DATA COLLECTION AND PROCESSING
Telephone interviewing for each MCAP poll is conducted by trained student interviewers under the direction of the center’s director. The MCAP’s interviewing facilities consist of sixteen sound-protected workstations equipped with state-of-the-art computers. When operating at capacity, the interviewing facility generates forty, fifteen-minute interviews per hour.
In order to increase the likelihood that hard-to-reach respondents are included in the sample, between 4 and 8 attempts (an initial call plus three or four callbacks) are made to each telephone number. Although interviews are generally conducted on weekdays between 6:00 and 9:00 PM, hard-to-reach respondents may be called during weekday mornings or afternoons, and on Sundays between 12:00 and 6:00 PM. When necessary, interviewers schedule appointments with respondents as a means of increasing cooperation rates.
The analysis of data is conducted by MCAP’s Director using state-of-the-art data statistical software. The software, SPSS, is specifically designed for the analysis of quantitative data. Prior to analyzing results, the data file is prepared by the center’s director. This process may include some or all of the following tasks:
Coding Responses to Open-Ended Questions. Questions for which the interviewer recorded verbatim responses using the CATI system’s voice capture software are often manipulated to produce quantitative data. To do this, three MCAP consultants and the center’s director sample thirty percent of the verbatim responses and create tentative codes for each response. Three MCAP staff members use the tentative coding categories to assign codes to each verbatim response. Once each staff members has completed coding the entire data file, his/her results are compared to the results obtained by the two other members of the staff. In instances where all three staff members agree on a particular code, the information is entered into the survey data file. In places where one or more coders disagree, the coders consult one another and, if necessary, create a new code category.
Data Cleaning. The CATI system utilized by MCAP minimizes the need for data cleaning as invalid entries by interviewers are not permitted by the software. In cases where anomalies exist (e.g. interviewer inputs an in accurate code), the interviewer makes note of their error and the center’s director modifies the response in the survey data file.
Weighting. In cases where minor variations exist between a quality of the study population and sample, the MCAP director may weight the results. Weighting occurs by means of software designed correct variations between a group’s representation in the study population and the sample.
Reporting. The results of MCAP polls are made available to the public via the Mercyhurst College Department of Political Science website.