Sensemakr: extending omitted variable bias
The sensemakr package implements a suite of sensitivity analysis tools that makes it easier to understand the impact of omitted variables in linear regression models, as discussed in Cinelli and Hazlett (2020). package
The main function of the package is sensemakr, which computes the most common sensitivity analysis results. After running sensemakr you may directly use the plot and print methods in the returned object.
You may also use the other sensitivity functions of the package directly, such as the functions for sensitivity plots (ovb_contour_plot, ovb_extreme_plot) the functions for computing bias-adjusted estimates and t-values (adjusted_estimate, adjusted_t), the functions for computing the robustness value and partial R2 (robustness_value, partial_r2), or the functions for bounding the strength of unobserved confounders (ovb_bounds), among others.
More information can be found on the help documentation, vignettes and related papers.
# loads dataset data("darfur") # runs regression model model <- lm(peacefactor ~ directlyharmed + age + farmer_dar + herder_dar + pastvoted + hhsize_darfur + female + village, data = darfur) # runs sensemakr for sensitivity analysis sensitivity <- sensemakr(model, treatment = "directlyharmed", benchmark_covariates = "female", kd = 1:3) # short description of results sensitivity # long description of results summary(sensitivity) # plot bias contour of point estimate plot(sensitivity) # plot bias contour of t-value plot(sensitivity, sensitivity.of = "t-value") # plot extreme scenario plot(sensitivity, type = "extreme") # latex code for sensitivity table ovb_minimal_reporting(sensitivity) # data.frame with sensitivity statistics sensitivity$sensitivity_stats # data.frame with bounds on the strengh of confounders sensitivity$bounds ### Using sensitivity functions directly ### # robustness value of directly harmed q = 1 (reduce estimate to zero) robustness_value(model, covariates = "directlyharmed") # robustness value of directly harmed q = 1/2 (reduce estimate in half) robustness_value(model, covariates = "directlyharmed", q = 1/2) # robustness value of directly harmed q = 1/2, alpha = 0.05 robustness_value(model, covariates = "directlyharmed", q = 1/2, alpha = 0.05) # partial R2 of directly harmed with peacefactor partial_r2(model, covariates = "directlyharmed") # partial R2 of female with peacefactor partial_r2(model, covariates = "female") # data.frame with sensitivity statistics sensitivity_stats(model, treatment = "directlyharmed") # bounds on the strength of confounders using female and age ovb_bounds(model, treatment = "directlyharmed", benchmark_covariates = c("female", "age"), kd = 1:3) # adjusted estimate given hypothetical strength of confounder adjusted_estimate(model, treatment = "directlyharmed", r2dz.x = 0.1, r2yz.dx = 0.1) # adjusted t-value given hypothetical strength of confounder adjusted_t(model, treatment = "directlyharmed", r2dz.x = 0.1, r2yz.dx = 0.1) # bias contour plot directly from lm model ovb_contour_plot(model, treatment = "directlyharmed", benchmark_covariates = "female", kd = 1:3) # extreme scenario plot directly from lm model ovb_extreme_plot(model, treatment = "directlyharmed", benchmark_covariates = "female", kd = 1:3, lim = 0.05)
Cinelli, C. and Hazlett, C. (2020), "Making Sense of Sensitivity: Extending Omitted Variable Bias." Journal of the Royal Statistical Society, Series B (Statistical Methodology).
Useful links