Purpose: There is a significant educational migration in Turkey, and if life satisfaction is not improved, it is expected that this migration may increase. The aim of this study was to determine the impact effective socio-economic and educational variables in migration using life satisfaction survey data of Turkish Statistical Institute, and to calculate the numerical coefficient values of these variables to be used by policy makers for investments.
Research Methods: Two types of econometric models were used to determine the effective variables in migration. Outlier observations were detected, and their negative effects were corrected with the help of robust regression methods. This paper provides evidence of how outliers changed the statistics and test results. In addition, multi-collinearity corrected estimates were calculated.
Findings: The most significant variables in migration were the gross domestic product per capita and education variable. Using life satisfaction index values, educational and related migrations can be reduced. This paper also provides evidence of how outliers in data changed the statistically significant variables, estimates, normality and heteroscedasticity in the test results.
Implications forResearch and Practice: Migration can be reduced by increasing life satisfaction and lowering dissatisfaction in essential and non-essential municipality service variables. Using the methods in this paper and using future indices that are going to be published it is possible to take countermeasures for migration using models with higher explanatory power.
Keywords:education, robust regression, multicollinearity, outlier, life satisfaction