Using Simulation to Analyze Interrupted Time Series Designs.

Area of Science:

• Epidemiology
• Biostatistics
• Public Policy Analysis

Background:

• Interrupted Time Series (ITS) designs are crucial for evaluating policy changes when control groups are unavailable.
• Criminal justice reforms, such as bail setting practice changes, often lack natural comparison groups, necessitating alternative evaluation methods.

Purpose of the Study:

• To present a methodological approach for Interrupted Time Series (ITS) designs using simulation to assess policy impacts.
• To accurately estimate uncertainty in projections when evaluating policy changes with single-unit interventions.

Main Methods:

• Utilizing simulation to generate plausible counterfactual trajectories for comparison with observed data.
• Modeling pre-policy trends with flexibility, including autoregressive components and seasonality.
• Incorporating time-varying covariates and employing nonparametric smoothing for multiple post-policy time points.

Main Results:

• Simulation provides a robust framework for capturing and visualizing uncertainty in ITS designs.
• The approach yields confidence intervals and point estimates for policy impacts.

Conclusions:

• Simulation-based methods offer a reliable way to assess policy impacts using ITS designs, especially in single-unit intervention scenarios.
• This methodology enhances the accuracy and realism of uncertainty assessment in policy evaluations.