Skip to main content

rdplot

Function rdplot 

Source
pub fn rdplot(
    y: &[f64],
    x: &[f64],
    c: f64,
    p: usize,
    kernel: Kernel,
    binselect: &str,
    scale_l: f64,
    scale_r: f64,
    nbins_l: Option<usize>,
    nbins_r: Option<usize>,
    wt: Option<&[f64]>,
) -> RdPlot
Expand description

Computes the data behind an RD plot for the evenly-spaced (ES) binselect family: the global polynomial curve coefficients, the IMSE/MV bin counts, and the per-bin scatter means. Ports R rdplot (rdplot.R), ES half only, no covariates and no CI.

Splits the sample at c (x < c left, x >= c right), fits one weighted order-p polynomial per side over the full side range, selects J_l/J_r via the order-4 ES pilot regressions, lays evenly-spaced breakpoints, assigns each point with R’s findInterval convention, and returns per-bin x/y means (empty bins dropped). Bin counts are exact vs R; coefficients and means match to ~1e-8.

§Arguments

  • y: outcome values.
  • x: running variable, parallel to y, need not be sorted.
  • c: cutoff; the point x == c falls on the right side.
  • p: global polynomial order for the curve (R default 4).
  • kernel: kernel for the polynomial-fit weights (uniform/triangular/Epanechnikov).
  • binselect: ES variant string, one of es, espr, esmv (also matched by ""), or esmvpr. Any other value, including the QS variants, triggers unimplemented!.
  • scale_l, scale_r: multipliers on the selected per-side bin count (R default 1). The scaled count is rounded up to an integer; R keeps a fractional count and produces fractional bin ids, so non-integer products diverge from R there.
  • nbins_l, nbins_r: optional manual bin counts that override binselect per side. A side whose outcome has zero variance always gets one bin, as in R.
  • wt: optional unit weights that multiply the kernel weights in the curve fit only; they do not enter the pilot regressions or bin construction.

§Panics

Panics via unimplemented! when binselect is not one of the four supported ES strings (the QS family is a later increment), and when a selector diverges to an infinite bin count (zero spacings variance with non-zero bias on a side, where R errors inside seq).