Abstract copyright UK Data Service and data collection copyright owner.
The Cardiff survey was undertaken for geographical/planning research purposes, in particular to create the geographical data base required by two associated research projects at the University of Bristol and UWIST. These projects are concerned with analysing the characteristics of urban shopping behaviour in the 1980s and developing predictive models which can be used for retail planning purposes.
Main Topics:
The Cardiff Consumer Panel Survey, funded by the SSRC, monitored the daily food and grocery shopping behaviour of approximately 500 households from January to July 1982 (24 weeks main monitoring period, plus 2 weeks 'running in' period). It is believed to be the largest, long-term, consumer panel survey of a single urban area which provides information on the exact location (i.e. which shop and shopping centre) of each purchase made by the panel members. In addition, it provides information on the products (and sometimes brands) purchased on each shop visit, the time of the visit, the amount spent at the shop, the travel mode, the shopping trip characteristics, the demographic, socio-economic and attitudinal characteristics of the panel member and the household to which the panel member belongs. The survey was successfully piloted with SSRC funding in 1980. The main survey had an initial panel size of 605 households. This was reduced by panel attrition over the 26-week monitoring period to 481. It was then further reduced, for purposes of analysis, to a sub-set of 454 'continuous reporters' for whom complete purchasing, demographic, socio-economic and attitudinal information is available. (Demographic and socio-economic information is not available for 3 of these, and for some purposes the sub-set must be reduced to 451). The data tape consists of: (1) 24 files containing the purchasing records of 454 'continuous reporters' over the main 24-week monitoring period (weeks 03-26); (2) 24 files containing the purchasing records of the 542 panellists who began Week 03 (i.e. 454 'continuous reporters', 27 who completed the main monitoring period but were subsequently excluded, 61 panellists who dropped out of the panel during Weeks 03-26) (3) 1 file containing second (or final) questionnaire data; (4) 1 file containing first (or initial) questionnare data.
Multi-stage stratified random sample
(a) 1st stage. The characteristics of a random sample of 50 enumeration districts from the city of Cardiff 1971 Census classification had been established in a previous survey. These districts were then stratified into 8 strata on the basis of 3 accessibility and potential mobility variables. Within each stratum one enumeration district (in north or east Cardiff) was selected at random. To each of these 8 enumeration districts, 2 contiguous enumeration districts belonging to the same stratum were added, creating 8 small clusters across the city. (b) 2nd stage. Within each of the 8 sub-areas, 255 housholds were selected at random by systematic sampling from specially created lists of electors (based on te 1981-82 electoral roll and adjusted to the 1971 area boundaries used in Stage 1). The sampling procedure was adjusted (using the method of `firsting') to take account of the fact that the electoral roll is not a roster of households but a list of electors who come from households of varying size. These households were then approached by the fieldworkers and invited to become members of the consumer panel. In 4 sub-areas this procedure successfully generated the target size for the initial sub-area panel, but in the other 4 areas it was necessary to generate additional lists of households (ranging in size from 22 to 91 and generated by the same method) to obtain (or approximate) the target initial panel size for those areas. This sampling procedure provided an initial panel of 605 households which was reduced by a 20% panel attrition rate to 481 by the end of the 26 week monitoring period. For purpose of analysis this was further reduced to a sub-set of 454 (or 451) 'continuous reporters'.
Face-to-face interview
Purchasing records (diaries)
