Web Survey Bibliography
Title Uncertainty In Web Based Polling
Author Greenberg, A., Bocian, M.
Year 2000
Access date 07.09.2004
Full text pdf (88k)
Abstract Ever since Converse (1964) raised the issue of “non-attitudes” in his classic, “The Nature
of Belief Systems in the Mass Public,” scholars have grappled with questions about how to
handle expressions of uncertainty in survey research. Media and political pollsters customarily
resolve this issue by compelling respondents to make choices about policy and politics regardless
of their level of certainty, interest or knowledge. Interviewers, in fact, are trained to probe
survey respondents to elicit answers and this social interaction has a profound effect on results.
In the Internet format, the absence of an interviewer means that researchers have to make choices
about how to permit respondents to express uncertainty. Using a series of experiments with
InterSurvey’s Internet-based panel, we explore different strategies for reducing the incidence of
the “don’t know” response and the consequences of such strategies for the substantive
interpretation of results. We show that respondents will answer questions in the absence of
direct social pressure. We argue, however, that such strategies mask real uncertainty about
complex public policy issues and political events such as vote preference in the early stages of a
presidential campaign.
of Belief Systems in the Mass Public,” scholars have grappled with questions about how to
handle expressions of uncertainty in survey research. Media and political pollsters customarily
resolve this issue by compelling respondents to make choices about policy and politics regardless
of their level of certainty, interest or knowledge. Interviewers, in fact, are trained to probe
survey respondents to elicit answers and this social interaction has a profound effect on results.
In the Internet format, the absence of an interviewer means that researchers have to make choices
about how to permit respondents to express uncertainty. Using a series of experiments with
InterSurvey’s Internet-based panel, we explore different strategies for reducing the incidence of
the “don’t know” response and the consequences of such strategies for the substantive
interpretation of results. We show that respondents will answer questions in the absence of
direct social pressure. We argue, however, that such strategies mask real uncertainty about
complex public policy issues and political events such as vote preference in the early stages of a
presidential campaign.
Year of publication2000
Bibliographic typeConferences, workshops, tutorials, presentations
