BUSINESS’ PARTICIPANTS MOTIVATION IN OFFICIAL SURVEYS BY FUZZY LOGIC

Abstract

Business statistics, collected and published by official institutes, are important for policy makers and for businesses in their management decisions. In spite of that, business participation in surveys is decreasing. This causes a decrease in quality of data and increases demands for data imputation procedures, endangering the value of survey based data collection. Improving the motivation of respondents could significantly influence the validity of business survey research. To achieve this, we suggest applying fuzzy logic to both dissemination tools for statistical data and adaptive survey design methods to tailor survey design to the characteristics of businesses. Adaptive survey designs (ASDs) take into consideration the fact that the impact of various design features varies significantly over respondents. When auxiliary information is available from e.g. registry data, survey designs may be tailored to optimize response rates through strategies designed to enhance motivation. In a crisp classification design businesses or respondents although having similar attribute values, may be classified into different classes and therefore receive different treatments. Applying fuzzy logic to an ASD may lead to a more robust and effective classification when aiming at improving unit and item response and decreasing measurement error through the stimulation of motivation. Statistical institutes provide data on their data portals. However, data users typically find it hard to distill the relevant data and information on these websites, which could demotivate them to cooperate in surveys. Therefore, in dissemination we should include methods capable of processing imprecise queries and methods for mining rules in the data.