Several latest elections and referendums were marked by the dramatic failure of electoral forecasts based on mass polls. To respond to the dissatisfaction of the public and politicians, alternative approaches like prediction markets, Implicit Attitude Test (IAT), expectationbased forecasts and so on were developed. IAT proves to be one of the most efficient ways to enrich the forecasting models and improve their accuracy. The problem is that the original form of IAT implies too rigid rules to be applied in the traditional mass poll. As a thorough laboratory-style measurement of nervous reactions to stimuli, IAT requires a special environment, for instance, nothing should disturb or distract respondents from performing experimental tasks. Such an environment is difficult to provide during the mass poll’s fieldwork; thereby, researchers usually implement IAT on small samples. This article presents the Graphic Associative Test of Attitude (GATA) as a tool for mass polls. It is the IAT’s functional analog developed by the author and tested in a wide range of preelectoral mass polls in Russia. GATA is easy to use even with inexperienced interviewers, and its simple and intuitive-clear tasks do not create additional barriers for respondents and do not decrease the response rate. At the same time, in a reliable way, GATA identifies implicit factors of behavior and helps to improve the accuracy of forecast. As a theoretical research, this study proves the ‘dual attitude’ concept of the structural theory of attitude.
Read full abstract