SURVEY RESEARCH AT ISR

TELEPHONE SURVEYS

For general population surveys, ISR selects telephone survey respondents by random digit dialing (RDD), followed by a random selection of respondents within households. While somewhat more expensive than sampling from telephone directories, this method ensures the inclusion of individuals with unlisted telephone numbers and those who have recently moved. Such individuals are known to have distinctive attitudes and demographic characteristics so excluding them would introduce significant bias in the survey results. To improve response rates and sample representativeness, at least 12 attempts at different times of the day and different days of the week are made to reach selected telephone numbers. In terms of data quality and response rates, ISR telephone surveys compare favourably with those using more traditional personal interviews. In addition, telephone surveys can often be completed more rapidly and at considerably lower costs than personal interviews.

COMPUTER ASSISTED TELEPHONE INTERVIEWING (CATI)

Since 1987 the majority of our projects have been conducted from our centralized Computer Assisted Telephone Interviewing (CATI) laboratory. Interviews are conducted at thirty stations equipped with personal computers and linked through a local area network (LAN). With this system, questions are read from video display terminals and respondents' answers are entered directly into a computer. To ensure the quality of the data, supervisors monitor interviews in progress.

CATI offers significant advantages over the traditional paper and pencil methods of conducting telephone interviews. Because there is no separate data entry step, data processing errors are minimized, along with the time and expense of data entry. Interviewer errors are dramatically reduced because the computer system controls the sequencing of questions and 'skips' around questions that are not applicable to a given respondent. Incorrect response codes are rejected by the system and the interviewer is prompted for a correction.

CATI dramatically increases the range of questions that can be included in an interview. In order to determine the extent to which alterations in question wording affect responses, it is possible to introduce random variation in the wording of questions. Similarly, the sequencing of questions can be altered to test for 'order effects'. The computer system also allows the content and sequence of questions to be altered on the basis of a respondent's answers to previous questions or on the basis of an experimental design requiring random variation of survey items among respondents.

MAIL SURVEYS

Mail surveys offer an attractive alternative for gathering data from well-defined populations. The Institute has obtained high response rates in mail surveys of physicians, lawyers, residents of housing projects, students, police officers and members of organizations. Mail surveys can provide good quality data and good response rates at a very low cost. For studying sensitive topics, mail surveys often have been found to be less subject to 'desirability bias' -- i.e., the tendency of some respondents to give favourable answers about themselves. Mail surveys are also an excellent way to collect substantial amounts of additional information from respondents who have completed telephone or personal interviews.

Data collection strategies, especially the number and timing of follow-up contacts after the initial questionnaire mailing, are critical to achieving high response rates. In a number of surveys, the Institute has increased final response rates, at limited cost, by using the telephone in a final effort to encourage respondents to return the questionnaire.

FOCUS GROUPS

The Institute uses focus group methodology in a variety of ways in applied social research projects. In addition to forming the base for qualitative studies of many types, focus groups are used to explore new issues or topics which are poorly understood and are a useful starting point for questionnaire design. As a tool to identify needs, to discover what people think when they make decisions, and to help understand why people think the way they do, focus groups are also useful in program and policy evaluation in all types of organizations.