Lots of readers of this happy Corner have emailed me on the subject. What have I learned? That we cannot say for certain that in 1960 Richard Nixon, and not John Kennedy, carried the popular vote. What can we say? That, no matter how thoroughly we examine the electoral records, we will never know for certain who did win the 1960 popular vote.

A thorough explanation from Fred Schwarz of American Heritage magazine:

In every state but three in 1960, you had a choice between Kennedy and Nixon (sometimes with a few minor-party candidates), and popular votes were counted as usual. In Louisiana and Mississippi, there was a three-way race: Kennedy, Nixon, or “uncommitted.” In Louisiana, Kennedy won, and in Mississippi, “uncommitted” won–but in both cases, people could vote for Kennedy (or Nixon) if they wanted, and their votes were recorded as Kennedy (or Nixon) votes. So there was no funny business in either of those states.
The problem occurs in Alabama. There, only two slates of electors were on the ballot (again excluding minor parties): One uncommitted and one for Nixon. “Uncommitted” won by 324,000 to 238,000, and of Alabama’s 11 electors, 6 voted for Byrd and 5 for Kennedy. So technically, Rakove [the Stanford history professor] is right–these 324,000 votes are usually counted as Kennedy votes, when in fact they were “uncommitted” votes. And since Kennedy won the election by only 120,000, you can call him a minority president.
Of course, if you want to be super technical, in Alabama and many other states, no one voted for either candidate but rather for electors for that candidate. Yes, this is an exceedingly minor quibble, and if the 11 Alabama electors had cast their votes for Kennedy, no one would object to counting the popular votes as Kennedy votes too. If only one elector had defected, as happened in Oklahoma, they would still have been counted as Kennedy votes. But since 6 of the Alabama electors voted for Byrd [the Virginia Dixiecrat], the question arises as to who should get credit for the 324,000 popular votes.
If you want to boost JFK’s total, you call them Kennedy votes and say the 6 electors were faithless, like the Oklahoma guy. If you want to boost Nixon, you say the “uncommitted” votes were for nobody and the Nixon votes were for Nixon. This would leave Nixon with a popular-vote majority of about 200,000….But you can still make a plausible case for counting those votes as Kennedy votes. And if you object that Kennedy’s name did not appear on the ballot in Alabama, well, neither did Nixon’s. So to be consistent, you would have to take away his 238,000 votes, and Kennedy squeaks ahead again.

And a nice summary statement by Professor Matt Franck, chairman of the department of political science at Radford University:

Now the margin for Kennedy nationally has long been said to have been about 113,000 votes nationally (the Clerk of the House gave him a margin of 119,000 in April 1961, but the lower figure is accepted today). What we would have to know in order to say that Kennedy actually received fewer votes nationwide than Nixon–and we simply cannot know it–is that at least 113,000 Alabama voters, or more than one-third of those voting Democratic, went to the polls and pulled the lever for the entire slate of eleven Democratic presidential electors while thinking “I want Byrd” (or at least “I don’t want Kennedy”) rather than “I want Kennedy.”
Since Alabama then, as now, used the winner-take-all method of allocating electors to candidates–the party winning the popular-vote plurality having all its electors seated in the college–there was only one way for either Kennedy voters or Byrd voters to get even one electoral vote out of the eleven the state had to cast, and that was to vote for the whole party ticket of electors, whether any presidential candidate’s name was on the ballot or not.
We know only that the six Alabama electors who had made no pledge to Kennedy voted for Byrd–probably their intent all along….Is it possible that more than 113,000 Alabama Democratic voters wanted Byrd over Kennedy? Yes. Do we know that for sure? No way….