This post has been added to the… California In Transition- Must read Threads! 
Want on our daily or major news ping lists? Freepmail DoctorZin
Posted on 10/16/2003 2:34:38 PM PDT by Jean S
The press relied heavily on exit-poll results in analyzing the results of the California recall election. That exercise left some analysts, including me, a little confused because the two major exit polls didn’t always agree.
Even as exit polling tried to recover from its Voter News Service (VNS) 2002 blotto hangover, there are reasons to remain suspicious of this genre of polling.
First, it should be noted that since the polls are based on huge samples, the theoretical margin of sampling error for them should be quite small. The Edison Media Research poll, directed by longtime exit pollster Warren Mitofsky and used by Fox and other major news outlets, interviewed 4,214 voters. The alternative Los Angeles Times exit poll was based on interviews with an even larger sample of 5,205 voters.
In most respects, the two polls agreed on the composition of the California electorate last week. But there are notable and disturbing discrepancies.
The Times poll recorded twice as many Asian voters, 6 percent, as did the Edison poll, 3 percent.
The Times poll said that 59 percent of its voters have completed college but just 51 percent of the Edison poll’s voters have done so.
The Times poll reported that 46 percent of its voters reside in households with incomes of $75,000 or higher. Only 42 percent of Edison’s sample had incomes in this range.
The Times poll said that 25 percent of recall election voters are “conservative Republicans” while the Edison poll found many more, 32 percent.
Those differences are substantively and statistically significant, leaving observers wondering about the accuracy of either or both polls. While some nuanced explanation might be offered to explain away some of those differences, the sense that we cannot completely trust exit polls is palpable.
The polls also revealed philosophical and analytical biases in their design that are potentially more disturbing.
In general, the Times polling was more “conventional” in its approach to selecting demographic questions for analysis. For example, it asked voters whether their religion is Catholic, Jewish or non-Catholic Christian. The Edison poll apparently ignored religion altogether, a curious and puzzling omission given the questions raised about Arnold Schwarzenegger’s ability to hang on to conservative religious voters given his liberal social views and late-developing sexual misconduct charges.
The Edison poll ignored religion and marital status but asked about sexual orientation: Are you gay, lesbian or bisexual?
Only 4 percent of the Oct. 7 electorate reported any homosexual or bisexual tendencies to Edison poll interviewers, making that small cross-tab more symbolic than useful. Meanwhile, the 6 percent of California voters who are Jewish, the 25 percent who are Catholic and the 29 percent who are married women were overlooked by the most quoted exit poll. In place of asking traditional analytical characteristics of the electorate, the apparent successor to the failed VNS operation decided that knowing gay opinion was more revealing than knowing Catholic opinion.
If Edison/Mitofsky, the apparent successor to VNS, insists on promoting an agenda that reflects those social, political and research values, we may soon be yearning for a return of the people who ran the now-defunct VNS. The VNS crowd couldn’t handle programming and was deaf to complaints about such issues as the morality of releasing results before the polls are closed.
By the way, by combining the two polls, I was able to learn that the new governor got stronger support from non-Catholic Christians (58 percent) and Roman Catholics (46 percent) than he got from homosexuals (32 percent). Maybe that will start a new reality show: queer views for the guys in the pews. That sounds like something that Fox and their new favored exit pollster could pull off.
Dr. David Hill is director of Hill Research Consultants, a Texas-based firm that has polled for Re-publican candidates and causes since 1988.
Want on our daily or major news ping lists? Freepmail DoctorZin
Memo to Governor-elect Schwarzenegger: Dance with the one that brung ya.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.