Military News, News From Iraq & Afghanistan - Military Times

Sampling the military

The Military Times 2008 Election Poll is the latest in a series of efforts to gauge the attitudes and opinions of a crucial but hard-to-measure group: members of the U.S. military.

To gather military opinion on this year’s presidential election, Military Times began with a list of more than 80,000 subscribers and former subscribers to Army Times, Navy Times, Marine Corps Times and Air Force Times who gave the papers their e-mail addresses.

Each was e-mailed an invitation to participate through a secure Internet page. Security measures ensured that readers could not submit multiple entries and that only those who received an invitation could respond.

From Sept. 22 to Sept. 29, 2,813 active-duty members, 1,480 members of the National Guard or reserve and 4,411 retirees participated. The figures include only those respondents who indicated they are registered and intend to vote in the election. Chart totals may not equal 100 percent because of the rounding of figures.

Although public opinion pollsters use random selection to survey the general public, the Military Times survey is based on responses from those who chose to participate. That means it is impossible to calculate statistical margins of error commonly reported in opinion surveys, because those calculations depend on random sampling techniques.

The voluntary nature of the survey could affect the results — if supporters of one candidate are more prone to express their opinions, for example. The dependence on e-mail could also affect the results, because e-mail users may have different characteristics than the military population as a whole.

Characteristics of Military Times readers may also affect the results. The group surveyed is significantly older than the military as a whole, and the survey group contains a higher percentage of officers than is present in the military.

Conversely, junior enlisted troops, women and racial and ethnic minorities made up a smaller share of the sample than of the military at large. While it is difficult to predict how those factors affect the results, those groups are generally regarded as more supportive of Democratic candidates.

More details, including some of the data from the survey, are available online at http://www.militarytimes.com.

Researchers interested in obtaining data can contact database reporter Brendan McGarry at bmcgarry@atpco.com.