Table 2. Parameters used in simulation scenarios for the meta-analysis power analysis.
Model | Scenario | β0 | β1 | β2 | n Studies |
---|---|---|---|---|---|
Non-Linear/categorical | 25% Linear Decrease | -0.29 | -0.41 | -1.1 | 5, 10, 20, 30 |
50% Decrease then Stable | -0.69 | 0 | 0 | 5, 10, 20, 30 | |
50% Decrease then 10% Decline | -0.69 | -0.22 | -0.51 | 5, 10, 20, 30 | |
50% Exponential Decrease | -0.69 | -0.69 | -0.99 | 5, 10, 20, 30 | |
Linear/continuous | 25% Linear Decrease | -0.59 | -0.59 | 5, 10, 20, 30 | |
50% Decrease then Stable | -0.52 | -0.27 | 5, 10, 20, 30 | ||
50% Decrease then 10% Decline | -0.7 | -0.5 | 5, 10, 20, 30 | ||
50% Exponential Decrease | -1.04 | -0.89 | 5, 10, 20, 30 |
A total of four scenarios for each model type were run with different β0, β1, and β2, values (the parameter estimates that define the relationship between fatality reduction and changes in cut-in speed) and differing numbers of studies per category (n Studies). Parameters are log-transformed.