I can't get my hands on the survey that Wallis is citing. The most recent study on this subject at the Pew web site (http://pewforum.org/docs/index.php?DocID=26) does have a few interesting tidbits, though.
"[S]ignificant numbers of Americans say they would be reluctant to vote for a presidential candidate even if generally well-qualified if the candidate was a member of a specific faith. Nearly four-in-ten (38%) say they would not vote for a well-qualified Muslim for president, and 15% express concern about voting for a well-qualified evangelical Christian. Far fewer say they would not vote for a Jewish (10%) or Catholic (8%) candidate. But fully half say they would not vote for a well-qualified atheist."
There go my presidential ambitions.
For what it's worth, a 2000 survey (http://people-press.org/reports/display.php3?ReportID=32) reports that 45% of respondents attend church at least once per week, and 61% attend at least once per month. These figures are self-reported, so they probably include a bit of wishful thinking on the part of respondents.
--tim francis-wright