Outlier analysis: Natural resources and immigration policy.

Seung-Whan Choi,Giray Gozgor

doi:10.1371/journal.pone.0261533

Abstract

This replication underlines the importance of outlier diagnostics since many researchers have long neglected influential observations in OLS regression analysis. In his article, entitled “Primary Resources, Secondary Labor,” Shin finds that advanced democracies with increased natural resource wealth, particularly from oil and natural gas production, are more likely to restrict low-skill immigration policy. By performing outlier diagnostics, this replication shows that Shin’s findings are a statistical artifact. When one outlying country, Norway, is removed from the sample data, I observe almost no significant and negative relationship between oil wealth and immigration policy. When two outlying countries are excluded, the effect of oil wealth completely disappears. Robust regression analysis, a widely used remedial method for outlier problems, confirms the results of my outlier diagnostics.

Highlights

Marcus Cicero, the famous Roman political theorist, once said, “by doubting we arrive at the truth.” In the spirit of Marcus Cicero, contemporary political scientists question influential theories and empirical findings that have been widely accepted in the discipline of political science
“Primary Resources, Secondary Labor,” Shin [15] argues that natural resource-rich democracies, those with high revenue from oil and natural gas production, are more likely to restrict low-skill immigration policy
The diagnostic analysis performed after the removal of only one outlying country, Norway, weakens the robustness of the core findings reported in replicated Table 1

Summary

Introduction

Marcus Cicero, the famous Roman political theorist, once said, “by doubting we arrive at the truth.” In the spirit of Marcus Cicero, contemporary political scientists question influential theories and empirical findings that have been widely accepted in the discipline of political science. “seeing is believing.” When I perform graphic outlier diagnostics on the OLS regression models appeared in Shin’s study, I notice flaws in his empirical findings and validity issues with his case selection of Norway that is an outlier rather than a typical oil-rich democracy.

Results

Conclusion