Laserfiche WebLink
converting refusals by respondents whose eligibility has been established involves two <br />steps. After the first refusal, we call back at a different time. We ask for the eligible <br />respondent, beginning by apologizing for bothering them again, and subsequently <br />attempting to explain the scientific reason for our interest in speaking with them. This <br />procedure converts about 20% of refusals known to be eligible for the survey. If this <br />results in a second refusal, we will try one additional time using a "please help" approach. <br />This generally yields a conversion rate of 10%. After tI-iree refusals, the telephone <br />number is retired and classified accordingly. This procedure is a modification of <br />Dillman's (1978) Total Design Method. <br />Our response rates for listed sample survey projects such as this generally exceed 65% <br />and supports precise inferences from the sample to the population. <br />Duality Cont1•ol a~zd Verifccatiotz o. f'Data Bnt~~~ Procedures <br />Data entry and consistency-checking features of CATI software, real-time random <br />monitoring of interviews in progress by SSRC shift supervisors, and daily review of the <br />data collected ensure high quality information that accurately depicts respondents' <br />answers to survey questions. <br />Survey questions and response options appear on a computer screen while the interviewer <br />is speaking to the respondent. Data are entered directly into the CATI system, which will <br />not accept out-of--range response options, so recording errors are reduced. In addition, <br />range and validity checks can be programmed to minimize other recording and survey <br />administration errors (e.g. verifying the number of jobs reported, whether income from <br />day labor was received, and the wages associated with each). <br />In addition to these technical features, SSRC supervisors are present during all <br />interviewing shifts. Each of our supervisors previously worked as an intervie«~er, and <br />each has been trained in telephone inten~iewing techiuques and methodological <br />considerations. Random call-monitoring is performed during each shift to verify the <br />accuracy of the data. Clients can connect to this call-monitoring system from anoff- <br />campus telephone number, allowing them to listen as data are collected. <br />Our 24-station CATI lab enables us to collect data quickly and our export function allows <br />for the rapid extraction of results into numerous fornlats; including ASCII, SPSS, and <br />SAS formats. The resulting quantitative data and open-ended text responses are reviewed <br />daily so that interviewing staff can~be "debriefed" regarding vague or imprecise data. <br />This is done throughout the life of the project. <br />14 <br />25N-40 <br />