tsdive.SwitchbackEstimate
dataclass
¶
SwitchbackEstimate(
covariates: tuple[str, ...],
kept_per_block: tuple[int, ...],
*,
estimate: float | None = None,
p_value: float | None = None,
lo: float | None = None,
hi: float | None = None,
reason: str | None = None,
detail: str | None = None,
)
One analysis of the target: direct (no covariates) or adjusted.
estimate is B minus A in the target's unit. lo and hi bound
the 95% interval and are -inf or +inf on a side it does not close on.
On a refusal reason is one of too_few, no_spread,
too_many_covariates, empty_block or collinear, detail
says what the data showed, and the numbers are None.
Attributes:
| Name | Type | Description |
|---|---|---|
covariates |
tuple[str, ...]
|
|
kept_per_block |
tuple[int, ...]
|
|
estimate |
float | None
|
|
p_value |
float | None
|
|
lo |
float | None
|
|
hi |
float | None
|
|
reason |
str | None
|
|
detail |
str | None
|
|
n_kept |
int
|
|
refused |
bool
|
|
lo_unbounded |
bool
|
|
hi_unbounded |
bool
|
|