Skip to content

Update PET-PEESE and WAAP-WLS - #485

Open
FBartos wants to merge 12 commits into
jasp-stats:masterfrom
FBartos:PET-PEESE
Open

FBartos wants to merge 12 commits into
jasp-stats:masterfrom
FBartos:PET-PEESE

Conversation

@FBartos

@FBartos FBartos commented Nov 2, 2026

Copy link
Copy Markdown
Collaborator

Changed

  • Rewrote PET-PEESE using metafor, including optional multilevel/multivariate models. Legacy analyses are converted; incompatible settings may reset on refresh.
  • Rewrote WAAP-WLS using metafor in the PET-PEESE style (moderators, subgroups, clustering, no-selection indicator, optional multilevel/multivariate models), with an adjustable power criterion evaluated at the pooled or moderator-specific predicted effects, and exports of the estimated power and WAAP inclusion. Legacy analyses are converted; incompatible settings may reset on refresh.
  • Selection Models, PET-PEESE, and WAAP-WLS forest plots can show unadjusted fixed and random effects estimates.
  • Unrestricted weighted least squares reports the multiplicative heterogeneity (γ and √γ).

Fixed

  • Clustered meta-analyses no longer offer z-tests; saved analyses are switched to t-tests.
  • Multilevel/multivariate meta-analysis fixes (profile plots, random effects confidence intervals, weighted estimation, models with multiple heterogeneity levels).
  • Precomputed variance-covariance matrices in semicolon-separated files (decimal commas) are read correctly.
  • Forest plot predicted effects are unavailable when the study information is aggregated.

FBartos and others added 7 commits November 2, 2026 09:16
PET-PEESE-style inputs (UWLS default, moderators, subgroups, clustering,
no-selection, multilevel/multivariate); power criterion field and power
conditional on moderators checkbox; WLS/WAAP estimates table; upgrade
0.98.0 -> 0.99.0. Drop legacy lm backend and correlation input helpers.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Standard classical output only; power from predict(fit)$pred / se in R code;
no decision comment or trailing predict.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- WAAP: predictions at moderator means of all estimates; V subset from full matrix
- PET-PEESE: no crash when subgroup drops all moderators
- drop NA-weight rows before bias/power terms
- forest: predicted effects off when aggregating
- power criterion max 99%, module and R code share power >= criterion rule
- R code: no PET-PEESE decision comment/predict; WAAP no-selection NA filter
- precomputed V: detect semicolon files (read.csv2) in module and R code
- guard legacy 0.17.3 renames for PET-PEESE / WAAP-WLS

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@FBartos
FBartos requested a review from juliuspfadt November 2, 2026 10:56
Ubuntu sysreqs query only repository packages; latest-packages job uses the
lockfile jaspTools commit (current jaspTools imports unresolvable jaspSyntax).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@codecov-commenter

codecov-commenter commented Nov 2, 2026 •

Copy link
Copy Markdown

Codecov Report

❌ Patch coverage is 90.10989% with 117 lines in your changes missing coverage. Please review.
✅ Project coverage is 59.42%. Comparing base (b6465ed) to head (56130f2).
⚠️ Report is 120 commits behind head on master.

Files with missing lines Patch % Lines
R/classicalmetaanalysis-bubble-plot.R 66.32% 33 Missing ⚠️
R/petpeese.R 91.47% 19 Missing ⚠️
R/waapwls.R 91.86% 10 Missing ⚠️
R/classicalmetaanalysis-model-summary-tables.R 73.07% 7 Missing ⚠️
R/classicalmetaanalysis-exports.R 73.91% 6 Missing ⚠️
R/forestplot-model-data.R 89.83% 6 Missing ⚠️
R/classicalmetaanalysis-diagnostic-plots.R 84.84% 5 Missing ⚠️
R/petpeese-tables.R 95.76% 5 Missing ⚠️
R/waapwls-r-code.R 93.33% 4 Missing ⚠️
R/waapwls-tables.R 89.47% 4 Missing ⚠️
... and 10 more
Additional details and impacted files
@@             Coverage Diff             @@
##           master     #485       +/-   ##
===========================================
+ Coverage   40.24%   59.42%   +19.17%     
===========================================
  Files          24       86       +62     
  Lines       13331    18538     +5207     
===========================================
+ Hits         5365    11016     +5651     
+ Misses       7966     7522      -444     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:
  • ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
  • 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

FBartos and others added 4 commits November 2, 2026 15:44
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- failed models reported without metafor call; PET failure names subgroup and number of estimates
- semicolon construct/distance/correlation matrix files read via shared reader
- failed variance-covariance matrix save reported
- plain-text profile likelihood axis labels
- footnote typos

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants