Shy Elephant Factor
In other countries, the term can be closely aligned with the British "Shy Tory Factor." "Shy Tory Factor" is a name given by British opinion polling firms to a phenomenon observed in the 1990s where the share of the vote won by the British Conservative Party, known as the "Tories," in elections was substantially higher than the proportion of people in exit and opinion polls who said they would vote for the party.
The elephant in the term is used to align conservative voters with the closest related U.S. Republican Party. The official symbol of the Republican party is the Elephant.
Statistics and the 2004 presidential election
Following the 2004 election, the National Election Pool (NEP) discovered a discrepency in exit poll results vs. the actual popular vote results. The national sample showed Kerry ahead 51% to 48%, but Bush won the national popular vote by a 2.5% margin (50.7% for Bush and 48.3% for Kerry). On average, the statewide exit polls showed a similar overstatement.
Warren Mistofsky, who headed the 2004 exit polls, appeared following the election and offered a theory: "We suspect that the main reason was that the Kerry voters were more anxious to participate in our exit polls than the Bush voters." Three months later, the report issued by Mitofsky and his partner Joe Lenski again argued that the discrepancy occurred because "Kerry voters were more likely to participate in the exit polls than Bush voters." They also offered "hypothetical completion rates of 56% among Kerry voters and 50% among Bush voters" that would have accounted for the entire discrepancy. This theory was unofficially coined the "reluctant Bush responder" hypothesis.
Years following the Mistofsky-Lenski "reluctant Bush responder" hypothesis, studies by various polling firms have found data that confirms that Democratic or Liberal voters are much more likely to respond to an exit poll than Republican or Conservative leaning voters. More Democrats (72%) than Republicans (66%) said they were likely to fill out an exit poll questionnaire. The gap was far bigger -- and highly statistically significant -- among those who felt strongly (typically a better predictor of actual behavior): 44% of Democrats said they would be "very likely" to participate in an exit poll compared to only 35% of Republicans.