W poniedziałek 21 listopada o 14:30 na seminarium POLPAN wystąpiła dr Barbara Jancewicz (Instytut Socjologii Uniwersytetu Warszawskiego), która przedstawiła referat pt. „Inequality and its perception: in search of sources of multidimensionality using multidimensional scaling”.
Abstrakt:
Researchers are trying to link income inequality with multiple negative social phenomena such as crime, distrust and even increased mortality (Wilkinson and Pikett, 2011). We evaluate income inequality using specially designed measures: each of them represents one of many concepts of what inequality is. Economists disagree as to what exactly inequality is, so they use several different definitions simultaneously. Amiel and Cowell (1999) decided to see how the definitions used by the experts correspond to those used by ordinary people: they initiated a series of studies measuring perception of income inequality.
Research on income inequality assumes that each respondent is rational and has a one-dimensional definition of inequality. It is understandable that sometimes respondents need time to think through their answers in order to express their opinion about inequality, but in the end a one-dimensional and coherent response is expected. Results of previous studies have shown that the reality is much more complex, and the responses are often vague. Inspired by the remark made by Amartya Sen (1973) – indicating that inequality has many aspects – a study was conducted, it showed that respondents’ answers were best reflected by a two-dimensional map (Jancewicz, 2014). This study continues this analysis: it tests whether the two-dimensionality results from differences between respondents or perhaps it appears already in the assessments of individual respondents.
A custom made questionnaire was prepared in order to verify whether individuals perceive and define inequality in many dimensions instead of one. Each respondent’s answer was visualized separately using multidimensional scaling in one, two and three dimensions. The errors of each representation (compared with errors corresponding to random answers) allowed to conclude that individual responses were best represented by a single axis. This result is in accordance with assumptions made by economists. While individual perceptions of inequality were one-dimensional, the differences in perceptions of different respondents required the visualization of joint answers to be two-dimensional.
Innovative questionnaire and unusual use of multidimensional scaling constitute the methodological innovation of this study. Formulated conclusions help us understand how people perceive income inequality.