Page 49 - An Introduction to Political Communication Second Edition

AN INTRODUCTION TO POLITICAL COMMUNICATION

The aforementioned are conceptual difficulties, arising from the
complexity of the communication process itself. They remind us that
successful communication of a message (political or otherwise) cannot
be taken for granted, but must be worked for by the sender.

The evidence
A further problem for political communication research concerns
the nature and quality of the evidence used to measure effects. There
are, in the final analysis, only three ways to assess the effects of
political communication on attitudes and behaviour. The first is to
ask people how they have responded to specific messages, and then
collate their responses into statistically significant aggregates, usually
in the form of public opinion polls. Secondly, one may observe voting
behaviour, relating this to the communication strategies of the
contestants in a political campaign. Thirdly, one may conduct
experiments intended to isolate the effects of particular elements of
the communication process. Each of these data-gathering techniques
has its methodological limitations.

Surveys
Public opinion polling, for example, a technique which originated
with commercially motivated survey sampling in the 1930s, depends
for its accuracy on the application of sampling procedures which
permit the survey to be ‘representative’. The questions asked of those
polled must be carefully formulated so as to avoid distortion,
simplification, and exaggeration of response. The timing of polls
must be taken into account, and results interpreted cautiously, with
allowances made for a variety of potential sources of error. While
the best-known and most frequently used polling organisations, such
as Gallup, MORI, and NOP take considerable time and money to
achieve the maximum degree of accuracy possible, many opinion
polls, particularly those conducted independently by print and
broadcast media, do not. As the 1992 British general election showed,
even the established pollsters may get it substantially wrong when
attempting to predict election outcomes. 3
Public opinion polls are not only a measure (however imperfect)
of political attitudes and intentions at a given point in time. Many
observers agree that they can become a causal factor in voting
behaviour. American news broadcasters have come under pressure
in recent presidential elections to delay releasing the findings of their

32