Abstract

Background: The socioeconomic dynamics and economic status of households are changing rapidly all over the world, especially for middle-income and developing countries such as Turkey. It is well known that income is one of the most important determinants of socioeconomic status (SES) and health. There is a need to analyze changing trends and relationships among SES variables over time. Unfortunately, such changes remain undetected in cross-sectional studies. However, panel data cover multiple sequential time periods for the same firms or individuals and allow analysis of these changing patterns. Objectives: This study aims to determine predictive factors of income on a decision tree and to observe changes of predictive role of SES and health indicators over time, to predict income using three different panel datasets.
Materials and Methods: Data from the Income and Living Conditions Survey in Turkey between 2008 and 2011 were used. A random effects and expectation maximization tree algorithm were also used in the analysis.
Results: This study results show that education is the variable that is the most influential in predicting income between 2008 and 2011 and between 2009 and 2011; on the other hand, current economic status is a determining factor of income between 2010 and 2011. Conclusion: Experimental results highlight that education outperforms SES and health indicators in the prediction of personal income.