Hybrid models of Random utility maximization and random regret minimization: Results from two empirical studies