PyMetaAnalysis 0.5.0__tar.gz → 0.6.0__tar.gz

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/.gitignore

@@ -1,5 +1,6 @@
 # Local design notes are intentionally not part of the published project.
 /DESIGN.md
+/articles/
 
 # Python environments and generated files
 /.venv/

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/CHANGELOG.md

@@ -6,6 +6,23 @@ Changes planned for the next release accumulate under `Unreleased`.
 
 ## Unreleased
 
+## 0.6.0 - 2026-08-13
+
+### Added
+
+- common-effect Peto one-step odds-ratio pooling, including Peto-specific
+  study estimates, O-minus-E heterogeneity, explicit approximation warnings,
+  provenance/report metadata, and independent R `metafor` references;
+- common-effect Mantel-Haenszel risk-difference pooling with the
+  Sato-Greenland-Robins sampling variance, explicit method metadata,
+  sparse-table policy integration, and independent R `metafor` references.
+
+### Fixed
+
+- isolated builds temporarily cap Hatchling below 1.32 so the release
+  workflow continues to produce Core Metadata 2.4 accepted by Twine 6.2;
+  the cap can be removed once Twine validates Metadata 2.5.
+
 ## 0.5.0 - 2026-07-25
 
 ### Added

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/CITATION.cff

@@ -8,8 +8,8 @@ authors:
   - family-names: Ding
     given-names: Zhaobo
     email: ding.zb@yahoo.com
-version: 0.5.0
-date-released: 2026-07-25
+version: 0.6.0
+date-released: 2026-08-13
 repository-code: https://github.com/ZhaoboDing/PyMetaAnalysis
 url: https://zhaoboding.github.io/PyMetaAnalysis/
 license: MIT

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: PyMetaAnalysis
-Version: 0.5.0
+Version: 0.6.0
 Summary: A pandas-first, auditable meta-analysis library for Python
 Project-URL: Documentation, https://zhaoboding.github.io/PyMetaAnalysis/
 Project-URL: Source, https://github.com/ZhaoboDing/PyMetaAnalysis
@@ -133,7 +133,7 @@ not individual-level or causal effects.
 | Input | Effects | Pooling/models |
 | --- | --- | --- |
 | Effect + sampling variance or standard error | Generic | Common/random inverse variance |
-| Two-group events + totals | OR, RR, RD | Common MH OR/RR; common/random IV |
+| Two-group events + totals | OR, RR, RD | Common MH; common Peto OR; common/random IV |
 | Two-group means + SDs + sizes | MD, Hedges' g | Common/random inverse variance |
 | Effect + variance/SE + moderators | Generic | Common/mixed Meta-regression |
 
@@ -153,6 +153,12 @@ Sparse binary behavior is explicit: study-level and Mantel-Haenszel continuity
 corrections are separate, relative-effect double-zero/double-all rows remain
 visible as exclusions, and RD exposes
 `rd_zero_variance="correct" | "exclude"`.
+Peto OR uses raw tables for pooling and a separate one-step study estimator;
+it always reports a caution that its approximation is intended for rare
+outcomes, similar within-study arm sizes, and effects that are not large.
+Common-effect MH risk differences use the Sato-Greenland-Robins sampling
+variance; random-effects binary analyses continue to use inverse-variance
+pooling.
 
 ## Inspect and report
 
@@ -252,7 +258,7 @@ python -m mkdocs serve
 
 The test suite combines hand calculations, statistical invariants, numerical
 edge cases, and committed R `metafor` reference fixtures. CI covers Python
-3.10–3.13, declared dependency lower bounds, strict typing/linting, docs, and
+3.10–3.14, declared dependency lower bounds, strict typing/linting, docs, and
 distribution builds.
 
 This is independent cross-software validation, not a formal external

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/README.md

@@ -83,7 +83,7 @@ not individual-level or causal effects.
 | Input | Effects | Pooling/models |
 | --- | --- | --- |
 | Effect + sampling variance or standard error | Generic | Common/random inverse variance |
-| Two-group events + totals | OR, RR, RD | Common MH OR/RR; common/random IV |
+| Two-group events + totals | OR, RR, RD | Common MH; common Peto OR; common/random IV |
 | Two-group means + SDs + sizes | MD, Hedges' g | Common/random inverse variance |
 | Effect + variance/SE + moderators | Generic | Common/mixed Meta-regression |
 
@@ -103,6 +103,12 @@ Sparse binary behavior is explicit: study-level and Mantel-Haenszel continuity
 corrections are separate, relative-effect double-zero/double-all rows remain
 visible as exclusions, and RD exposes
 `rd_zero_variance="correct" | "exclude"`.
+Peto OR uses raw tables for pooling and a separate one-step study estimator;
+it always reports a caution that its approximation is intended for rare
+outcomes, similar within-study arm sizes, and effects that are not large.
+Common-effect MH risk differences use the Sato-Greenland-Robins sampling
+variance; random-effects binary analyses continue to use inverse-variance
+pooling.
 
 ## Inspect and report
 
@@ -202,7 +208,7 @@ python -m mkdocs serve
 
 The test suite combines hand calculations, statistical invariants, numerical
 edge cases, and committed R `metafor` reference fixtures. CI covers Python
-3.10–3.13, declared dependency lower bounds, strict typing/linting, docs, and
+3.10–3.14, declared dependency lower bounds, strict typing/linting, docs, and
 distribution builds.
 
 This is independent cross-software validation, not a formal external

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/adr/0002-statistical-policy.md

@@ -4,6 +4,10 @@
 - Date: 2026-07-15
 - Amendment: the prediction-interval variance decision is superseded by
   [ADR 0004](0004-hartung-knapp-prediction-intervals.md).
+- Amendment: the decision to defer Mantel-Haenszel risk differences is
+  superseded by [ADR 0005](0005-mantel-haenszel-risk-difference.md).
+- Amendment: the decision to defer Peto pooling is superseded by
+  [ADR 0006](0006-peto-odds-ratio.md).
 
 ## Context
 
@@ -44,13 +48,15 @@ option, and a provenance transformation.
 Mantel-Haenszel pooling remains a common-effect estimator for OR and RR. It
 uses raw tables by default and has a correction setting separate from the one
 used for individual-study effects. Random-effects Mantel-Haenszel, RD
-Mantel-Haenszel, and Peto pooling are outside the current scope.
+Mantel-Haenszel, and Peto pooling were outside the scope of this original
+decision; ADRs 0005 and 0006 supersede the latter two deferrals.
 
 ### Heterogeneity
 
 Cochran's Q, degrees of freedom, and p-value always use common-effect inverse-
 variance weights. Common-effect and Mantel-Haenszel analyses use Q-based
-I-squared and H-squared.
+I-squared and H-squared. ADR 0006 adds Peto's estimator-specific Q while
+retaining those Q-based inconsistency transformations.
 
 Random-effects analyses use:
 

pymetaanalysis-0.6.0/docs/adr/0005-mantel-haenszel-risk-difference.md

@@ -0,0 +1,65 @@
+# ADR 0005: Mantel-Haenszel risk difference
+
+- Status: Accepted
+- Date: 2026-08-13
+- Supersedes: the RD exclusion from the Mantel-Haenszel scope in
+  [ADR 0002](0002-statistical-policy.md)
+
+## Context
+
+ADR 0002 limited Mantel-Haenszel pooling to common-effect odds ratios and risk
+ratios while the risk-difference estimator and its variance convention were
+still undecided. This left inverse variance as the only pooling method for RD,
+although conventional R implementations provide a common-effect MH RD.
+
+The point estimator is straightforward, but several variance estimators have
+appeared in the literature. The selected rule must work under both large-
+stratum and sparse-data limiting models, preserve treatment/control symmetry,
+and remain explicit in reports and cross-software validation.
+
+## Decision
+
+PyMetaAnalysis supports `measure="RD", method="MH", model="common"`. With
+treatment total `n1_i`, control total `n0_i`, total `N_i`, and
+`w_i = n1_i n0_i / N_i`, the estimate is:
+
+```text
+RD_i = a_i / n1_i - c_i / n0_i
+RD_MH = sum(w_i RD_i) / sum(w_i)
+```
+
+The normal confidence interval uses the Sato-Greenland-Robins sampling
+variance. The resolved method options record
+`mh_rd_variance="Sato-Greenland-Robins"`.
+
+Raw tables are used for MH pooling by default. `mh_continuity_correction` and
+`mh_correction_scope` remain the only settings that alter MH pooling tables;
+the separate study-effect correction continues to control displayed study
+uncertainty and the inverse-variance heterogeneity calculation. The existing
+`rd_zero_variance` policy determines whether boundary studies enter all
+synthesis calculations. A non-positive pooled Sato variance raises a domain
+error instead of silently adding a correction.
+
+Random-effects MH remains unsupported. Random-effects RD continues to use
+inverse-variance pooling with an explicit tau-squared estimator.
+
+## Validation
+
+- direct formula tests cover the estimate, Sato variance, weights, and normal
+  interval;
+- treatment/control swaps negate the estimate and mirror its interval without
+  changing the standard error;
+- row reordering and common count scaling preserve the expected invariants;
+- extreme finite counts exercise the overflow-safe scaled implementation;
+- boundary-policy, explicit-correction, subgroup, leave-one-out, and
+  cumulative paths are covered; and
+- fixed-version `metafor::rma.mh(measure="RD")` fixtures cover ordinary,
+  sparse, and explicitly corrected tables.
+
+## Consequences
+
+- common-effect OR, RR, and RD all support MH or inverse-variance pooling;
+- study-table MH weights for RD are proportional to `n1_i n0_i / N_i`;
+- the selected RD variance convention is recoverable from method metadata and
+  generated reports; and
+- documentation must continue to distinguish MH RD from random-effects IV RD.

pymetaanalysis-0.6.0/docs/adr/0006-peto-odds-ratio.md

@@ -0,0 +1,87 @@
+# ADR 0006: Peto one-step odds ratio
+
+- Status: Accepted
+- Date: 2026-08-13
+- Supersedes: the Peto deferral in
+  [ADR 0002](0002-statistical-policy.md)
+
+## Context
+
+Peto's one-step method is a conventional common-effect estimator for binary
+outcomes and is particularly associated with rare-event meta-analysis. It is
+not interchangeable with an ordinary inverse-variance odds ratio: it derives
+both the study contribution and pooled estimate from observed-minus-expected
+events and hypergeometric information.
+
+The approximation can be biased when treatment and control group sizes differ
+substantially within studies, effects are large, or events are not rare. Its
+zero-cell behavior also differs from ordinary log odds ratios, so pooling and
+display corrections must not be conflated.
+
+## Decision
+
+PyMetaAnalysis supports `measure="OR", method="Peto", model="common"` with
+`ci_method="normal"`. `"peto_one_step"` is an accepted alias and the resolved
+pooling method is `"peto"`.
+
+For stratum `i`, let `O_i = a_i`, `m_i = a_i + c_i`, treatment and control
+totals be `n1_i` and `n0_i`, and `N_i = n1_i + n0_i`. Define:
+
+```text
+E_i = m_i n1_i / N_i
+V_i = m_i (N_i - m_i) n1_i n0_i / (N_i^2 (N_i - 1))
+```
+
+The individual and pooled model-scale estimates are:
+
+```text
+y_i = (O_i - E_i) / V_i
+Var(y_i) = 1 / V_i
+y_Peto = sum(O_i - E_i) / sum(V_i)
+Var(y_Peto) = 1 / sum(V_i)
+```
+
+Pooling always uses raw 2-by-2 tables. The existing
+`continuity_correction` and `correction_scope` settings affect only displayed
+study estimates and variances; there is no Peto pooling correction parameter.
+Double-zero and double-all rows are excluded before every synthesis
+calculation because they contain no relative-effect information.
+
+Peto heterogeneity uses the fitted pooled coefficient and the same
+observed-minus-expected contributions:
+
+```text
+Q = sum(((O_i - E_i) - y_Peto V_i)^2 / V_i)
+```
+
+Q-based I-squared and H-squared follow the project's common-effect
+conventions. Method options record `peto_pooling_tables="raw"` and
+`peto_heterogeneity="O-minus-E"`. Every Peto result carries an approximation
+warning naming the rare-outcome, balanced-arm, and modest-effect conditions.
+
+Random-effects Peto and Peto RR/RD are unsupported. Users requesting those
+estimands must select an implemented inverse-variance or Mantel-Haenszel
+combination explicitly.
+
+## Validation
+
+- direct formula tests cover study estimates, pooled estimate, variance,
+  weights, confidence interval, and Peto Q;
+- treatment/control swapping reverses and exponentiates the log-OR limits as
+  expected, while row order leaves the fit unchanged;
+- count scaling and extreme finite-count tests exercise overflow-safe
+  arithmetic;
+- sparse tables verify that study-level correction does not alter raw Peto
+  pooling and that double-zero/double-all rows are excluded; and
+- fixed-version `metafor::escalc(measure="PETO")` and `metafor::rma.peto()`
+  fixtures cover ordinary and sparse datasets.
+
+## Consequences
+
+- common-effect binary OR now offers MH, Peto, and inverse-variance pooling;
+- Peto results remain on the log-OR model scale and use exponentiated display
+  values like other OR results;
+- sensitivity, subgroup, provenance, reporting, and plotting workflows reuse
+  the same public result contracts; and
+- documentation and reports must preserve the Peto applicability warning
+  rather than presenting it as a general sparse-data default.

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/development.md

@@ -11,7 +11,7 @@ cd PyMetaAnalysis
 python -m pip install -e ".[test,dev,docs,plot]"
 ```
 
-Use a supported Python version (3.10–3.13). Keep changes focused and preserve
+Use a supported Python version (3.10–3.14). Keep changes focused and preserve
 unrelated worktree modifications.
 
 ## Run checks
@@ -31,7 +31,7 @@ python -m build
 python tools/inspect_distribution.py dist
 ```
 
-The CI matrix also tests Python 3.10–3.13 and declared dependency lower bounds.
+The CI matrix also tests Python 3.10–3.14 and declared dependency lower bounds.
 Install the `notebook` extra before running the notebook executor.
 
 ## Statistical changes

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/binary-outcomes.md

@@ -35,8 +35,33 @@ result = ma.meta_binary(
 print(result.summary())
 ```
 
-Mantel-Haenszel pooling currently supports OR and RR with `model="common"` and
-`ci_method="normal"`.
+Mantel-Haenszel pooling supports OR, RR, and RD with `model="common"` and
+`ci_method="normal"`. MH risk differences use the Sato-Greenland-Robins
+sampling variance, recorded in `dict(result.method.options)`.
+
+## Peto common-effect odds ratio
+
+Peto's one-step estimator is an explicit alternative for common-effect OR:
+
+```python
+peto = ma.meta_binary(
+    studies,
+    event_treat="events_t",
+    n_treat="total_t",
+    event_control="events_c",
+    n_control="total_c",
+    measure="OR",
+    method="Peto",
+    model="common",
+)
+```
+
+Peto pooling uses the raw 2-by-2 tables and its own observed-minus-expected
+heterogeneity statistic. The study table contains Peto one-step effects;
+`continuity_correction` affects those displayed study effects but not the
+pooled result. Peto is intended for rare outcomes when treatment/control arm
+sizes are similar within studies and effects are not large. The result always
+retains this caveat in `warnings` and Methods text.
 
 ## Random-effects analysis
 
@@ -71,7 +96,8 @@ them to the generic inverse-variance model.
 For OR and RR, `result.estimate` and `result.ci` stay on the log model scale.
 Use `display_estimate` and `display_ci` for exponentiated ratios.
 
-RD is available through inverse-variance pooling:
+For a common-effect RD, choose either MH or inverse-variance pooling explicitly.
+The MH form is:
 
 ```python
 result = ma.meta_binary(
@@ -81,7 +107,7 @@ result = ma.meta_binary(
     event_control="events_c",
     n_control="total_c",
     measure="RD",
-    method="IV",
+    method="MH",
     model="common",
     rd_zero_variance="correct",
 )
@@ -91,6 +117,8 @@ result = ma.meta_binary(
 their raw RD and uses corrected counts only for sampling variance. Use
 `"exclude"` for a protocol that excludes these studies before all synthesis
 calculations. See [zero-event studies](zero-events.md) for details.
+Use `method="IV"` for random-effects RD or when inverse-variance common-effect
+pooling is the prespecified estimator.
 
 ## Input validation
 
@@ -104,5 +132,5 @@ Sparse tables require additional decisions. Read
 settings.
 
 See [statistical methods](../methods/statistical-methods.md#binary-study-effects)
-for the OR/RR/RD and Mantel-Haenszel equations, and
+for the OR/RR/RD, Mantel-Haenszel, and Peto equations, and
 [validation](../validation.md) for cross-software coverage.

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/method-selection.md

@@ -36,15 +36,21 @@ protocols.
 | Data and model | Available pooling method |
 | --- | --- |
 | Generic effects, common or random | Inverse variance |
-| Binary OR/RR, common effect | Mantel-Haenszel or inverse variance |
-| Binary OR/RR, random effects | Inverse variance |
-| Binary RD, common or random | Inverse variance |
+| Binary OR, common effect | Mantel-Haenszel, Peto, or inverse variance |
+| Binary RR/RD, common effect | Mantel-Haenszel or inverse variance |
+| Binary OR/RR/RD, random effects | Inverse variance |
 | Continuous MD/SMD, common or random | Inverse variance |
 
 Mantel-Haenszel and inverse variance are different estimators, not aliases.
 PyMetaAnalysis does not extrapolate its common-effect Mantel-Haenszel weights
-into an undocumented random-effects procedure. Mantel-Haenszel OR/RR pooling
-uses raw tables by default and is not offered for RD.
+into an undocumented random-effects procedure. Mantel-Haenszel OR/RR/RD
+pooling uses raw tables by default. MH RD uses the Sato-Greenland-Robins
+sampling variance; it is not a random-effects estimator.
+
+Peto is a distinct common-effect OR estimator based on observed-minus-expected
+events. Consider it only when outcomes are rare, treatment/control group sizes
+are similar within each study, and effects are not large. It always uses raw
+tables for pooling and is not an automatic sparse-data default.
 
 ## Estimating tau-squared
 
@@ -63,9 +69,9 @@ Failure to converge raises `ConvergenceError`; it does not silently fall back
 to DL.
 
 The public default is `tau2_method=None`: it resolves to REML for a random- or
-mixed-effects fit. Common-effect and Mantel-Haenszel fits reject an explicitly
-supplied tau-squared method so analysis code cannot appear to request an
-estimator that was not used.
+mixed-effects fit. Common-effect, Mantel-Haenszel, and Peto fits reject an
+explicitly supplied tau-squared method so analysis code cannot appear to
+request an estimator that was not used.
 
 ## Confidence intervals
 
@@ -89,9 +95,10 @@ cautiously: choosing one does not remove the small-sample uncertainty.
 
 ## Heterogeneity definitions
 
-Cochran's Q, its degrees of freedom, and its p-value always use common-effect
-inverse-variance weights. Common-effect and Mantel-Haenszel analyses derive
-I-squared and H-squared from Q.
+Cochran's Q for IV and Mantel-Haenszel fits uses common-effect
+inverse-variance weights; MH residuals are centered on its pooled estimate.
+Peto instead uses its documented observed-minus-expected Q. All common-effect
+fits derive I-squared and H-squared from the applicable Q.
 
 For random-effects analyses, I-squared and H-squared instead use the estimated
 tau-squared and a typical within-study variance:

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/provenance-reporting.md

@@ -50,6 +50,16 @@ Binary analyses record:
 - non-informative relative-effect exclusions;
 - a separate Mantel-Haenszel correction record when MH pooling is used.
 
+A Peto OR transformation additionally records
+`study_estimator="peto_one_step"`. Its method options state that pooling used
+raw tables and heterogeneity used O-minus-E contributions. The study-level
+continuity-correction record remains separate because it does not alter Peto
+pooling.
+
+An MH risk-difference result additionally records
+`mh_rd_variance="Sato-Greenland-Robins"` in its method options, and generated
+Methods text names that variance convention.
+
 This separation matters because exact Mantel-Haenszel pooling can use raw
 tables even when a corrected individual effect is needed for display and
 heterogeneity calculations.
@@ -73,6 +83,8 @@ The text identifies the model, pooling estimator, tau-squared estimator,
 confidence interval, heterogeneity statistics, prediction interval,
 correction settings, missing-data policy, numerical controls, and package
 version as applicable.
+For Peto it also names the O-minus-E heterogeneity calculation, raw-table
+pooling, and the estimator's applicability caveat.
 
 It is a starting point for a manuscript Methods section. Review and adapt it to
 the analysis protocol, field conventions, and journal requirements.

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/r-interoperability.md

@@ -16,6 +16,7 @@ fixtures used by this project.
 | Generic effects and variances | `meta_analysis()` | `rma.uni(yi, vi, ...)` | `metagen(TE, seTE, ...)` |
 | Binary 2x2 tables, inverse variance | `meta_binary(..., method="IV")` | `escalc()` then `rma.uni()` | `metabin(..., method="Inverse")` |
 | Binary 2x2 tables, Mantel-Haenszel | `meta_binary(..., method="MH")` | `rma.mh()` | `metabin(..., method="MH")` |
+| Binary 2x2 tables, Peto OR | `meta_binary(..., measure="OR", method="Peto")` | `rma.peto()` | `metabin(..., sm="OR", method="Peto")` |
 | Continuous group summaries | `meta_continuous()` | `escalc()` then `rma.uni()` | `metacont()` |
 | Subgroups | `subgroup=` on a high-level call | separate fits or a moderator model | `subgroup=` |
 | Leave-one-out | `result.leave_one_out()` | `leave1out()` for supported fits | `metainf()` |
@@ -68,7 +69,8 @@ remain explicit attributes in Python.
 | `model="common"` | `rma.uni(..., method="EE")` | `common=TRUE, random=FALSE` | Inverse-variance common-effect fit |
 | `model="random"` | random-effects `rma.uni()` | `random=TRUE` | Requires a tau-squared policy |
 | `method="IV"` | inverse-variance weighting | `method="Inverse"` | Binary API only; generic and continuous fits are IV |
-| `method="MH"` | `rma.mh()` | `method="MH"` | Common-effect OR/RR only |
+| `method="MH"` | `rma.mh()` | `method="MH"` | Common-effect OR/RR/RD |
+| `method="Peto"` | `rma.peto()` | `method="Peto"` | Common-effect OR; raw pooling tables and O-minus-E heterogeneity |
 | `tau2_method=None` (resolved as `"REML"`) | `method="REML"` | `method.tau="REML"` | PyMetaAnalysis random-effects default |
 | `tau2_method="PM"` | `method="PM"` | `method.tau="PM"` | Paule-Mandel |
 | `tau2_method="DL"` | `method="DL"` | `method.tau="DL"` | DerSimonian-Laird |
@@ -78,6 +80,9 @@ handling, or heterogeneity definitions. In particular, PyMetaAnalysis records
 `i2_method`; random-effects inverse-variance results use the documented
 tau-squared/typical-variance definition, while common-effect and MH results use
 the Q-based definition.
+Peto also reports Q-based inconsistency, but its Q is formed from the Peto
+observed-minus-expected contributions rather than ordinary study-log-OR
+inverse-variance residuals.
 
 `result.tau2_confidence_interval()` corresponds to
 `confint(fit, type="QP")` for an `rma.uni` random-effects fit. Both return
@@ -128,6 +133,19 @@ not implement. For OR/RR, double-zero and double-all studies are excluded from
 all model and heterogeneity calculations by default while remaining visible in
 the result table. RD uses its separate `rd_zero_variance` policy.
 
+For `measure="RD"`, PyMetaAnalysis and `metafor::rma.mh(measure="RD")` use
+the Sato-Greenland-Robins sampling variance. A separate explicit
+`mh_continuity_correction` changes the MH tables; the study-effect
+`continuity_correction` still affects only displayed study variances and
+heterogeneity.
+
+For Peto, PyMetaAnalysis's default study-level correction corresponds to
+`escalc(measure="PETO", add=0.5, to="only0", drop00=TRUE)` for the displayed
+study rows. Its pooled fit corresponds to `rma.peto()` with raw pooling tables;
+the study correction never changes the pooled result. Compare Peto's
+rare-outcome, within-study arm-balance, and modest-effect assumptions before
+porting it solely because a table contains zeros.
+
 Read [zero-event studies](zero-events.md) before translating sparse analyses.
 
 ## Worked generic translation
@@ -194,6 +212,7 @@ as numerically equivalent.
 - [`metafor::rma.uni`](https://wviechtb.github.io/metafor/reference/rma.uni.html)
 - [`metafor::predict.rma`](https://wviechtb.github.io/metafor/reference/predict.rma.html)
 - [`metafor::rma.mh`](https://wviechtb.github.io/metafor/reference/rma.mh.html)
+- [`metafor::rma.peto`](https://wviechtb.github.io/metafor/reference/rma.peto.html)
 - [`metafor::escalc`](https://wviechtb.github.io/metafor/reference/escalc.html)
 - [`meta::metagen`](https://search.r-project.org/CRAN/refmans/meta/html/metagen.html)
 - [`meta::metabin`](https://search.r-project.org/CRAN/refmans/meta/html/metabin.html)

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/sensitivity-analysis.md

@@ -160,7 +160,7 @@ model, and the boundary is recorded in `cumulative.warnings`.
 Every refit reuses the original result's:
 
 - common- or random-effects model;
-- inverse-variance or Mantel-Haenszel pooling method;
+- inverse-variance, Mantel-Haenszel, or Peto pooling method;
 - tau-squared and confidence-interval methods;
 - confidence level, absolute tolerance, and iteration limit;
 - missing-data policy;

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/guides/zero-events.md

@@ -1,8 +1,8 @@
 # Zero-event studies
 
-Zero cells in 2-by-2 tables affect OR, RR, RD, and Mantel-Haenszel estimators in
-different ways. PyMetaAnalysis therefore separates study-level effect
-corrections from the pooled Mantel-Haenszel correction.
+Zero cells in 2-by-2 tables affect OR, RR, RD, Mantel-Haenszel, and Peto
+estimators in different ways. PyMetaAnalysis therefore separates study-level
+effect corrections from table-based pooling rules.
 
 ## Default study-level correction
 
@@ -71,7 +71,8 @@ RD-specific; setting `exclude` for OR or RR is rejected.
 
 ## Mantel-Haenszel correction is separate
 
-Exact common-effect Mantel-Haenszel OR/RR pooling uses raw tables by default:
+Uncorrected common-effect Mantel-Haenszel OR/RR/RD pooling uses raw tables by
+default:
 
 ```python
 mh_continuity_correction = None
@@ -83,6 +84,29 @@ and heterogeneity. It does not silently alter the pooled Mantel-Haenszel
 estimator. If an exact pooled estimator is undefined, choose an explicit
 positive `mh_continuity_correction` and report that decision.
 
+For MH RD, `rd_zero_variance="exclude"` removes zero-variance boundary rows
+before every synthesis calculation. With the default `"correct"` policy, the
+raw table still enters the MH point estimate unless an explicit
+`mh_continuity_correction` is supplied. If the Sato-Greenland-Robins variance
+is non-positive, the uncorrected fit raises instead of silently changing the
+tables; a positive MH correction is an explicit protocol choice.
+
+## Peto pooling always uses raw tables
+
+Peto's observed-minus-expected pooling contribution remains defined for many
+single-zero tables, so `method="Peto"` does not apply a continuity correction
+to pooling or Peto Q. The general `continuity_correction` and
+`correction_scope` settings apply only to the one-step study effects and
+variances shown in the study table. There is deliberately no separate Peto
+pooling-correction option.
+
+Double-zero and double-all studies have zero Peto information and are excluded
+before pooling, Q, I-squared, H-squared, and weights. The result retains the
+same structured exclusion reasons used for other relative-effect analyses.
+Peto's lack of a single-zero pooling correction does not make it universally
+preferable: its rare-outcome, balanced-arm, and modest-effect assumptions must
+still be considered.
+
 ## Inspect what happened
 
 ```python
@@ -104,3 +128,5 @@ Resolved correction values and scopes also appear in
 zero-variance policy and affected row IDs in provenance. This makes it possible
 to distinguish a corrected analysis from an exact or exclusion-based one after
 fitting.
+Peto analyses additionally record `peto_pooling_tables="raw"` and
+`peto_heterogeneity="O-minus-E"`.

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/index.md

@@ -84,11 +84,13 @@ lists unsupported methods explicitly.
 
 ## Project status
 
-PyMetaAnalysis 0.4.0 extends Meta-regression with exact leave-one-out and
-influence diagnostics, VIF/GVIF and weighted condition diagnostics, explicit
-linear contrasts, and opt-in Riley prediction intervals. The project has not
-undergone a formal external statistical audit. Pin the package version for
-consequential work and independently check important analyses. See the
+PyMetaAnalysis 0.6.0 adds common-effect Mantel-Haenszel risk-difference and
+Peto one-step odds-ratio pooling, with explicit sparse-table policies,
+statistical metadata, and independent R references. It also retains the
+Q-profile heterogeneity intervals and Meta-regression diagnostics introduced
+in earlier releases. The project has not undergone a formal external
+statistical audit. Pin the package version for consequential work and
+independently check important analyses. See the
 repository
 [changelog](https://github.com/ZhaoboDing/PyMetaAnalysis/blob/main/CHANGELOG.md)
 and [contribution guide](development.md). For manuscripts and archived

{pymetaanalysis-0.5.0 → pymetaanalysis-0.6.0}/docs/installation.md

@@ -1,6 +1,6 @@
 # Installation
 
-PyMetaAnalysis supports Python 3.10 through 3.13. The numerical core depends
+PyMetaAnalysis supports Python 3.10 through 3.14. The numerical core depends
 on NumPy, pandas, and SciPy; Matplotlib is optional.
 
 ## Install from PyPI