Skip to the main content

Original scientific paper

Case Study: Analysis of the Puls Agency Exit Poll Methodology

Dragan BAGIĆ orcid id orcid.org/0000-0002-0371-838X ; Faculty of Philosophy; Zagreb
Vesna LAMZA POSAVEC ; Institute of Social Sciences Ivo Pilar; Zagreb


Full text: croatian pdf 246 Kb

page 695-721

downloads: 1.470

cite


Abstract

In an attempt to improve exit polls in Croatia, a detailed
analysis was carried out of the exit poll methodology used by
the agency Puls in the last few elections. The results of the
analysis showed that the main biases of the applied
methodological approaches were related to the coverage of
potential respondents, mostly due to survey non-response,
biased sampling procedures and, consequently to inefficient
statistical correction procedures aimed at reducing such
research biases. In addition to a more cautious interpretation
and public presentation of results, it would be best in the
future to improve the existing and develop some new
methodological solutions, especially predictive models of
election outcomes in which the exit polls data would be one
of its constituent elements.

Keywords

exit polls; public opinion research; survey methodology

Hrčak ID:

29432

URI

https://hrcak.srce.hr/29432

Publication date:

31.10.2008.

Article data in other languages: croatian german

Visits: 3.577 *