Telephone Interviewing

The UW Survey Center (UWSC) and its staff have a great deal of experience with Computer-Assisted Telephone Interviewing. We have been using CATI procedures in all of our telephone surveys since we began operations in 1987. Since the early 1990s, UWSC has been using the CASES (Computer Assisted Survey Execution System) software, which was developed in the early 1980s by the Computer Assisted Survey Methods Program at the University of California Berkeley. CASES has evolved into a comprehensive and powerful CATI system that is rapidly becoming the survey research industry standard.

The UWSC has several staff members who are highly skilled CASES programmers. Any of these individuals has the skill to program the instrument and manage its development and implementation. Several of these staff members are also highly trained and experienced in the coding and cleaning phase of a CATI project.

CASES is a comprehensive computer-assisted interviewing system that can be used in all stages of data collection and manipulation. CASES instruments can be used to collect, check, clean and code data, and write special-purpose reports. It also has excellent survey-management capability. It is relatively easy to convert an interview into CASES code. The logic is quite simple to apply, and we have found that it permits us to do anything that we have ever wanted to do in an interview.

During an interview, interviewers can back up to any previous point in the interview to review and, if necessary, correct answers. When an answer is changed the computer directs the interviewer to the next appropriate unanswered question. Open-ended responses of any length can be recorded. Interviewers may leave notes at any time. Responses must be entered for all appropriate questions. An interviewer may not jump forward and skip any appropriate unanswered question. An interview can be interrupted and resumed at a later date.

Execution of the data collection instrument results in the creation of two data files. The first is a structured data record for each case. The second file consists of a history of events associated with the case, including responses to open-ended questions and spontaneous interviewer notes with a record of the question numbers where the notes were produced.

A modified version of the data collection instrument is used to check, clean and code the data after it has been collected. The data checking, cleaning, and coding instrument can be highly customized, offering the user the ability to perform a variety of post-entry processes. Open-ended questions can be coded categorically, new categories can be added to existing items, new items can be created and selected entry instrument codes can be flagged for coder review.

Accompanying CASES is SDA, a set of programs designed to simplify common data analysis and documentation tasks. SDA provides extremely rapid execution of commonly used statistical operations and full documentation for survey data sets. SDA is able to read variable and value labels and full question text from the CASES survey instrument and use this information to create SPSS, SAS or STATA syntax files. SDA also excels at creating codebooks and provides tools for making data sets web-accessible.

The University of Wisconsin Survey Center also retains a highly trained and professional interviewing staff. Before hiring an interviewer the UWSC conducts a personal interview, thoroughly checks the candidate's references, performs a mock interview, and administers a typing competency test. Only those who perform well in all areas are considered for the interviewing position. Besides a multi-day training period immediately after hiring, all interviewers attend study-specific briefings before a new project begins to ensure the interviewer is well prepared to deal with issues and answer questions from respondents pertaining to the study when they call. Interviewers are also monitored monthly by a shift leader to ensure high quality data collection.

CASES also provides comprehensive survey management procedures including sample management, call scheduling, and procedures to monitor the progress of a survey. As a fully featured CATI package, CASES has all the programs necessary to install sample (including importing pre existing data into the sample records); prepare an "entry instrument" (their term for a CATI interview schedule); monitor survey progress (by case, or by interviewer, or by project, as needed); automatically send into the field those cases which require calling at a specific time or date (also known as automatic call scheduling); code and clean data; produce reports; and output data into rectangular files for analysis. The system's program tools for monitoring survey progress are particularly strong. These management programs are fast, accurate and easy to use.

We have developed our own program to schedule interviewers. The program tracks the minimum shift and peak calling hour requirements that we ask of employees. This insures that a thorough spread of calling hours for our projects, with an emphasis on evening and weekend hours for maximum yield.

Instrument Design and Testing

Many of our project directors have special training in instrument design and are familiar with the extensive literature on this complex topic. A few clients come to the UWSC with an instrument that is developed, tested, and ready to be fielded. Other clients find a period of instrument development - which could include individual interviews or focus groups - to be beneficial to the success of their project. In addition to such development efforts, UWSC staff have experience in multiple methods of testing and evaluating survey instruments, including cognitive interviewing, designing questions to debrief respondents as part of a pretest, and debriefing pretest interviewers. Because our telephone interviews can easily be recorded digitally, clients can usually listen to pretest interviews at their convenience.