"Prediction is very difficult, especially about the future." - Niels Bohr.
Richard Nixon was involved in two particularly close elections. In 1960, Nixon lost to John Kennedy by the razor thin margin of 118,000 thousand votes out of over 78 million cast. Had not Nixon lost precious campaign time due to a minor car accident or if Mayor Richard Daly of Chicago had not been so effective in delivering a few crucial precincts and hence Illinois to Kennedy, Nixon might have won both the popular vote and the vote in the Electoral College.
In 1968, Nixon squeaked by Hubert Humphrey by 500,000 votes out of the 72 million cast. In the waning days of the campaign, the polls were rapidly shifting in favor of Humphrey. If the election were held a couple of weeks later or if George Wallace did not pull 9.9 million votes, many from Democratic strongholds in the South, Richard Nixon might have never served in office.
If Nixon had never become president, there would have been no Vice-President Spiro Agnew who was convicted of accepting kickbacks while governor of Maryland and ultimately there would have been no President Gerald Ford. If there had been no Gerald Ford, there would have been no pardon of Nixon and no narrow victory by Jimmy Carter of Ford in 1976 with a 37-vote margin in the Electoral College. It is in this context, of rapidly changing events influenced by unexpected and unrepeatable circumstances, that some political scientists claim to be able to predict presidential elections.
These models are based on a number of parameters that one would reasonably believe are related to election outcomes: the economic trends in the country, the favorability ratings of the incumbent, whether one party has served in office over a couple of election cycles, whether polls indicate people feel better off than they used to. However, there have only been 13 presidential elections since 1948 when much of the background data have been available and given enough input parameters it is always possible to construct a model to "hindcast" election results. Hindcast is a word used by meteorologists, who try to understand and deal with random rapidly changing weather, when they try to reconstruct events and explain observed behavior. Hindcasts are always better than forecasts, but this nuance is lost when modelers of presidential election outcomes speak about ``predictions.'' Their predictions are really hindcasts, made after the fact. There is an old saying in science, "give me enough parameters and I can fit an elephant [or in some cases a donkey]" and it seems that modelers of presidential forecasts have done just that.
Go To Page: 1 2
| Here's the follow-up discussion on this article: | View all related messages |
For a complete listing of article comments, questions, and other discussions related to Frank Monaldo's Conservative Politics topic, please visit the Discussions page.