Donald Trump, master showman, took the usual Kabuki of selecting a vice presidential running mate and blew it up into something like
Former House Speaker Newt Gingrich, Indiana Gov. Mike Pence, New Jersey Gov. Christie and others all had tryouts on the road on live TV and multiple meetings with the presumptive GOP presidential nominee and endured leaks and hints about where Trump was leaning.
Who would be left standing to hear, "You're hired"?
Trump picked Pence, an even-keeled Midwesterner with congressional experience and appeal to the religious right, but there was confusion at the end of the show. After the terrorist attack in France, Trump postponed Friday's formal rollout and even said he had not made a "final, final" choice. Then, Friday morning, he tweeted the news.
Yet one thing has been conventional. As the search narrowed, the real estate mogul and his allies, accompanied by a chorus of pundits, applied the usual strategic considerations to the contenders, speculating, for instance, that this one or that one might help Trump carry an important state, put a noncompetitive state into play, or boost him with an important demographic.
Political science, however, tells us that vice presidential candidates rarely make a decisive difference.
Voters in a VP candidate's home state were no more likely to support that candidate's ticket than voters elsewhere, discovered political scientists Kyle C. Kopko of Elizabethtown (Pa.) College and Christopher Devine of the University of Dayton in Ohio.
And, they found, election returns showed that those home states did not deviate significantly from their historic partisan voting patterns - while presidential candidates enjoyed a home-state advantage of 3 to 7 percentage points. The scholars ran the statistical traps on more than a hundred years of state-level election results, as well as hundreds of thousands of individual voter interviews conducted from 1952 to 2008, lodged in the massive American National Election Studies (ANES) database.
"Political implications for VP candidates are often overstated, not as great as a lot of folks think," Kopko said, noting that the research in the ANES found voter perceptions of a presidential nominee carried three times more weight in their decisions that did opinions of a vice presidential candidate.
"If you had a running mate who was exceptionally popular or unpopular relative to the presidential candidate, that could make a difference," Kopko said. "That typically doesn't happen because campaigns take a 'first do no harm' approach."
Belief in the home state effect of veeps dies hard. In 2012, for example, both sides largely ignored Wisconsin, but when Republican Mitt Romney picked Rep. Paul Ryan (R., Wis.) as his running mate, the state was targeted. By the end of the campaign, it was the fifth most popular state for candidate visits and ad spending - despite being only the 10th most competitive in terms of the results.
"That says to me that the people who run these campaigns really believed that Romney's chances to carry Wisconsin were improved by his choice of Ryan," Devine said. "Political people are hardwired to think that way."
The research found that running mates can and do make a difference in their home states in limited circumstances: if they've been a political figure a long time and the state is small. As Obama's 2008 running mate, Joe Biden overperformed in Delaware, where he was a senator for three decades. And in 1968, Edmund Muskie helped the Democrats win Maine, where he was a senator and former governor.
Trump himself seemed to recognize the folly of deciding on the basis of geographic and demographic factors.
"History has said nobody ever helps," he said Tuesday. "I've never seen anybody that helps."
Except, he said, everyone knows that John F. Kennedy's choice of Texas Sen. Lyndon B. Johnson delivered Texas to the Democrats in 1960, helping win a narrow victory.
Turns out that is a myth, too.
Kennedy was more popular than Johnson in Texas and in heavily Catholic Louisiana, as well as throughout the rest of the former Confederacy, the NAES and archived internal campaign polls showed, Kopko said.
Johnson, he said, was a typical conservative Democrat and segregationist but came to be reviled in many quarters of the South after he became Senate majority leader in 1957 and pushed through an early civil rights bill.
Recently, the researchers tested whether running mates could "deliver" votes from a targeted demographic group, if they were from the same group. In 1984 (Geraldine Ferraro) and 2008 (Sarah Palin) did not bring a big bump in female voters for their tickets. The same was true of Catholics in 1972, 1984, 2008, and 2012. And in 2000, Jewish voters did not vote for the Democratic ticket in greater than normal numbers because Sen. Joseph Lieberman (D., Conn.) was on the ticket.
Myth busting aside, Kopko and Devine are not saying that a presidential candidate's choice means nothing. The decision sends messages that may not be captured in statistical analysis.
"When a vice president helps, it's in the margin and in a more subtle way," agreed another scholar of VPs, Joel Goldstein of St. Louis University Law School. "It can influence how people view the decision maker, the presidential candidate." (For instance, he said, running mate Walter Mondale probably helped Jimmy Carter in several industrial states with more liberal electorates in 1976.)
Some wonder if this is the year a No. 2 pick could make a real difference. After all, Trump and Democrat Hillary Clinton both have historically high unfavorable ratings - in polls, majorities of voters say that she is untrustworthy and that he is unqualified to be president. Could the one who picks the more popular running mate could have an advantage?
Probably not, the available data suggest. Then again, the 2016 campaign already has been a long, strange trip.