A Robust Approach for Identifying the Major Components of the Bribery Tolerance Index

The paper aims to emphasize the advantages of several advanced statistical and data mining techniques when applied dense literature on corruption measurements determinants. For this purpose, we used all seven waves World Values Survey employed Naive Bayes technique in SQL Server Analysis Services 2016, LASSO package together with logit melogit regressions raw coefficients Stata 16. We further conducted different types tests cross-validations wave, country, gender, age categories. eliminating multicollinearity, predictor correlation matrices. Moreover, assessed maximum computed variance inflation factor (VIF) against a acceptable threshold, depending model’s R squared Ordinary Least Square (OLS) regressions. Our main contribution consists methodology for exploring validating most important predictors risk associated bribery tolerance. found significant role three influences corresponding questions about attitudes towards property, authority, public services, other people terms anti-cheating, anti-evasion, anti-violence. scobit, probit, average marginal effects build test index based these attitudes. successfully tested using also prediction nomograms accuracy (AUCROC > 0.9).