Interviewing for The Roanoke College Poll is conducted by The Institute for Policy and Opinion Research (IPOR) at our call center on the campus of Roanoke College. Typically, about 600 people are interviewed for each poll. This yields a margin of error of +4 percent at the 95 percent level of confidence. Sample sizes for some election polls are 800 or more, resulting in a margin of error of +3.5 percent or lower. The margin of error is not adjusted for survey design effects unless so stated.
Interviews are conducted through random-digit-dialing with both land lines and cell phones to better represent the Commonwealth. RDD samples are created by ASDE Survey Sampler. Quotas are used to ensure that different regions of the Commonwealth are proportionately represented, and the data are usually statistically weighted for key demographic characteristics such as gender, race, and age. As noted, some surveys are also weighted for political party. Consumer sentiment polls are weighted to reflect demographics from the U.S. Census in Virginia. Election polls are weighted to reflect Exit Poll results for the most recent appropriate election-Presidential, Gubernatorial/statewide, or midterm. Polls taken several months prior to an election typically include all Virginia residents who are eligible to vote, while those taken closer to an election may screen for registered voters or for likely voters. For land lines, we ask to speak to the youngest male over the age of 18; only age screens (over 18) are applied to cell phone calls. Each IPOR poll press release has its own specific Methodology section.
IPOR subscribes to the American Association for Public Opinion Research (AAPOR) Code of Professional Ethics and Practices. IPOR is non-partisan and non-ideological. Unless otherwise noted on a specific report or press release, IPOR surveys are funded as a community service by Roanoke College.