Skip to contents

Wald inference on a linear combination of fixed effects

Source: R/inference.R
mm_lincomb.Rd

Convenience helper for the common case of testing \(H_0:\, c^\top \beta = 0\) where c is a sparse, named weight vector. The estimate is \(c^\top \hat\beta\), the standard error is \(\sqrt{c^\top V c}\) where V is the model's fixed-effect covariance, the statistic is the Wald ratio, and the interval is the symmetric Wald CI at level.

Usage

mm_lincomb(fit, weights, level = 0.95, method = NULL, ...)

# Default S3 method
mm_lincomb(fit, weights, level = 0.95, method = NULL, ...)

# S3 method for class 'mm_glmm'
mm_lincomb(fit, weights, level = 0.95, method = NULL, ...)

# S3 method for class 'mm_lmm'
mm_lincomb(
  fit,
  weights,
  level = 0.95,
  method = c("auto", "satterthwaite", "kenward_roger", "asymptotic"),
  ...
)

Arguments

fit

A fitted mm_lmm or mm_glmm.

weights

A named numeric vector (or named list / single-row data.frame coercible to one). Names must match names(fixef(fit)) exactly.

level

Confidence level for the Wald interval. Default 0.95.

method

For mm_lmm, the degrees-of-freedom method passed to df_for_contrast(). Defaults to "auto" (Satterthwaite when available). For mm_glmm, only "asymptotic" is accepted.

...

Reserved for future methods.

Value

A single-row data.frame with columns estimate, std_error, statistic, statistic_name ("t" or "z"), df, p_value, lower, upper, and method. The result carries an "mm_status" attribute reflecting the underlying vcov reliability (status, method, reliability, reason).

Details

For mm_glmm, the statistic is the asymptotic Wald z (no df). For mm_lmm, the default is Satterthwaite-approximated t via df_for_contrast(); pass method = "asymptotic" to force Wald z.

Weight names must be a subset of names(fixef(fit)). Coefficients not named in weights contribute zero. Pass the long-form contrast() front door if you need multiple contrasts or non-default rhs.

See also

contrast() for the long-form, Rust-routed contrast surface with full estimability / reliability reporting.

Examples

if (FALSE) { # \dontrun{
# Difference-in-differences contrast at a focal SOA = 25 ms
# (Loo et al. 2026 aphantasia primary estimand, glmm path)
soa_s_25 <- (log(0.025) - mean(fit$data$soa_log)) / sd(fit$data$soa_log)
mm_lincomb(fit, c(
  "group: aphant:mask: masked"        = 1,
  "group: aphant:mask: masked:soa_s"  = soa_s_25
))
} # }