CA STATE UNIVERSITY, FULLERTON (SOCIAL SCIENCE RESEARCH CENTER) 14 - 2008

Home Browse Search

TECHNICAL APPROACH TO DATA COLLECTION The SSRC's computer assisted telephone interviewing (CATI) software and standard operating procedures ensure a superior response to issues of response bias, callback, quality control; and the verification of data entry procedures essential for all telephone survey research endeavors. These are discussed in detail below. Response Bias Response bias is a major source of sampling error that can be minimized by preserving the random nature of the selected sample. By attempting sample records that result in a busy, no answer, or answering machine disposition up to 20 times, a representative sample of WORK Center "Core A" or universal access customers that used the Center between July 1, 2003 and June 30, 2004 will be surveyed. Up to 20 attempts at contact are also made for any sample record known to connect to a residential household, such as requests to phone the individual another time/day. These "callbacks" are explained in the next section. Callbacks The CATI system includes a sophisticated call tracking and callback scheduling procedure. A call history is developed for each number and remains active until a final disposition is assigned. If no contact is made, the call record will note the time of day and the interviewer who attempted the call. The call will then be automatically reassigned at a later time based upon an algorithm that reduces the probability that the call will come up again on the same day and time. When a contact is made, but the interview is not completed, call information is recorded that includes whether a callback has been scheduled, who the interviewer spoke with, and the current disposition of the call (for example, immediate refusal, answering machine, mid-interview termination, etc.). In addition, the time of each call, the number of call attempts, and any interviewer-generated notes are recorded. For multi-lingual studies such as this, records with phone numbers lalown to connect to a Spai>ish-speaking household (generally established by a prior call attempt) are dispositioned accordingly and automatically routed to an interviewer with that language capability. If no linguistically competent interviewer is available, the system holds the record until such an interviewer is free. The CATI system allows the researcher to set the number of times a telephone number is to be called before it is retired. SSRC standard operating procedure dictates 20 attempts at contact. If contact is not established after 20 calls, the number is transferred to a holding queue. Our policy regarding immediate refusals is to call back at random intervals after the initial refusal. Experience shows that about 30% of initial refusals can be converted simply by contacting the individual another day. Our procedure for 13