Choosing mode of data collection for surveys to maximize data quality
Stanford University, USA
Survey researchers can now collect data in one of four modes: face-to-face
interviewing, telephone interviewing, self-administered paper questionnaires,
and self-administered computer-presented questionnaires. It might seem that
question-answering will be cognitively equivalent in all four modes, but
psychological theory anticipates differences between the modes in response
accuracy for a variety of reasons. This lecture will begin with a review of
this theoretical background and will then review the findings of empirical
comparing data collected in various modes in order to make recommendations
about mode selection. Part of the discussion will compare the findings of
online surveys of the general public done with probability samples,
recruited through RDD phone calls or face-to-face visits to respondents'
homes (as has been done by organizations in the U.S. Sweden, and Germany,
and is being done in Norway and perhaps other nations as well) vs. online
surveys done with non-probability samples of people who volunteer to answer
questionnaires occasionally for money or gifts or prizes (in response to
online ads or email invitations sent to non-representative groups of people).
The quality of data obtained from these various methods will be contrasted.