Assessing the heterogeneity of social connectedness index via quantile regression mixture model

Abstract

This study aims to visualize a network of Social Connectedness Index (SCI) in Organization for Economic Co-operation and Development (OECD) countries and then explores the importance of socio-demographic, economic, religion, and distance metrics between countries on SCI using a non-parametric test. The final dataset is aggregated from 3 different data sources: Worldbank, OECD, and Facebook. Drawing upon a data set from Facebook Inc. is used to visualize and understand the network structure among OECD countries. Furthermore, the aggregated dataset used in this study is the first usage of Quantile Regression Mixture Models (QRMIX) to determine factors affecting SCI. As a result of the QRMIX model, 4 clusters are identified in different quantiles where the impact of independent factors are differentiated. Based on the variable importance analysis, almost the least important variable at the lower level of SCI value is religion while it is the second most important factor at the highest level of SCI value. SCI mostly shows up as strong relationships between countries with residents of similar ages and education levels where using common language and having same religions, as well. Also, based on the literature review, it is shown that countries with a higher proportion of similar connections to other countries have more positive economic connections among OECD countries. Thus, given the variable importance of SCI for different subgroups of based on SCI quantiles, this study suggests that different action plans about improving import-export and other financial transactions for the country pairs might be created. To sum up, according to different social connection power of OECD countries, this study can help policymakers.