Bounce or No Bounce ,Maybe even a False Bounce?

Trolling through the Archives at Gallup this week, I came across the records of their polling trends for past Presidential elections. It’s fascinating to read how many people are predicting bounces for Kerry and/or Bush, when the history might not say what they think. Using the Gallup Polls as reference, I looked at the elections beginning in 1944 (in 1936 and 1940, there were no polls before the conventions, so the effects of the conventions cannot be measured).  I got the dates of the conventions from InfoPlease.

In 1944, polls before both conventions had FDR ahead, 47-45.  They were unchanged after the conventions, no bounce.  

In 1948, before the Republican Convention, Dewey led Truman 49-38.  After the Republican Convention, Dewey led Truman 48-37, no bounce.  Before and after the Democratic Convention, Truman trailed Dewey 48-37, no bounce.  

In 1952, before both conventions, Eisenhower led Stevenson 59-31.  After the conventions, Eisenhower led Stevenson 50-43, a 21-point bounce for Stevenson, but Eisenhower won easily in November anyway.  

In 1956, before both conventions, Eisenhower led Stevenson 61-37 (!).  After the conventions, Eisenhower led Stevenson 54-41, a 13-point bounce for Stevenson, but again, in November Eisenhower routed Stevenson.  

In 1960, before both conventions, Kennedy led Nixon 50-44.  After the conventions, they were tied 47-47, a 6-point bounce for Nixon.  

After 1960, the Republicans and Democrats began to hold their conventions further apart by date, and there are then polls in between the two conventions.

In 1964, before the Republican Convention, Goldwater trailed Johnson 76-20 (!!!).  After the convention, Goldwater trailed Johnson 59-31, a 28-point bounce for Goldwater.  Before the Democratic Convention, Johnson led Goldwater 65-29.  After the convention, Johnson led Goldwater 62-32, for a loss of six points, a negative bounce.   

Note that in 1952, 1956, and 1964, where a candidate was far ahead of his opponent, the trend was to close the gap.

In 1968, before the Republican Convention, Nixon led Humphrey 40-38.  After the convention, Nixon led Humphrey 45-29, a 14-point bounce.  Before the Democratic Convention, Humphrey trailed Nixon 45-29.  After the convention, Humphrey trailed Nixon 43-28, a 1-point bounce.  Note that the 1968 campaign is the first instance where the Convention shows a candidate clearly pulling away from his opponent.  
 
In 1972, before the Democratic Convention, McGovern trailed Nixon 53-37.  After the convention, McGovern trailed Nixon 57-31, a ten-point loss, or negative bounce.  Before the Republican Convention, Nixon led McGovern 64-30 (!).  After the convention, Nixon led McGovern 61-33, for a 14-point convention bounce.  

In 1976, before the Democratic Convention, Carter led Ford 53-36.  After the convention, Carter led Ford 62-29, a 16-point bounce.  Before the Republican Convention, Ford trailed Carter 51-36.  After the convention, Ford trailed Carter 51-40, for a modest 5-point bounce.  The 1976 campaign was the first election where both parties clearly got a bounce from their convention, though Carter got much more than Ford did.  

In 1980, before the Republican Convention, Reagan led Carter 37-34.  After the convention, Reagan led Carter 45-29, for a 13-point bounce.  Before the Democratic Convention, Carter trailed Reagan 45-29.  After the convention, Carter trailed Reagan 39-38, for a 15-point bounce. 1980 is the first year where substantial bounces cancelled each other out in the main.  Note also that by the election, Reagan had regained a substantial lead.  

In 1984, before the Democratic Convention, Mondale trailed Reagan 53-39.  After the convention, Mondale trailed Reagan 53-41, for a 2-point bounce.  Before the Democratic Convention, Reagan led Mondale 52-41.  After the convention, Reagan led Mondale 56-37, for an 8-point bounce.  Note here, that the closure in August did not keep Reagan from destroying Mondale in November.  

In 1988, before the Democratic Convention, Dukakis led GHW Bush 47-41.  After the convention, Dukakis led GHW Bush 54-37, for an 11-point bounce.  Before the Republican Convention, GHW Bush trailed Dukakis 49-42.  After the convention, GHW Bush led Dukakis 48-44, for an 11-point bounce.  This is the first election where the convention bounces cancelled each other out, yet the lead changed sides.  

In 1992, before the Democratic election, Clinton led GHW Bush 56-36.  After the convention, Clinton led GHW Bush 56-37, for a net loss of one point.  Before the Republican Convention, GHW Bush trailed Clinton 54-39.  After the convention, GHW Bush trailed Clinton 51-42, for a 6-point bounce.  Bush was not, however, able to close ground after that point.  

In 1996, before the Republican Convention, Dole trailed Clinton 53-36.  After the convention, Dole trailed Clinton 53-37, giving Dole a 1-point bounce.  Before the Democratic Convention, Clinton led Dole 54-36.  After the convention, Clinton led Dole 53-37, for a net loss of 2 points.  That was not nearly enough to allow Dole to close in on Clinton.  

In 2000, before the Republican Convention, GW Bush trailed Gore 47-46.  After the convention, GW Bush still trailed Gore 47-46, no bounce at all.  Before the Democratic Convention, Gore led GW Bush 47-44.  After the convention, Gore led GW Bush 47-45, for a net loss of 1-point.  

In the fifteen elections starting with 1944, the Republicans have averaged 4.8% bounce from the convention, the Democrats have averaged 3.5%.  However, from this record, we can also see that while large bounces are possible, small bounces are as well, and so are no changes, and even losses.  Further, there does not appear to be a direct relation to gains from the party convention, and actual victory in the elections.  So, even a big bounce could end up being a false bounce.