A “single step” methodology will be used in the National exit poll-2012
“During the National exit poll-2012 we will use the same methodology as during 2010 presidential election – “single step” methodology”, told director of the Kyiv international institute of sociology Volodymyr Paniotto.
This methodology differs from those of other exit polls in 2010 (except from Savik Shuster’s exit poll), which fixed the quantity of voters needed to be surveyed at the polling stations calculated from the previous exit polls. Such methodology could be valid only if the turnout on the current election does not differ from that during previous one.
““Single step” methodology is just a fixed selection, for instance survey of every 15th voter who comes to the polling station. Such method would guarantee necessary proportions. For example, if on current election only half of those who voted on the previous election come to the polling stations, but we use “single step” methodology, then we will poll two times fewer voters“, the sociologist explained. “On current election, we select 300 polling stations and create two representative subsamples. Razumkov Centre and KIIS will have two separate samples, which will ensure additional control. Then, we will determine the step and – according to that – select voters to poll. Consequently, if the turnout is lower, we will poll fewer voters. During the 2nd round of 2010 presidential elections, such methodology gave the lowest error – less than one per cent”.
Volodymyr Paniotto added that statistical error will also depend on several others conditions, since, for instance, exit poll does not take into consideration foreign polling stations and specialized polling stations, such as military departments, hospitals, prisons etc. “If we are to suppose that someone could guarantee unanimous voting at such places, our results would differ from official ones more than by one per cent”.