In 2004, John Kerry appeared to be a clear winner based on the exit polls. But the recorded vote counts deviated sharply from the polls and were too one-sided to attribute to mere chance. The political pundits claimed the polls "behaved badly", calling the three million Bush margin a "mandate". They failed to report the impossible late adjustments made to the National, Florida and Ohio exit polls that were necessary to force them to match the recorded vote.
Spreadsheet-wielding Internet bloggers analyzed the statistical anomalies. Online election forums were aflame with debates about state and national pre-election and exit polls, Bush approval ratings, margins of error, non-response, past vote recall, correlation between vote swing from 2000 and the exit poll shift, the counter-intuitive Bush gain over his 2000 vote share in urban locations compared to the decline in rural areas (the "Urban Legend").
But the "Smoking Gun" was the Final National Exit Poll, which indicated that there were four million more returning Bush 2000 voters than were alive in 2004. This "phantom voter" anomaly has never been discussed by pundits in the media or political scientists in academia, even though similar anomalies occurred in 1988, 1992 and 2008. The media pundits still claim that the recorded vote is correct (and the exit polls wrong) despite millions of uncounted votes in every election.
This book is a comprehensive resource for analyzing presidential elections from 1968 to 2008, including the 2006 midterms. It is written for readers of virtually all backgrounds. The only requirement is an inquisitive, open mind. The True Vote is estimated using basic statistical modeling that is for some reason avoided in the media and academia. Internet links to several election analysis spreadsheet models are provided in the book and are free to download.