Mogući utjecaj objavljenih rezultata predizbornih anketa na izborno ponašanje birača
The Possible Influence of Published Results of Pre-Election Polls on the Behaviour of Voters in Elections

Summary/Abstract: In the research the authors tried to assess the possible influence of published results of pre-election polls on voters' preferences and intentions to vote. The starting hypothesis was that a direct influence on voters was possible if the following conditions were fulfilled: that they had seen or 95 heard the results of such polls, that they knew their results and believed in their credibility. The data were collected on representative samples of the Croatian adult population, in the last three days of the 2003 election campaign for the Croatian Parliament and the first round of the 2005 Presidential Elections. The results suggest firstly that even the theoretical influence of published polls at the time was relatively small. It was further indicated that in the 2003 research those who could have been theoretically influenced were mostly among the potential voters for the SDP, which came in second at the elections, and in the research for the 2005 elections those influenced were among the possible voters for the leading presidential candidate Stjepan Mesić. However, differences have been observed that positively correlate with the level of education and some other demographic characteristics of the examinees, but not with the indicators of the possible influence of pre-election polls. The results are in accordance with findings of most of the research conducted worldwide, stating that there is no conclusive scientific evidence of a significant direct influence of published results of public opinion research on the electoral behaviour of individuals.

• Details
• Contents