Statistical Issues in Modelling Happiness Level of Immigrants: An Investigation with World Happiness Report, 2018

Authors

  • Ardhendu Banerjee Department of Statistics, Calcutta University, Kolkata, India
  • Subrata Chakraborty Department of Statistics, Dibrugarh University, Dibrugarh, India
  • Aniket Biswas Department of Statistics, Dibrugarh University, Dibrugarh, India

Keywords:

Binary regression, link function, cross-validation, AUC

Abstract

World Happiness Report (WHR) released in 2018 among others, ranked the countries around the world with respect to the happiness level of immigrants measured in ladder-score from 0 to 10. Regression analysis with happiness score as response and several important determinants (covariates) has also been reported in that study with usual least square assumptions for finding important covariates and prediction purposes. First, we point the statistical problem out in doing so and attempt modeling this happiness level by first dichotomizing the response (as either happy or unhappy) and then employing binary regression with the given covariates. The risk associated with miss-specification of the link functions is demonstrated by considering four popular choices and a new data driven computational routine based on assessment metrics and cross validation is prescribed to choose the best link function. Important covariates are reported thereafter considering the best choice.

Downloads

Download data is not yet available.

Downloads

Published

2021-12-30

How to Cite

Banerjee, A. ., Chakraborty, S. ., & Biswas, A. . (2021). Statistical Issues in Modelling Happiness Level of Immigrants: An Investigation with World Happiness Report, 2018. Thailand Statistician, 20(1), 195–206. Retrieved from https://ph02.tci-thaijo.org/index.php/thaistat/article/view/245859

Issue

Section

Articles