Remember the recent web-based survey that the American Family Association intended to present to Congress as proof of Americans' overwhelming opposition to gay marriages and civil unions? Well, it turns out the poll didn't quite turn out the numbers the AFA planned:

The conservative American Family Association (AFA) said it will not take the results of its marriage poll to Capitol Hill after a majority of respondents favored same-sex marriage, according to a Thursday report in Wired News.

The AFA posted the poll online in December with a stated intention to forward the results to Congress as evidence of U.S. opposition to same-sex marriage. Respondents could select one of these three choices:

* " I favor legalization of homosexual marriage."
* "I favor a 'civil union' with the full benefits of marriage except for the name."
* "I oppose legalization of homosexual marriage and 'civil unions.'"

But as of press time, the numbers support same-sex unions: Sixty percent favored same-sex marriage and 8 percent favored civil unions, leaving just 32 percent opposed.

AFA representative Buddy Smith complained to Wired News: "It just so happens that homosexual activist groups around the country got a hold of the poll -- it was forwarded to them -- and they decided to have a little fun, and turn their organizations around the country (on to) the poll to try to cause it to represent something other than what we wanted it to. And so far, they succeeded with that."

Matt Foreman, the Executive Director of the National Gay and Lesbian Taskforce was dumbfounded. "The abject hypocrisy of these people never ceases to amaze me," he exclaimed.

Previously, the AFA had claimed the poll would back up their claim that "99 percent of Americans oppose gay marriage."

Interesting. You'd think that "99%" wouldn't have had much of a hard time putting their votes in, no? Maybe they're just not "web-savvy"? I took the poll myself, and can vouch for the fact that the site was configured to reject repeat votes from the same computer or IP address, which would tend to discount the theory that a small number of users "loaded" the survey.

---