EXPERIMENTAL DESIGN
In the first stage of the test, an initial sample was split into three groups: Two experimental groups and one control group (see Table 1). All respondents were first contacted by telephone. Next, the respondents in the control group and in experimental group 1 were interviewed immediately by telephone, whilst the respondents in experimental group2 were visited by an interviewer. In both interviews, a small selection of questions was asked twice within the same interview. At the end of the interviews, each respondent was asked to participate in a second interview that would take place one month later. In the second stage, one month after the first interview, the respondents were re-interviewed.
The two experimental groups were crossed over: Those interviewed using CATI the first time were interviewed using CAPI the second time, and vice-versa (table 1). It was explained to these respondents that we ask them the same questions in two different ways because we want to compare two data collection strategies. The control group, which had been interviewed using CATI the first time, were interviewed a second time using CATI again. This control group was included to control for change in opinions over the time interval of one month. In the experimental groups such a change in opinion would be confounded with the change in interview method. As in the first interview, a small selection of questions was asked twice within the second interview to all groups of respondents.
Method instrument: Standardised interviews face-à-face, Standardised interviews by telephone
Population, unit of observation: The sample unit was the household. All members of the household who were 15 years or older were to be interviewed.
Size of sample/selections: Initial sample of 1452 telephone numbers plus a second complementary sample of 727 telephone numbers. In total, 864 interviews were carried out.
Participant selection or sampling method: A simple random sample of telephone numbers within the agglomeration of Bern was drawn from a file containing all valid Swisscom telephone numbers. Next, the names and addresses belonging to these telephone numbers were searched, using direct access to the Swisscom's electronic dictionary.
Survey design: two-wave panel design
Collected by: The commercial survey institute IPSO in Dübendorf, Switzerland, carried out all fieldwork.
Other significant comments: Within each interview, a number of questions was repeated using a different response scale the second time.
