Stats pointbiserial test

Development build. This page describes main, not a released package. The latest published Lodestar.Stats is 0.5.0 — read its documentation.

Home › Stats › Hypothesis tests

PointBiserial.Test

Correlates a binary variable with a continuous one, scipy.stats.pointbiserialr.

public static TestResult Test(ReadOnlySpan<bool> x, ReadOnlySpan<double> y, NanPolicy nanPolicy = NanPolicy.Propagate)

Parameters — x is the binary variable and y the continuous one, the same length. nanPolicy is what to do with a NaN in y; NanPolicy.Omit drops the pair.

Returns — TestResult: the coefficient and its two-sided p-value, both NaN when either side is constant or fewer than two pairs remain, where scipy answers rather than raises.

Exceptions — ArgumentException when the samples differ in length, or when a NaN meets NanPolicy.Raise.

Example — whether the students who passed studied longer.

using Lodestar.Stats;

bool[] passed = [true, false, true, true, false, true, false, false, true, true];
double[] hours = [12.0, 4.5, 9.0, 14.0, 6.0, 10.5, 3.0, 7.5, 11.0, 8.0];

TestResult result = PointBiserial.Test(passed, hours);

double r = Math.Round(result.Statistic, 6);   // => 0.824775
double p = Math.Round(result.PValue, 6);      // => 0.003319

Remarks — scipy offers no alternative here, and neither does this; for a one-sided reading, Pearson.Test on the coded variable takes one.

Applies to — net10.0, netstandard2.0.

See also — Pearson.Test.