Tests linear hypotheses of the form `contrast `beta` is the coefficient vector from one `misskappa_estimate` or the stacked coefficient vector from several fits on the same subjects.
With one fit, `vcov(fit)` is used. With several fits, `joint_vcov()` is used, so every fit must expose per-subject influence functions and have rows aligned to the same subjects.
Arguments
- ...
One or more `misskappa_estimate` objects. Multiple fits may be named; names become prefixes such as `"ac.Conger"` in the stacked coefficient vector. Unnamed multiple fits are prefixed as `"fit1"`, `"fit2"`, and so on.
- contrast
Numeric vector or matrix defining the linear contrast. For multiple fits, named vector entries or matrix column names are matched against the stacked coefficient names.
- value
Null value for each contrast row. A scalar is recycled.