Surveying the military - Air Force News | News from Afghanistan & Iraq - Air Force Times

Surveying the military

From Nov. 11 through 30, Military Times conducted a voluntary confidential survey on “don’t ask, don’t tell.”

Military Times e-mailed invitations to about 45,000 subscribers whose accounts listed a valid e-mail address and randomly inserted about 40,000 invitations into newsstand copies of Army Times, Air Force Times, Navy Times and Marine Corps Times. Each included a unique password and directed respondents to a third-party Web site to complete the questionnaire.

In response, 8,001 subscribers who received e-mailed invitations took the survey, as did 213 readers from newsstand inserts. Participation rates adjusted for undelivered invitations were 20 percent for subscribers and 1 percent for newsstand readers.

Except where noted in certain graphics, survey data were filtered to include only the 3,030 responses from active-duty members, including currently mobilized members of the National Guard and reserve.

Respondents reflect the career-oriented core of the Military Times readership; that is, they are, on average, older, more experienced, and more senior in rank than the overall military population. It is this population of officers and noncommissioned officers who are the day-to-day leaders who make the military run, and who ultimately are responsible for implementing policies formulated in the Pentagon and laws crafted on Capitol Hill.

Military Times consulted with several outside experts to construct the questionnaire. Before the survey results were tallied, some gay rights advocates, including the Palm Center, a think tank at the University of California, Santa Barbara, and the Servicemembers Legal Defense Network, criticized the wording and order of the survey questions. They argued that these factors might have biased responses.

But Dr. Laura Miller, a military sociologist at the Rand Corp. who has studied troop attitudes on this issue since 1992, and who was consulted on the wording and order of the questions, said: “It’s unlikely that military personnel would change their opinion on ‘don’t ask, don’t tell’ mid-survey because of question order. Most already have well-formed opinions on this topic, even if the opinion is that they don’t care either way.”

Public opinion pollsters use random selection to survey the general public, whereas the Military Times survey is based on responses from readers who chose to take part. Statistical margins of error commonly reported in opinion surveys cannot be calculated for this survey because those calculations depend on random sampling.

However, prior to launching the full survey, Military Times conducted three pilot polls — two via telephone and one via e-mail — to measure selection bias. An analysis of the results by the American University Statistical Consulting Center showed no statistically significant differences between active-duty members who responded at random and those who responded voluntarily. The survey data, including those from the pilot polls, are posted online.

Reporters and researchers with questions about the methodology and analysis can contact Database Reporter Brendan McGarry at bmcgarry@atpco.com or 703-750-8665.

Related reading

* Troops speak out on gays serving openly

* Stressed by a double life

* Research more likely than repeal this year

* ‘Don’t ask’ debate heats up online