compare_models: Compare two models using the debiased score test

In dScoreTest: Debiased Score Tests for Goodness of Fit and Model Comparison

Compare two models using the debiased score test

Description

Compare two models using the debiased score test

Usage

compare_models(fit.0, ...)

Arguments

fit.0	A fitted null-model object. Methods are provided for glm, lm and mgcv::gam fits.
...	Additional arguments passed to the dispatched method, notably the alternative-model fit fit.1.

Value

An object of class "dScoreTest": a list whose key elements are the debiased test statistic t.stat and the one-sided p-value p.val (right tail of the standard normal), along with the test-set score residuals, the hunted direction, and the call. It has print, summary and plot methods.

References

Dhawan, A., Guo, F. R. and Shah, R. D. (2026). The debiased score test: hunt-and-test for semiparametric hypotheses. arXiv:2607.28861. https://arxiv.org/abs/2607.28861

See Also

compare_models.glm, compare_models.lm, compare_models.gam, gof_test, dScoreTest

Examples

 set.seed(42)
 n <- 500
 dat <- data.frame(x1 = rnorm(n), x2 = rnorm(n), x3 = rnorm(n))
 dat$x3 <- dat$x3 + (dat$x1 + dat$x2) / 3
 dat$y <- 5 * exp(dat$x1 + dat$x3) + rnorm(n) * 3
 fit.0 <- glm(y ~ x1 + x3, family = gaussian(link = "log"),
              data = dat, start = rep(1, 3))
 fit.1 <- glm(y ~ x1 + x2 + x3, family = gaussian(link = "log"),
              data = dat, start = rep(1, 4))
              
 # test fit.0 against fit.1: should not be rejected
 compare_models(fit.0, fit.1)
 compare_models(fit.0, fit.1, hunt.style="wls")
 anova(fit.0, fit.1)

 # test a misspecified null model: should be rejected
 fit.00 <- glm(y ~ x2, family = gaussian(link = "log"),
              data = dat, start = rep(1, 2))
 compare_models(fit.00, fit.1)
 plot(compare_models(fit.00, fit.1))
 compare_models(fit.00, fit.1, hunt.style="wls")
 plot(compare_models(fit.00, fit.1, hunt.style="wls"))
 anova(fit.00, fit.1)

dScoreTest documentation built on Sept. 3, 2026, 1:06 a.m.