WASHINGTON - A vote-by-vote review of untallied ballots in the 2000 Florida presidential election indicates George W. Bush would have narrowly prevailed in the partial recounts sought by Al Gore, but Gore might have reversed the outcome - by the barest of margins - had he pursued and gained a complete statewide recount.

Bush eventually won Florida, and thus the White House, by 537 votes out of more than 6 million cast. But questions about the uncounted votes lingered.

Almost a year after that cliffhanger conclusion, a media-sponsored review of the more than 175,000 disputed ballots underscored that the prize of the U.S. presidency came down to an almost unimaginably small number of votes.

The new data, compiled by The Associated Press and seven other news organizations, also suggested that Gore followed a legal strategy after Election Day that would have led to defeat even if it had not been rejected by the U.S. Supreme Court. Gore sought a recount of a relatively small portion of the state's disputed ballots while the review indicates his only chance lay in a course he advocated publicly but did not pursue in court - a full statewide recount of all Florida's untallied votes.

"We are a nation of laws and the presidential election of 2000 is over,"Gore said yesterday in a prepared statement. "Right now, our country faces a great challenge as we seek to successfully combat terrorism. I fully support President Bush's efforts to achieve that goal."

Said Bush press secretary Ari Fleischer: "The election was settled a year ago, President Bush won and the voters have long since moved on."

Against the backdrop of the disputed Nov. 7, 2000, election, the news organizations set out earlier this year to examine as many as possible of the ballots set aside as either undervotes or overvotes. Undervotes involved about 62,000 ballots where voting machines were unable to detect a choice for any presidential candidate, while about 113,000 overvotes were read by machines as possibly containing more than one choice.

The goal of the news organizations was not to learn who really "won" Florida; the Electoral College already had determined Bush was the winner following a U.S. Supreme Court decision that ended further counting and led to Gore's concession.

The aim was to provide a valuable historical record by thoroughly assessing tens of thousands of ballots that no one had fully examined.

Much of the legal wrangling focused on how votes were defined, and the ballot review did, too, calculating results under different standards - for example, whether to count as votes "hanging chads" on punch-card ballots or ballots marked with an "X" instead of the required filled-in oval on optical scan ballots.

Completing two partial recounts that Gore unsuccessfully pursued in court showed Bush maintaining a lead ranging between 225 and 493 votes - meaning Bush still would have won if the U.S. Supreme Court had allowed a partial statewide recount to continue.

Under any standard that tabulated all disputed votes statewide, however, Gore erased Bush's advantage and emerged with a tiny lead that ranged from 42 to 171 votes.

Strikingly, all these outcomes were closer than even the narrow 537 votes of Bush's official victory. With numbers that tiny, experts said it would be impossible to interpret the survey results as definitive.

Under the most inclusive standards, the study showed up to 24,653 potentially salvageable overvotes and undervotes among the 12 presidential candidates who ran in Florida.

The Florida election review was developed by the AP, CNN, The New York

The Wall Street Journal and The Washington Post. Tribune newspapers include the Chicago Tribune, Los Angeles Times, Newsday, the Orlando Sentinel and the South Florida Sun-Sentinel in Ft. Lauderdale.

This media consortium hired the National Opinion Research Center at the

University of Chicago to view each untallied ballot and gather information about how it was marked. Founded in 1941, the center is a not-for-profit corporation whose charter is to advance the methodology of public opinion surveys and provide accurate survey data.