A new poll from Mercyhurst Center for Applied Politics showed Governor Corbett with low approval ratings, outscored by each of the Democratic challengers for his seat.

A majority, 52%, disapprove of the job Corbett is doing as governor, and only 29% approve.

But by comparison, very few even know the Democrats who are running.

Majorities of voters don’t recognize the name Katie McGinty (79%), Rob McCord (56%), or Allyson Schwartz (76%). Among the candidates, McCord is viewed most favorably, with 34% (11% strong favorable, 23% somewhat favorable) having a favorable opinion, and only 4% (3% somewhat unfavorable, 1% strongly unfavorable) having an unfavorable opinion. An equal number of Erie County voters have a favorable opinion of McGinty (11%) and Schwartz (11%), and very small minorities have unfavorable opinions of the candidates (4% McGinty, 5% Schwartz).

Wolf surprisingly does not have the best name recognition in the pack despite his dominance over the television market, and 64% don’t recognize him.

On the issues:

When asked whether they favored allowing gay and lesbian couples to marry legally in Pennsylvania, 55% say they favor (34% strongly favor, 21% somewhat favor) the change and only 36% oppose (13% strongly oppose, 23% somewhat oppose).

On the issue of marijuana legalization, 75% (37% strongly favor, 28% somewhat favor) favor making it legal for patients to use marijuana that was prescribed by a doctor, while only 23% (11% strongly oppose, 12% somewhat oppose) oppose. This is a decrease from their poll released on February 28, 2014 found that a large majority of registered voters in Pennsylvania (85%) favor legalizing the use of marijuana for medicinal purposes.

Interviews were completed with 426 registered voters in Erie County, Pennsylvania. For a sample size of 426, there is a 95 percent probability that our survey results are within plus or minus 4.75 points (the margin of error) of the actual population distribution for any given question. For subsamples the margin of error is larger (depending on the size of the subsample). The data were weighted on age, race and gender to correct for minor discrepancies between the sample and population. Question order was randomized whenever possible. Due to rounding, column totals may not equal 100 percent.