摘要：The article concerns elections to the Senate of the Czech Parliament. Primarily, it refers to potential shifts of voter preference between the first and second rounds of the Senate elections. The fundamental tool for determining this trend is the Bayesian ecological inference model, through the use of which the authors attempt to answer the question of how the voters from the first round voted in the second round one week later when the options were reduced to only two candidates. The text analyzes these processes in detail in five electoral districts.
