A U.S. firm's exit poll that said Venezuelan President Hugo Chavez (search) would lose a recall referendum has landed in the center of a controversy following his resounding victory.

"Exit Poll Results Show Major Defeat for Chavez," the survey, conducted by Penn, Schoen & Berland Associates (search), asserted even as Sunday's voting was still on. But in fact, the opposite was true — Chavez ended up trouncing his enemies and capturing 59 percent of the vote.

Any casual observer of the 2000 U.S. presidential elections knows exit polls can at times be unreliable. But the poll has become an issue here because the opposition, which mounted the drive to force the leftist leader from office, insists it shows the results from the vote itself were fraudulent. The opposition also claims electronic voting machines were rigged, but has provided no evidence.

Election officials banned publication or broadcast of any exit polls during the historic vote on whether to oust Chavez, a populist who has sought to help the poor and is reviled by the wealthy, who accuse him of stoking class divisions.

But results of the Penn, Schoen & Berland survey were sent out by fax and e-mail to media outlets and opposition offices more than four hours before polls closed. It predicted just the opposite of what happened, saying 59 percent had voted in favor of recalling Chavez.

Cesar Gaviria, secretary general of the Organization of American States (search) who monitored the referendum, said the poll must have had a tremendous impact on Chavez's opponents, who felt they were about to complete their two-year drive to oust him.

"They were told they had a lead of 20 points and then when the results came, they lost by 20 points," Gaviria said. "It's very difficult to deal with that."

Both Gaviria and former President Jimmy Carter (search), another election monitor, endorsed the vote, saying the results coincided with their own independent samplings.

Mark Penn, of Penn, Schoen & Berland Associates, said Wednesday he has limited knowledge of the exit poll. He said his partner, Doug Schoen, "believes there were more problems with the voting than with the exit poll."

Schoen could not immediately be reached, and another employee familiar with the poll declined to comment.

"We have to let the authorities do their investigation of the election," said Marcela Berland, with the firm. "It would be irresponsible to interfere with that."

Critics of the exit poll have questioned how it was conducted because officials have said Penn, Schoen & Berland worked with a U.S.-funded Venezuela group that the Chavez government considers hostile.

Penn, Schoen & Berland had members of Sumate (search), a Venezuelan group that helped organize the recall initiative, do the fieldwork for the poll, election observers said.

Roberto Abdul, a Sumate official, acknowledged in a telephone interview that the firm "supervised" an exit poll carried out by Sumate. Abdul added that at least five exit polls were completed for the opposition, with all pointing to a Chavez victory.

Abdul said Sumate — which has received a $53,400 grant from the National Endowment for Democracy (search), which in turn receives funds from the U.S. Congress — did not use any of those funds to pay for the surveys.

The issue is potentially explosive because even before the referendum, Chavez himself cited Washington's funding of Sumate as evidence that the Bush administration was financing efforts to oust him — an allegation U.S. officials deny.

Venezuelan Minister of Communications Jesse Chacon said it was a mistake for Sumate to be involved in the exit poll because it might have skewed the results.

"If you use an activist as a pollster, he will eventually begin to act like an activist," Chacon told The Associated Press.

Chris Sabatini, senior program officer for the National Endowment for Democracy, defended Sumate as "independent and impartial."

"Exit polls are notoriously unreliable," Sabatini said by telephone from Washington. "Just because they're off doesn't mean that the group that conducted them is partial to one side."