The results of the MC simulations with a nonnormal Y variable are in Table 4. With m = 750, the relative parameter bias values are outside of the specified limits for all three variables, as is the relative SE bias and coverage for X1, and the coverage X3. The aberrant bias values decrease minimally with a m of 10,000, indicating that using a typical regression model for this data will produce biased regression coefficients.
Compared to the results from Example 1, the power decreased for X3 and the 95% CI half-widths increased for all three predictors. As X1's relative SE bias is somewhat large, its CI will likely be inaccurate so the half-width should be interpreted cautiously.
The results of the MC simulations with a nonnormal Y variable are in Table 4. With m = 750, the relative parameter bias values are outside of the specified limits for all three variables, as is the relative SE bias and coverage for X1, and the coverage X3. The aberrant bias values decrease minimally with a m of 10,000, indicating that using a typical regression model for this data will produce biased regression coefficients.Compared to the results from Example 1, the power decreased for X3 and the 95% CI half-widths increased for all three predictors. As X1's relative SE bias is somewhat large, its CI will likely be inaccurate so the half-width should be interpreted cautiously.
การแปล กรุณารอสักครู่..
