Criminology and Criminal Justice Commons^™

Assessing Bias In Regression Estimates Using Monte Carlo Simulations: Examples In Criminal Justice Research, Matthew P. West, Melissa Rorie, Mark A. Cohen

Graduate Research Symposium (2018 - present)

Can we trust published results? Problems with bias in reported results: “Do social scientists even know anything?” Failed replications (“repligate”). Inaccurate inferences about important relationships (Type I and Type II errors). Inaccurate power analyses for future studies. To avoid these problems, researchers need tools to rigorously evaluate statistical models. The Monte Carlo method is one tool that can be used to evaluate bias in model estimates