Washington (CNN) - A new poll suggests that GOP Senator Orrin Hatch of Utah may be looking over his shoulder for a 2012 challenge, and not just from the Democrats.
A Salt Lake Tribune/Mason-Dixon survey released Monday indicates that a plurality of likely voters in Utah favor electing a fresh face for Utah's U.S. Senate seat in 2012.
In a poll conducted one week before the 2010 midterm elections, 48 percent of likely voters would favor another candidate if the 2012 elections were held today. Forty percent said they would re-elect Hatch and 12 percent were unsure, though they have two years to make a decision.
Earlier this year three-term Sen. Bob Bennett, a fellow Republican who supported the Troubled Asset Relief Program (TARP) and sponsored a bi-partisan heath care reform bill, was targeted by Tea Party activists and was defeated in his bid for re-election. At the Utah GOP state convention he lost his bid for his party's re-nomination at the hands of Tea Party conservatives, who eventually backed conservative Mike Lee, who subsequently won the GOP primary and then the general election.
According to an April Mason-Dixon poll, 19 percent of Republican delegates said they'd support a seventh term for Hatch, but a majority–71 percent– said they'd favor another candidate for Senate. Hatch has a history of reaching across the aisle to work with Democrats, which could spell trouble for 2012. The most recent Salt Lake Tribune/Mason-Dixon poll indicates that 60 percent of Republican likely voters would support Hatch's bid for re-election, though it did not gauge his chances in a contest with any other candidate.
If Hatch does survive any possible primary challenge, he could face off in the general election against Democratic Representative Jim Matheson, who considered a run this year against Bennett before winning re-election to a sixth House term. Matheson may look to challenge Hatch in 2012.
The Salt Lake Tribune poll of 625 registered Utah voters was conducted by Mason-Dixon Polling & Research from October 25-27. It has a sampling error of plus-or-minus four percentage points.