Method for examining the variability of socio-economic objects quoting the example of Baltic Sea countries Kesra Nermend Faculty of Economics and Management University of Szczecin ul. Mickiewicza 64, 71-101 Szczecin, Poland kesra@wneiz.pl Abstract The article presents the method for construing vectorial synthetic measure which allows to determine the effect of indicators’ variability on the result of ordering. In order to construe the measure, the authors used ordered pair, mean value increment and standard deviation increment. As a consequence, additional information was gathered for every value of the measure. This information allowed to define how the position of a given subject in the ranking might change. The authors present exemplary ways of construing vectorial synthetic measure for creating the ranking of regions situated in Baltic Sea countries taken the education of children and young people into account. Keywords: Vectorial synthetic measure, synthetic measure, taxonomy, vector calculus, arithmetic of increments, image processing, regional analysis, linear ordering methods. 1. Introduction Vectorial synthetic measure belongs to the group of the linear objects ordering methods, forming the branch of taxonomy. One of the currently most popular method of linear objects ordering has been proposed by Hellwig [Hellwig 1968], so called Hellwig's synthetic measure. Hellwig has introduced the basic notions of the method, e.g. stimulants and de- stimulants. The Hellwig's method finds many applications, e.g. in construction of synthetic variables in the process of econometric modeling [Bartosiewicz 1984], determination of the product quality [Borys 1984], survey of regions development [Mlodak 2006], investigations of stock investment attractiveness [Tarczynski 2006], determination of social status [United 2010], poverty [Social 2010], famine [GHI 2010], ease of running a business [World Bank 2010], competitiveness [World Economic 2010]. Application of the Hellwig's synthetic measure is slightly limited by the assumption that the template has to be the object, being the best in all aspects. This excludes the usage of real templates, for which as a rule not all coordinates in the feature space are greater than for all other objects. Also it is the reason of the constructed measures sensitivity to existence of objects with exceptionally big values of coordinates, in the set of investigated objects, what can greatly influence the ranking result. To allow using the real templates, which are not