Draft V2.1 – do not distribute – do not quote without permission
1
Media effects in Voting and Polling:
E-Democracy, Trust and Social Identity
Abstract: Internet voting enables citizens to vote from different places on various
moments, and this is generally seen as one of the advantages that may positively
influence voting turnout. However, social psychological theory suggests that social
identity is dependent on the situation a person is in, and the voting decision (as well as
other behavior) may be different depending on these different identities. If this is true,
the outcome of a vote or opinion poll may be influenced by the deployment of Internet
based media. Voters may also have different levels of trust in the various voting media,
and this also may influence voting behavior. Findings from some field experiments with
an e-voting system seem to corroborate these expectations.
Introduction
Voting procedures and technologies have to guarantee that elections are free and
secure. For example, the polling station creates a social environment in which one can
vote without direct or indirect interference by others. Therefore political propaganda is
not allowed in the polling station, and in some countries it is forbidden to have polls in
the days before the election, to avoid that voters will be indirectly influenced by (the
vote of) others. The paper-based voting technology guarantees that no link can be
made between the individual voter and his/her vote. And when the public is allowed to
observe the counting, this procedure guarantees that votes are counted correctly.
Finally, comparing the number of votes with the number of voters easily shows whether
all votes have been counted. Most of these safeguards are based of procedures, and
rely on the social context of the ballot. E-voting may seriously change this. If voters
consider the new technology as less privacy protecting, as less secure, and as less
accountable, they may change their voting behavior. This may influence their decision
whether to participate in the vote, but also what they vote, and therefore the outcome of
the ballot.
1
In this paper, we try to investigate this empirically, based on an experiment
with an internet voting system.
1
The e-voting technology may also be used for opinion polling. A similar set of problems exists here, as
one does not want the polling media to have an independent influence on the participation in the poll, nor
on the result. After thinking about the influence of the interviewers’ behavior, of the posing of the
Page 1 of 13 
Convention