
Comprehensive RD Assumption Tests
rd_assumption_tests.RdRuns a battery of standard robustness and assumption tests for regression discontinuity designs: density continuity (McCrary/rddensity), covariate balance at the cutoff, bandwidth sensitivity, placebo cutoffs, and polynomial order sensitivity.
Usage
rd_assumption_tests(
data,
outcome,
running_var,
cutoff = 0,
covariates = NULL,
bw_seq = c(0.5, 0.75, 1, 1.25, 1.5),
placebo_cutoffs = NULL,
poly_orders = 1:4,
kernel = "triangular",
verbose = TRUE
)Arguments
- data
A data frame.
- outcome
Character. Outcome variable name.
- running_var
Character. Running (forcing) variable name.
- cutoff
Numeric. RD cutoff. Default
0.- covariates
Character vector or
NULL. Covariates to test for balance at the cutoff.- bw_seq
Numeric vector. Bandwidth multipliers to test in sensitivity analysis. Default
c(0.5, 0.75, 1, 1.25, 1.5).- placebo_cutoffs
Numeric vector or
NULL. Placebo cutoff values (should be far fromcutoff). IfNULL, auto-selected.- poly_orders
Integer vector. Polynomial orders to test. Default
1:4.- kernel
Character. RD kernel. Default
"triangular".- verbose
Logical. Whether to print test results. Default
TRUE.
Value
A list with:
- density_test
Output from
rddensity::rddensity().- covariate_balance
Data frame of RD estimates on each covariate.
- bandwidth_sensitivity
Data frame of estimates at different bandwidths.
- placebo_results
Data frame of placebo cutoff estimates.
- poly_sensitivity
Data frame of estimates by polynomial order.
- summary_report
Character. Text summary of all tests.