Public opinion polls conducted by the Mercyhurst Center for Applied Politics (MCAP) employ either registration list-based sampling or random digit dialing.
A registration list-based sampling methodology, or RBS, is utilized for most MCAP polls related to local, state and federal elections. In these cases, a list of registered voters who reside in the relevant geographic area (e.g. city, county, congressional district, state) is purchased from a vendor, such as Voter Contact Services (http://www.vcsnet.com/). The list provides a telephone number, address, county of residence, age, and other demographic information on nearly all registered voters in a particular geographic area. The telephone numbers associated with each registered voter are acquired from voter registration records and by cross referencing names and addresses against telephone directories. Although most telephone numbers are associated with landlines, an increasingly large percentage of numbers are for cell phones. Registered voters who list their cell phone numbers as their home phone number are not excluded from the sample frame.
Since the RBS sample frame is drawn from a list of individuals, no respondent selection procedure is necessary. Likewise, RBS sample frames allow stratification along a variety of dimensions which facilitates the analysis of poll results and research by academics. The RBS sample frame has the added benefit of allowing MCAP to contact respondents by mail at a later date for the purpose of follow-up interviews or inviting them to participate in focus groups.
A random digit dialing sampling methodology, or RDD, is employed most frequently for MCAP public opinion polls that are not directly related to elections. Telephone numbers for these surveys are generated using RDD software, and respondents are randomly selected from within each household. Cell phone numbers are excluded from the sample by imputing parameters into the random number generator, and interviewer screening procedures ensure that only residential telephone numbers are included in the sample. Since the RDD methodology allows each household telephone number an equal chance at being selected for inclusion in the sample, RDD is ideal for gauging public opinion as it pertains to issues confronting localities, states and the nation.