Thursday, November 29, 2007
Experts Question HRC's ENDA Survey
TransGriot Note: Seems like the TransGriot isn't the only person who thinks that the HRC survey that came out in the middle of the ENDA debate is shady and bogus.
Researcher says methodology 'doesn't make sense'
By JOSHUA LYNSEN | Nov 28, 4:47 PM
from the Washington Blade
Polling experts are questioning a recent Human Rights Campaign survey that asked gays about the Employment Non-Discrimination Act.
The survey's results, circulated last month by HRC when many gays were locked in heated debate over the measure's lack of transgender protections, show most people who responded support the bill as written.
But John Stahura, who specializes in survey research and directs the Purdue University Social Research Institute, said the survey's methodology is problematic.
"They're playing games," he said after reviewing survey excerpts at the Blade's request. "It doesn't make sense."
Conducted for HRC by Knowledge Networks, the survey shows most respondents believe national gay groups should support ENDA despite its lack of protections for transgender workers "because it helps gay, lesbian and bisexual workers and is a step toward transgender employment rights."
According to survey excerpts, about 68 percent of respondents chose that scripted statement among three offered lines to best represent their "point of view."
Another 16 percent of respondents indicated national gay groups should oppose ENDA "because it excludes transgender people," and 13 percent wanted groups to take a neutral stance "because while it helps gay, lesbian and bisexual workers, it also excludes transgender people."
About 3 percent of respondents refused to answer the question. The survey offered no margin of error.
Stahura said he "never would" structure a survey to include such explanatory clauses "because what you're asking people to evaluate is the because."
"I don't know why they didn't go with a straightforward, 'Here's the act. Should we support it, should we oppose it, or should we take a neutral stance?'" he said.
Brad Luna, the HRC communications director, said each scripted statement included explanatory clauses to focus respondents on the measure's omission of transgender protections.
"With complicated proposals such as this, if you don't link opposition to a reason, you might get people opposing for a variety of reasons," he said. "We chose this method because we wanted to know specifically if people supported or opposed ENDA because of the transgender exclusion."
The survey, conducted Oct. 2-5, polled a roster of people who are gay, lesbian or bisexual, and were previously located by Knowledge Networks through random-digit- dialing methodology.
Among the survey's 514 respondents, 246 are male, 262 are female, five are female-to-male transgender and one is male-to-female transgender. Survey excerpts provided to the Blade by HRC did not disclose the sexual orientation breakdown among the respondents.
Luna said HRC is confident in the work that Knowledge Networks performs.
"While all surveys have limitations, Knowledge Networks surveys are very high in quality," he said. "They have a stellar reputation, and I have full confidence in their work."
But experts said the survey could have been done better, and the excerpts released earlier this month by HRC left some questions unanswered.
"I don't know based upon this response that you could say how the community — the gay, lesbian, bisexual community — feels about the legislation, " Stahura said. "I don't think those questions give you that answer."
Christopher Barron, a Washington political consultant Log Cabin's former political director, who is gay and does survey interpretation, agreed. He said the methodology, which he described as "bizarre," might not allow the results to be projected nationally.
"It may be that it's completely and totally sound," he said. "But there's nothing there that tells us that it is, so you can't assume it's a nationally representative sample."
Luna told the Blade this week that the survey is nationally representative.
Barron and Stahura, who reviewed a two-page memorandum and three data sets prepared by Knowledge Networks and provided to the Blade, also noted they could not determine whether the survey is scientific. Both experts said that lingering question would preclude them from using the survey's findings in their work.
"I would not approve it for publication, " Stahura said. "I think with the 'becauses,' you're really pushing people toward particular responses in this instance."
Luna disagreed. He said there was "never any intent to influence survey respondents. "
"We wanted to gain an understanding as best we could of where people were on the issue," he said. "A number of voices were claiming to speak for the LGBT population, but no one in fact had done the research to know."
Joshua Lynsen can be reached at firstname.lastname@example.org