fix(results): exact fractional confidence-level labels family-wide

Every text surface that names a confidence level computed the percent with
int((1 - alpha) * 100) (truncation), int(round(...)) (rounding), or :.0f, so
alpha=0.025 printed "97%" or "98%" over a 97.5% interval. All 34 sites now
route through one shared formatter in results_base (_coverage_level /
_coverage_pct, plus _alpha_pct for the dCDH "Significant at" line, which
printed 2% for alpha=0.025). Default-alpha output is byte-identical.

BusinessReport headline.ci_level carries the exact level: int when integral
(95 unchanged), float otherwise (97.5). No schema-version bump (REPORTING.md
Note). tests/test_coverage_label.py pins the formatter, the representative
surfaces at alpha=0.025, and adds a source guard against reintroducing an
inline percent computation. Retires the TODO.md row.

Author: igerber

TODO.md

@@ -57,7 +57,6 @@ Related tracking surfaces:
 | `WooldridgeDiD` does not apply the W2025 Sec 5.4 `D_{G_max} x X` covariate normalization, and three sibling covariate rank deficiencies are pre-existing. Measured with the period range pinned and only the never-treated units toggled: (1) time-invariant `exovar` is absorbed by the unit FE, 4 of 26 columns, IDENTICALLY with and without never-treated units; (2) `xgvar`'s cell x covariate block, 19 of 41, identical on both panels; (3) `xtvar` under `demean_covariates=False` does exhibit the `sum_g D_g x = x` dependency that the default demeaning removes; (4) the newly-reachable case -- time-VARYING data passed through `exovar`, which its own docstring reserves for time-invariant covariates -- where the paper's `dT_i` rule would give a deterministic `D_{G_max} x X` drop instead of QR's arbitrary pick (coefficients unaffected, `1.35e-14`; `rank_deficient_action="error"` raises). REGISTRY's narrowed Sec 5.4 note cross-references this row. **Trap for whoever takes it:** `xtvar` under the DEFAULT `demean_covariates=True` is FULL RANK -- the raw block carries demeaned values while `D_g x X` carries raw ones -- and forcing the drop there moves `overall_att` 1.11903 -> 1.46269. Pinned as-is by `TestComparisonSupportFiltering::test_cells_derived_groups_did_not_leak_into_the_design`. | `diff_diff/wooldridge.py` | #729-followup | Heavy | Medium |
 | `WooldridgeDiD.n_control_units` counts never-treated UNITS on `control_group="never_treated"` regardless of method, but on the nonlinear paths (`logit`/`poisson`) treated units' pre-treatment rows ARE the identifying comparison -- only the OLS path absorbs them into their own cells. So the reported count under-states the comparison pool exactly where the REGISTRY control-pool asymmetry note applies. Widen to `not_yet_treated or (never_treated and method != "ols")`, or document the count as never-treated-units-by-definition. Behavior is PRE-EXISTING; documented for now in the REGISTRY control-pool Note rather than changed, because widening moves a public results field and wants its own ledger row and test matrix. | `diff_diff/wooldridge.py` | #729-followup | Mid | Low |
 | `WooldridgeDiD` has no opt-out for comparison-support period filtering: a user who would rather see the refusal than a reduced sample cannot ask for it. Adding one means a constructor parameter (`get_params`/`set_params` propagation, transactional validation), a ledger row, and a test matrix across both predicate branches and all three `rank_deficient_action` modes -- deliberately out of scope for the change that introduced the filter. The always-on warning is the interim answer. | `diff_diff/wooldridge.py` | #729-followup | Mid | Low |
-| Fractional confidence-level display unification: every `summary()` header computes `int((1 - alpha) * 100)` (~19 truncating sites incl. the shipped M-146 staggered family), so a fit `alpha=0.025` prints "97% Confidence Interval" over a 97.5% interval; `plot_dose_response`'s band label now renders exact fractional coverage via its `_coverage_label` helper - unify the summary surfaces on the same formatter (cross-family display change: moves every summary golden/doctest that pins a header, so it wants one sweep with its own test recapture, not a per-estimator drip) | `diff_diff/results_base.py`, `diff_diff/results.py` | alpha-guard review | Quick | Low |
 
 ### Performance
 

changelog.d/20260903-fractional-confidence-label.md

@@ -0,0 +1,19 @@
+### Fixed
+- **Exact fractional confidence-level labels, family-wide**: every text surface
+  that names a confidence level now prints the exact coverage (`97.5%` for
+  `alpha=0.025`; previously truncated to `97%` by `int((1 - alpha) * 100)` or
+  rounded to `98%` by `int(round(...))` / `:.0f`) via one shared
+  `results_base._coverage_pct` formatter: the 14 `summary()` headers
+  (DiD/TWFE/MultiPeriod/SyntheticDiD, CallawaySantAnna, staggered and 2x2x2
+  TripleDifference, StackedDiD, ImputationDiD, TwoStageDiD, EfficientDiD,
+  ContinuousDiD, dCDH, SunAbraham, TROP), the `EventStudyResults` / HAD / RDD /
+  ETWFE / LWDiD / LPDiD / ChangesInChanges table headers, the CS and dCDH sup-t
+  band labels, the dCDH HonestDiD block (whose "Significant at" line printed
+  `2%` for `alpha=0.025`; now `2.5%`), `WildBootstrapResults` and the LWDiD
+  wild-cluster-bootstrap summaries, `HonestDiDResults` / `PlaceboTestResults`
+  summaries, and BusinessReport / DiagnosticReport prose. The BusinessReport
+  headline `ci_level` field carries the exact level as an `int` when integral
+  (`95` is byte-unchanged) and a `float` otherwise (`97.5`); no schema-version
+  bump (REPORTING.md Note). Default-alpha output is byte-identical. A source
+  guard (`tests/test_coverage_label.py`) rejects any reintroduced inline
+  percent computation.

diff_diff/business_report.py

@@ -48,7 +48,7 @@
 
 from diff_diff._reporting_helpers import describe_target_parameter
 from diff_diff.diagnostic_report import DiagnosticReport, DiagnosticReportResults
-from diff_diff.results_base import Diagnostic
+from diff_diff.results_base import Diagnostic, _coverage_level, _coverage_pct
 
 BUSINESS_REPORT_SCHEMA_VERSION = "2.0"
 
@@ -534,7 +534,7 @@ def _build_schema(self) -> Dict[str, Any]:
                 "ci_upper": None,
                 "alpha_was_honored": True,
                 "alpha_override_caveat": None,
-                "ci_level": int(round((1.0 - self._context.alpha) * 100)),
+                "ci_level": _coverage_level(self._context.alpha),
                 "p_value": None,
                 "is_significant": False,
                 "near_significance_threshold": False,
@@ -683,7 +683,7 @@ def _extract_headline(self, dr_schema: Optional[Dict[str, Any]]) -> Dict[str, An
                 f"for the confidence interval because this fit uses "
                 f"{inference_label} inference; the displayed CI remains "
                 f"at the fit's native level "
-                f"({int(round((1.0 - result_alpha) * 100))}%). The "
+                f"({_coverage_pct(result_alpha)}%). The "
                 f"significance phrasing still uses the requested alpha."
             )
 
@@ -700,7 +700,7 @@ def _extract_headline(self, dr_schema: Optional[Dict[str, Any]]) -> Dict[str, An
         )
         if att is None or not np.isfinite(att):
             sign = "undefined"
-        ci_level = int(round((1.0 - display_alpha) * 100))
+        ci_level = _coverage_level(display_alpha)
         # bool(...) coerces away numpy bool_ — when ``p`` is a numpy NaN (e.g.
         # SyntheticControl, whose analytical p_value is always NaN), ``np.isfinite``
         # yields a numpy bool that is NOT JSON-serializable in the schema.
@@ -2073,7 +2073,7 @@ def _significance_phrase(p: Optional[float], alpha: float) -> str:
     """
     if p is None or not np.isfinite(p):
         return "statistical significance cannot be assessed (p-value unavailable)"
-    ci_level = int(round((1.0 - alpha) * 100))
+    ci_level = _coverage_pct(alpha)
     if p < 0.001:
         return "the direction of the effect is strongly supported by the data"
     if p < 0.01:

diff_diff/chaisemartin_dhaultfoeuille_results.py

@@ -34,7 +34,13 @@
 from diff_diff._deprecation import deprecated_field_property
 from diff_diff.aggregation import AggregationMixin, AggregationResult
 from diff_diff.results import _get_significance_stars
-from diff_diff.results_base import BaseResults, _require_fit_alpha, build_event_study_surface
+from diff_diff.results_base import (
+    BaseResults,
+    _alpha_pct,
+    _coverage_pct,
+    _require_fit_alpha,
+    build_event_study_surface,
+)
 
 __all__ = [
     "ChaisemartinDHaultfoeuilleResults",
@@ -935,7 +941,7 @@ def summary(self, alpha: Optional[float] = None) -> str:
             decomposition diagnostic, and a footer of significance codes.
         """
         alpha = _require_fit_alpha(alpha, self.alpha)
-        conf_level = int((1 - alpha) * 100)
+        conf_level = _coverage_pct(alpha)
         width = 85
         sep = "=" * width
         thin = "-" * width
@@ -1590,7 +1596,7 @@ def _render_path_effects_section(
             if self.path_sup_t_bands is not None and path in self.path_sup_t_bands:
                 crit_p = self.path_sup_t_bands[path].get("crit_value", np.nan)
                 if np.isfinite(crit_p):
-                    conf_level = int((1 - self.alpha) * 100)
+                    conf_level = _coverage_pct(self.alpha)
                     lines.append(
                         f"  Sup-t critical value: {crit_p:.4f} "
                         f"(simultaneous {conf_level}% bands)"
@@ -1605,7 +1611,7 @@ def _render_honest_did_section(self, lines: List[str], width: int, thin: str) ->
         method_label = hd.method.replace("_", " ").title()
         m_val = hd.M
         sig_label = "Yes" if hd.is_significant else "No"
-        conf_pct = int((1 - hd.alpha) * 100)
+        conf_pct = _coverage_pct(hd.alpha)
         lines.extend(
             [
                 thin,
@@ -1625,7 +1631,7 @@ def _render_honest_did_section(self, lines: List[str], width: int, thin: str) ->
                 f"{'Identified set:':<35} " f"[{_fmt_float(hd.lb)}, {_fmt_float(hd.ub)}]",
                 f"{'Robust ' + str(conf_pct) + '% CI:':<35} "
                 f"[{_fmt_float(hd.ci_lb)}, {_fmt_float(hd.ci_ub)}]",
-                f"{'Significant at ' + str(int(hd.alpha * 100)) + '%:':<35} " f"{sig_label:>10}",
+                f"{'Significant at ' + _alpha_pct(hd.alpha) + '%:':<35} " f"{sig_label:>10}",
                 thin,
                 "",
             ]

diff_diff/changes_in_changes_results.py

@@ -7,7 +7,7 @@
 import pandas as pd
 
 from diff_diff._deprecation import deprecated_field_property
-from diff_diff.results_base import BaseResults
+from diff_diff.results_base import BaseResults, _coverage_pct
 
 _ESTIMATOR_TITLES = {
     "cic": "Changes-in-Changes (Athey & Imbens 2006) Results",
@@ -159,7 +159,7 @@ def summary(self) -> str:
         """Fixed-width text summary: headline ATT block plus the quantile-effects table."""
         from diff_diff.results import _get_significance_stars
 
-        ci_pct = int(round((1 - self.alpha) * 100))
+        ci_pct = _coverage_pct(self.alpha)
         width = 88
         bar = "=" * width
         dash = "-" * width

diff_diff/continuous_did_results.py

@@ -18,6 +18,7 @@
 from diff_diff.results_base import (
     _SUMMARY_ALPHA_MESSAGE,
     BaseResults,
+    _coverage_pct,
     _require_fit_alpha,
     build_event_study_surface,
 )
@@ -274,7 +275,7 @@ def summary(self, alpha: Optional[float] = None) -> str:
             or relabeling. Re-fit at the desired alpha instead.
         """
         alpha = _require_fit_alpha(alpha, self.alpha, message=_SUMMARY_ALPHA_MESSAGE)
-        conf_level = int((1 - alpha) * 100)
+        conf_level = _coverage_pct(alpha)
         w = 85
 
         lines = [

diff_diff/diagnostic_report.py

@@ -49,7 +49,7 @@
 import pandas as pd
 
 from diff_diff._reporting_helpers import describe_target_parameter  # noqa: E402 (top-level import)
-from diff_diff.results_base import Diagnostic
+from diff_diff.results_base import Diagnostic, _coverage_pct
 
 DIAGNOSTIC_REPORT_SCHEMA_VERSION = "2.0"
 
@@ -4461,9 +4461,9 @@ def _render_overall_interpretation(schema: Dict[str, Any], labels: Dict[str, str
         # stays consistent with the rendered interval when alpha != 0.05.
         headline_alpha = headline.get("alpha") if isinstance(headline, dict) else None
         if isinstance(headline_alpha, (int, float)) and 0 < headline_alpha < 1:
-            ci_level = int(round((1.0 - headline_alpha) * 100))
+            ci_level = _coverage_pct(headline_alpha)
         else:
-            ci_level = 95
+            ci_level = "95"
         ci_finite = (
             isinstance(ci, (list, tuple))
             and len(ci) == 2

diff_diff/diagnostics.py

@@ -20,7 +20,7 @@
 from diff_diff._deprecation import NOT_SUPPLIED, require_arg, resolve_renamed_kwarg
 from diff_diff.estimators import DifferenceInDifferences
 from diff_diff.results import _get_significance_stars
-from diff_diff.results_base import Diagnostic
+from diff_diff.results_base import Diagnostic, _coverage_pct
 from diff_diff.utils import safe_inference, validate_binary
 
 
@@ -87,7 +87,7 @@ def significance_stars(self) -> str:
 
     def summary(self) -> str:
         """Generate formatted summary of placebo test results."""
-        conf_level = int((1 - self.alpha) * 100)
+        conf_level = _coverage_pct(self.alpha)
 
         lines = [
             "=" * 65,

diff_diff/efficient_did_results.py

@@ -21,7 +21,12 @@
 from diff_diff.efficient_did_aggregation import _EfficientAggregationMixin
 from diff_diff.efficient_did_bootstrap import EfficientDiDBootstrapMixin
 from diff_diff.results import _format_survey_block, _get_significance_stars
-from diff_diff.results_base import BaseResults, _require_fit_alpha, build_event_study_surface
+from diff_diff.results_base import (
+    BaseResults,
+    _coverage_pct,
+    _require_fit_alpha,
+    build_event_study_surface,
+)
 
 if TYPE_CHECKING:
     from diff_diff.efficient_did_bootstrap import EDiDBootstrapResults
@@ -647,7 +652,7 @@ def summary(self, alpha: Optional[float] = None) -> str:
         never recomputed or relabeled - re-fit at the desired alpha).
         """
         alpha = _require_fit_alpha(alpha, self.alpha)
-        conf_level = int((1 - alpha) * 100)
+        conf_level = _coverage_pct(alpha)
 
         lines = [
             "=" * 85,

diff_diff/had.py

@@ -90,7 +90,7 @@
     BiasCorrectedFit,
     bias_corrected_local_linear,
 )
-from diff_diff.results_base import BaseResults, build_event_study_surface
+from diff_diff.results_base import BaseResults, _coverage_pct, build_event_study_surface
 from diff_diff.survey import (
     SurveyMetadata,
     compute_survey_metadata,
@@ -444,7 +444,7 @@ def __repr__(self) -> str:
     def summary(self) -> str:
         """Formatted summary table."""
         width = 72
-        conf_level = int((1 - self.alpha) * 100)
+        conf_level = _coverage_pct(self.alpha)
         lines = [
             "=" * width,
             "HeterogeneousAdoptionDiD Estimation Results".center(width),
@@ -884,7 +884,7 @@ def __repr__(self) -> str:
     def summary(self) -> str:
         """Formatted per-horizon summary table."""
         width = 80
-        conf_level = int((1 - self.alpha) * 100)
+        conf_level = _coverage_pct(self.alpha)
         lines = [
             "=" * width,
             "HeterogeneousAdoptionDiD Event-Study Results".center(width),