Public Policy Polling just released new data from their most recent Pennsylvania Senate poll (1/3-5; 547 registered PA voters), but the small size of the sampling universe leads us to question the validity of the results.
The survey shows Sen. Bob Casey, Jr. (D) easily defeating all potential opponents including former Sen. Rick Santorum (R), the man he ousted from the seat in 2006. Though Santorum has made no overt move to seek a re-match, he actually polls the best in the field of potential Republican candidates. Mr. Santorum, however, is the only person in the Republican field who has significant statewide name identification.
What makes the poll suspect, however, is not that it shows Sen. Casey to be performing well — that’s believable, since Pennsylvania still favors Democrats in statewide races (the election of 2010 notwithstanding), and he has not been the focal point of any controversy or scandal during his first term in office. Rather, it is his potential opponents’ favorability scores that seem wildly out of whack.
According to the PPP poll, Casey would defeat Rep. Charlie Dent (R-PA-15) 51-31%; Rep. Jim Gerlach (R-PA-6) 49-33%; Santorum 48-41%; former Lt. Gov. Mark Schweiker 47-34%; and unknown attorney Mark Scaringi, the only announced Republican candidate, 50-27%. For an incumbent re-election race, especially in a situation where the senator is a member of the state’s majority party, these seem to be credible numbers. But, the depicted views of his opponents are not.
In our opinion, the poll’s accuracy factor seriously deteriorates when looking at the potential Casey opponents. Both Republican congressmen, Dent and Gerlach, score very poorly within this sampling universe. Dent gets an incredibly low 6:18% favorability rating, and Gerlach is only slightly better at 9:17%. The fact that 3/4 of the respondents haven’t heard of them is believable, but what could each have done to make them so unpopular?
The answer is nothing, hence, these numbers make little sense. The likely reason for the faulty results is that only 136 members of the polling universe could even identify them. In raw number terms, it is likely that only eight people said something positive about Dent versus just 24 who viewed him unfavorably. These are far below the minimum cell size to accurately forecast a result, especially in a state the size of Pennsylvania. Thus, the poll is trying to suggest that the opinions of 32 people are accurately depicting the feelings of an electorate of almost nine million voters.
Small-sample polling can be tricky because it bases conclusions upon very small pools of data. Though the Casey ballot test numbers certainly are in the realm of the possible, it is unlikely that all of the potential Republican candidates are viewed so negatively, especially when name ID is exceedingly low and their party just scored a big victory. More Pennsylvania polls will have to be taken before an accurate picture of Sen. Casey’s political health can be determined.