Predictive models of unemployment: an empirical contribution

Repalust, Anja. (2018). Predictive models of unemployment: an empirical contribution. Diploma Thesis. Filozofski fakultet u Zagrebu, Department of Sociology. [mentor Klasnić, Ksenija and Bagić, Dragan].

Preview

PDF (Croatian) Download (1MB) | Preview

Abstract

This paper discusses the idea that members of certain social groups aren't at the same risk of unemployment. In three measuring points, on aggregated data between 2008 and 2010, between 2011 and 2013 and between 2014 and 2016 which are collected as a part of the Omnibus research conducted by Ipsos, multivariate logistic regression was applied to test the hypothesis that unemployment can be explained with prediction set made up of sociodemographic variables – gender, age, education, occupation, marital status and cohabitation with their own underage children. The hypothesis is confirmed in all three measuring points, but obtained models relatively poorly explain the criterion variable. The occupation of unskilled and low skilled workers, as well as skilled workers, including nurses, and lower level of education, have been shown to be predictors which increase the chances of a person being unemployed the most.

Actions (login required)

View Item