Computational consideration for selection of social classes in Romania

Andoria Ioniţă 
National Institute of Statistics, Romania


Middle class is a subject discussed by almost everyone, judging it in most cases from the visible living standard’s point of view: having the ownership of the dwelling, a car, making trips inside country or abroad, buying good quality and expensive goods or services and so on. But, at least in the case of our country, very often there is not a quantitative measurement of middle class, due to the fact that defining correct and reliable criteria to separate this social class from the others isn’t an easy task.
Which are the “latent” factors which ensure each person’s capability to belong to the middle class? How much this affiliation depends on the individual characteristics and how much it depends on external factors like the characteristics of the society in which the persons are living in?
A subtle definition of the middle class has to take into consideration several aspects, some of them more easily or more difficult to measure from the quantitative point of view. We are taking about some quantitative criteria like incomes or the number of endowment goods owned by a person, which are criteria relatively easy to estimate thought statistical methods, but also about aspects like wellbeing or social prestige, variables with a strong subjective specificity, on which there is very difficult to find an accord regarding methods of measurement between different specialists.
This paper presents the results of an attempt to define social classes for Romania, in order to highlight the dimensions and the social importance of the middle class in our country. The elaboration of the methodology to build the social classes starts from the definition of 11 professional categories, based on the Classification of Occupation in Romania. By using the professional categories defined, which can be considered a first instrument (or a first step) for the separation of middle class from the other ones, the present paper presents a first image of the middle class in Romania, analyzing the “chances” one person has to belong to middle class, more exactly by examining the probabilities persons with different characteristics have to belong of this part of population.
This paper can represent an example of good practice in using R product for the analysis of social and economic phenomena in Romania.
Keywords: R, middle class, logistic regression, multinomial regression

[Full Text]

Romanian Statistical Review 3/2015