tasmd_rand

tasmd_rand(x, designs, *, treat_lab=1, mean_tar=None, sd_tar=None, denom='target')[source]

TASMD of one group under one or more designs.

Parameters:

  • x (DataFrame) – Covariates, one column per covariate.

  • designs (dict[str, ndarray]) – Named designs. Each value is one assignment (a vector) or a set of assignments (one per row); all values must be the same kind.

  • treat_lab (object) – Label of the group assessed.

  • mean_tar (ndarray | None) – Target covariate means.

  • sd_tar (ndarray | None) – Target covariate standard deviations.

  • denom (Literal['target', 'group']) – Which standard deviation to divide by.

Returns:

For vectors, a covariate-by-design table of TASMDs. For sets of assignments, their mean and standard deviation across assignments.

Raises:

ValueError – If designs mix single assignments and sets.

Return type:

DataFrame | TasmdSummary