(Repeats with no change in text)
* 52 percent of Republicans say Perry should stay in race
* Perry, Cain setbacks reshaping race
WASHINGTON (Reuters) - Republican U.S. presidential
candidate Rick Perry's disastrous debate performance last week
has hurt his campaign, but may not be fatal, a Reuters/Ipsos
poll found.
After watching a video of the Texas governor stumbling
while trying to name three government agencies he would
eliminate if elected, 48 percent of registered Republicans
thought less favorably of him, the poll found.
But while 31 percent of Republicans believed Perry should
drop out of the race, many more -- 52 percent -- felt he should
stay in.
Perry's embarrassing memory lapse further roiled a
Republican race already hit by the travails of Herman Cain, who
was accused of sexual harassment against four women. Among
those polled, 37 percent of Republicans thought Cain, who
denies the allegations, should leave the race.
The setbacks to Perry and Cain are reshaping the Republican
field of potential challengers to President Barack Obama in the
2012 election. Former Speaker of the House of Representatives
Newt Gingrich, once considered an also-ran, has benefited as
conservative Republicans cast about for an alternative to the
more moderate Mitt Romney.
The Reuters/Ipsos poll had Romney in the lead among
Republicans, with 28 percent saying they would vote for him.
Perry was fourth with the support of 12 percent, while Cain was
second at 20 percent and Gingrich third at 16 percent.
The Texas governor once sat atop national polls and led the
money race during the most recent reporting quarter, but a
series of stumbles in the debates has eaten into his backing.
His fundraising has slowed significantly, sources close to his
campaign say.
Perry's campaign went into damage control overdrive
Thursday, putting the governor front and center on news shows
and allowing him to mock himself in an appearance on comedian
David Letterman's late-night talk show.
The online survey was taken among 461 registered Republican
voters Thursday and Friday.
Since it was an online poll, typical margins of error do
not apply. Despite that, various recognized methods were used
to select as representative a sample as possible and weigh the
results. If it were a traditional random survey, it would have
a margin of error of 4.5 percentage points.
(Reporting by Mark Felsenthal; Editing by Peter Cooney)
(mark.felsenthal@thomsonreuters.com; +1-202-898-8329; Reuters
Messaging: mark.felsenthal.reuters.com@reuters.net))
© 2013 Thomson/Reuters. All rights reserved.