer the candidate’s popularity the closer the election.

These may tell us something important about Pennsylvania. But what does comparable national data show about national results?

Indeed, quite a bit!

In fact, national presidential election data back to 1992 tells the same tale: popularity matters.

In 1992, Bill Clinton won nationally with 58 percent favorables to George H. W. Bush’s 47 percent favorable.

In 1996, Clinton won nationally with 56 percent favorables to Bob Dole’s 50 percent favorable.

In 2000, Gore won nationally (The popular vote) with 56 percent favorable to Bush’s 55 percent favorable.

In 2004, George W. Bush won with 51 percent favorables to Kerry’s 52 percent, the only election where the less popular candidate won the popular vote albeit, popularity that year was a virtual wash.

In 2008, Obama won with 56 percent favorables to McCain’s  52 percent favorable.

In 2012, Obama won with 50 percent favorables to Romney’s 49 percent favorable.

These national results combined with state data confirm that popularity does forecast winners and losers of presidential elections – although a single election (2004) suggests the correlation may not be perfect when popularity scores are close – while 2000 remind us that national popularity can be eclipsed by machinations of the Electoral College.

What about 2016 when both candidates are likely to be more unpopular than popular on Election Day?

If past popular vote elections have been won by the most popular candidate, this one will be won by the least unpopular candidate. The contest will be won by what the late legendary media analyst Tony Schwartz called “the least objectionable politician.” Some Americans might update that term to the “least obnoxious politician” in the race.

Electing a president because he or she is least obnoxious fails to inspire. Worse, however, doing so means we are going to live through perhaps the most negative presidential election in American history – one in which behind a virtual blizzard of negative ads, both campaigns will depict their candidate as the best of the worse.

Our politics have now reached the point where voters must choose for president the least unpopular, and least objectionable. We are told that the best of the worse is the best we can do.

If it is, we need to stop using the phrase “popular vote” to describe our elections.

Maybe we should call it the “unpopular vote.”