 |
Thursday, Aug 28 |
|
| ||||
PSID continued its strong measurement of economic, demographic, and sociological core content areas. As in prior waves, the changing demographics of the sample families were accounted for by the updating of marital and fertility information, the movement of individuals into and out of the family, and the formation of new sample families (i.e., "split-offs").
The minor changes made to the 2003 question included the following: A special series of questions on cohabitation was added to account for between-wave cohabiters who may have lived with the family for a period of at least four months. Additional questions were added about changes in residence since last wave. An additional series on computer and Internet usage in the home was introduced. Labor force participation and detailed employment history going back to January 2001 were collected in a newly redesigned series utilizing an event history calendar methodology. As in past waves, questions were asked about food costs, Food Stamp program participation, and food security. Detailed income and wealth measures at the individual and family level, the mainstay of the PSID, were collected along with details on the primary adults’ pension and retirement plans. New questions on hospital stays and IADLs of primary adults were added to the already extensive series on general health and chronic conditions assessment, health behaviors of primary adults, and health care costs for the entire family. The series on 30-day emotional distress, started in 2001, was continued. In addition, two questions were added about the experience of depression over the preceding 12 months. Both the 30-day distress and 12-month depression questions were asked only of the respondent. The series on nutritional knowledge was dropped for 2003. The series on philanthropy and volunteering, begun in 2001, was repeated with a much-expanded series on volunteering behavior of the primary adults. The series on welfare program participation and gross measures of income and labor for adults over the past two years was repeated.
The 2003 wave saw the biggest change in instrument format since the interview went from paper-and-pencil to computer-assisted interviewing (CAI) in 1993. Two CAI softwares were used for the instrument. Most of the interview retained a standard structured question list format. This portion of the interview was programmed in three instruments in Blaise, the new standard CAI software for SRC surveys, and presented to the interviewer and respondent in much the same way as in prior years. However, the residency changes and employment history series utilized a new ‘event history calendar’ (EHC) methodology that allowed the interviewer to record spells of time directly onto a calendar display and interview in a less structured manner. The EHC methodology was found to enhance respondent recall making for potentially more accurate data and a more enjoyable interview experience for interviewer and respondent.
Along with changes in CAI software, the study also saw a complete transition in case sample management and related systems, reprogrammed in new software. The systems and programs that generate, distribute, and monitor sample and interview data to and from interviewers, and contact and pay our respondents all were updated.
No changes in sample eligibility were made in 2003, and no supplemental sample was added. As in 2001, the sample was divided into two parts. The Core Sample consists of original 1968 sample families and all related families generated from them. Within this part were 6,915 families who were successfully interviewed in 2001 (Reinterviews) and 340 families who were part of a responding family in 1999 or 2001, but were nonresponse in 2001 on their own interview (Recontacts). These 7,255 families were the initial Core Sample eligible for interviewing in 2003. During interviewing, new sample families are generated as members move out of response families and establish their own families. In 2003, 792 such new families, called Splitoffs, were generated from response Reinterview families. An additional 33 families, called Recontacts or SplitRecons, were generated from responding Recontact families.
The second part of the sample is usually called the “New Sample”, but more correctly called the Immigrant Refresher Sample, or just Immigrant Sample. These are a supplemental sample added in 1997 and 1999 to make the PSID nationally representative given the changes in population due to post-1968 immigration, plus new families generated from those responding in 1997-2001. Four hundred and ninety-one such families were initially loaded eligible for a Reinterview in 2003. An additional 83 families were loaded as eligible Recontacts. Responding Reinterview families generated 60 new potential Splitoff families; responding Recontacts generated 10 new ReContacts or SplitRecons.
Interviewer training was conducted in Ann Arbor in two sessions, March 6-10 and March 12-16 by a team of SRO and PSID staff. Training was comprehensive and covered all aspects of field procedures, study-specific content and methodology, and CAI equipment operations. All interviewers were given extensive practice, help sessions, and self-study prior to certification. Interviews were conducted by over 90 interviewers from the SRC National Field Team using CAI primarily via telephone (96.8%) from their homes. The data collection field period ran from March 17, 2003 to November 9, 2003.
A total of 7,870 interviews were completed for an overall response rate (RR%) of 93.5%. Within the Core Sample, 6,554 Reinterviews were completed for a wave-to-wave RR of 96.8%; 561 interviews (RR 83.1%) were completed with the Splitoff families generated from these Reinterviews. Also within the Core Sample, 215 Recontact interviews (of 333 eligible) were completed for RR of 64.6%. Within the Immigrant Refresher Sample, 459 Reinterviews were completed for a Reinterview RR of 94.3%; 36 Splitoffs were completed for a RR of 61%. Recontact interviews in this subgroup were completed with 45 families (of 93 eligible) for a RR of 50.6%.
Average interview length over all respondents was 68.7 minutes. Additional time updating changes in the family composition and collecting respondent contact and payment information made the total average R burden 78.7 minutes. The respondent incentive was $55.00.
Institute for Social Research | University of Michigan | Privacy | Conditions of Use
