Abstract copyright UK Data Service and data collection copyright owner.
The aim of this project was to examine the role the interviewer in professional survey organisations plays in the survey nonresponse process. The 18-month-long programme was supported by a grant from the Economic and Social Research Council (ESRC) and was carried out in co-operation with the ESRC Centre on Micro-Social Change and NOP Research. It was divided into three sub-projects. The aim of the first sub-project was to tease out the interviewer effect on nonresponse from other effects, such as the effect of the area, and individual respondents, by using a specially designed experiment from an existing source - the <i>British Household Panel Study (BHPS)</i>. The second sub-project investigated when people are likely to be home and how interviewers might develop efficient calling strategies using two large pre-existing datasets - the <i>Family Resources Survey (FRS)</i> and <i>British Social Attitudes (BSA)</i>. The third sub-project focused on the initial doorstep interaction between interviewer and address residents, which can be crucial in determining whether or not the prospective respondent will agree to co-operate. The purpose was to gain a better understanding of what takes place during this interaction by collecting a small sample of such data using 32 interviewers from two different organisations - SCPR and NOP. Data collection for this sub-project covered addresses from NOP's <i>Political Tracking</i> survey, and SCPR's random half of the <i>FRS</i>. A total of 11 datasets are included in this study, numbered a - k: a. Interviewers and Survey Nonresponse : Household Level Doorstep Interactions, Political Tracking, 1996. b. Interviewers and Survey Nonresponse : Household Level Doorstep Interactions, FRS (SCPR portion), March 1996. c. Interviewers and Survey Nonresponse : Matched Contact Form/Tape Data at Call Level, Political Tracking, 1996. d. Interviewers and Survey Nonresponse : Matched Contact Form/Tape Data at Call Level, FRS (SCPR portion), March 1996. e. Interviewers and Survey Nonresponse : Interviewer Questionnaire on Survey Participation, 1996. f. Interviewers and Survey Nonresponse : Individual Transcripts, Political Tracking, 1996. g. Interviewers and Survey Nonresponse : Individual Transcripts, FRS (SCPR portion), March 1996. h. Interviewers and Survey Nonresponse : BHPS Household Level Wave 1 Call Record Subset, 1991. i. Interviewers and Survey Nonresponse : BHPS Household Level Wave 2 Call Record Subset, 1992. j. Interviewers and Survey Nonresponse : FRS (SCPR portion) Address Level Call Records, 1995-1996. k. Interviewers and Survey Nonresponse : BSA Address Level Call Records, 1994.
Main Topics:
Datasets a - b (see list under Abstract) : what interviewers and respondents said on the doorstep as collected by tape recordings and by a contact description form completed by the interviewer; also situational and environmental aspects of the interaction, physical details of the dwelling unit and neighbourhood as observed by the interviewer, and main survey final outcome variable. These are household level files - the tape data and contact description form data are matched at the household level, but not by individual call. Call level matching is only available in datasets c and d. Datasets c - d : what interviewers and respondents said on the doorstep as collected by tape recordings matched at the call level to how the interviewer described the same interaction on a contact description form (non-contacts and non-matches have been excluded). Dataset e : Interviewers' views on what they typically say and do as part of their doorstep behaviour, their views on persuasion strategies and non-contact strategies, and general background variables on the interviewer. Datasets f - g : actual transcripts of what interviewers and respondents say on the doorstep as collected by tape recordings. Datasets h - i : Wave 1 and Wave 2 call record data; call number, day, date, time, outcome; NOP interviewer characteristics; variables for matching to the <i>BHPS</i>. Datasets j - k : call record data; SCPR interviewer characteristics.
No sampling (total universe)
Multi-stage stratified random sample
Datasets j - k (see list under Abstract) - all call records from the original sample were used. 'Meta' observations and recordings in data files were drawn from multi-stage stratified random sampling. Datasets a - d, f - g used other non-random sampling - 16 interviewers and their respective assignment areas were selected to allow for geographic spread (but excluding Scotland) and to allow for a range of experience levels. Dataset e: 16 interviewers each from two organisations were selected to allow for geographic spread (but excluding Scotland) and to allow for a range of experience levels. Datasets h - i: call records were keyed for the households involved in the interpenetrated sample design sub-sample in Wave 2 of the <i>BHPS</i>. This sub-sample was originally generated by clustering adjacent PSUs into 'geographic pools' and selecting a systematic sample of these pools.
Face-to-face interview
Self-completion
'Meta' observations and recordings in data files were based on face-to-face interviews. Dataset e (see list under Abstract): a self-completion form was completed by the interviewer. Datasets a - d: tape recording and contact description forms completed by interviewer. Datasets f - g: tape recording and later transcription. Datasets h - k: coversheets/address record forms completed by interviewer.
