History

In 1960, John F. Kennedy won Illinois by less than 10,000 votes.  He would have had an Electoral College majority even without Illinois, but he won Missouri by about 10,000 and New Jersey by only about 20,000, and the margins were even smaller in several smaller states.  So there are several ways he could have lost the election given small shifts in votes.  Unsurprisingly, there were charges that the election had been stolen.  Richard Nixon and the Republican party leadership accepted the results, but there were some legal challenges by private citizens and local officials.  The Gallup Poll didn't ask any questions about allegations of fraud, but they did have an allusion to the controversy just before the Electoral College vote:  "Suppose that Kennedy is unable to get a majority of votes in the Electoral College when it meets December 19 and the question of who is to be President has to be turned over to the House of Representatives.  Which man would you like to have the House declare the winner--Kennedy or Nixon?"*  55% said Kennedy, 35% Nixon, 9% don't know.  Supposing that 50% voted for Kennedy and 50% Nixon (which was approximately true in the population, but I don't know about recalled vote in this sample), and all the Kennedy voters said it should be Kennedy, that would mean that 70% of the Nixon voters said Nixon, 10% Kennedy, and 18% were unsure.  So only a small minority went with the principle that it should select the winner of the popular vote rather than the candidate they preferred.  On the other hand, you could say that almost 30% of Nixon voters didn't just stick with their man.  Also, Kennedy's lead in the popular vote was very small (plus, there's a complicated question of how you count the vote in Alabama, where some Democratic electors were pledged to Kennedy and other's were not).

So this example suggests that the principle that whatever was good for your candidate was right was widely followed, even at a time when party differences were small.  However, I think the most interesting thing is that there were no survey questions about the allegations of voter fraud.  Gallup had asked a question on the issue in 1959, so it would seem to have been an excellent opportunity to repeat it (they did, a few years later).  The omission doesn't seem to have been because there was a lot of other stuff going on in the news--the December survey included a question on "what is your favorite season of the year." So my guess is that they just regarded it as inappropriate under the circumstances. 

The 1968 and 1976 elections were close, but I don't think that there were any significant claims of fraud.  The next disputed election was 2000, when the issue wasn't fraud, but how long the counting and recounting should continue.  I'll write about opinions on that election that in a future post. 

*Apparently there was an effort to get some southern Democrats to defect so that they could deny Kennedy a majority and then bargain with the candidates to drop any civil rights programs.