The Use of Regression Trees in the Study of Dependencies between the Employment Efficiency Index and the Structure of the Unemployed in County Employment Offices

Abstract

The aim of the article is to analyse the relationship between the employment efficiency index and the structure of unemployment in county ('powiat') employment offices in Poland. Five research hypotheses connected with the research goal were formulated. The most important features characterizing the unemployed were: education, age, the duration of unemployment, gender, the place of residence. Regression trees were used as the research tool, which enabled to group county employment offices with differing values of the employment efficiency index with a simultaneous indication of independent variable variants (diagnostic features) affecting its level. After conducting the research, the authors managed to verify positively and unquestionably the hypothesis that the higher the percentage of the unemployed with higher education, the easier it is to achieve higher employment efficiency. In a less explicit way, only under certain conditions, hypotheses regarding the age of the unemployed and the place of residence were verified. Nevertheless, it was not possible to verify positively the hypotheses regarding the impact of gender and the duration of unemployment on the employment efficiency index.