Repository navigation
Add residual plots - #231
Open
JohnnyDoorn wants to merge 9 commits into
Open
Add residual plots #231JohnnyDoorn wants to merge 9 commits into
JohnnyDoorn wants to merge 9 commits into
Conversation
Adds "Residuals vs. predicted" and "Q-Q plot standardized residuals" to each model in the classical and Bayesian Process analysis, with a "Set for All Models" default under Advanced. The plots are shown for every dependent variable of a model (the outcome and each mediator), so users can check the homoscedasticity and normality assumptions of each regression equation without refitting it in Linear Regression. Predicted values are computed from the parameter table, because lavPredict(type = "yhat") returns the observed values for models without latent variables. Standardized residuals use the estimated residual variance; Bayesian models use the posterior means. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Picks up the Q-Q plot fixes: the confidence band follows the drawn y = x line (jasp-stats#167), and missing values are dropped before the quantiles are computed (jasp-stats#169). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Adds "Confidence interval" under "Q-Q plot standardized residuals" in the classical analysis, as in Linear Regression, ANOVA and the t-tests: a nested qqPlotCi checkbox with a qqPlotCiLevel field, passed as ciLevel to jaspGraphs::plotQQnorm, plus its "Set for All Models" default under Advanced. The Bayesian analysis hides the option, like the other Bayesian analyses, whose Q-Q plots have no band. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Shows the "Credible interval" option under "Q-Q plot standardized residuals" in the Bayesian analysis (the same qqPlotCi/qqPlotCiLevel options as the classical confidence band), plus its "Set for All Models" default. As in Bayesian ANOVA, the standardized residuals are computed for each posterior draw: the points are their posterior means and the error bars their central credible intervals. The points previously used the posterior means of the parameters; the two differ by about 0.02 on the standardized scale. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Bootstrapped standard errors / confidence intervals and bootstrapped local
tests were not reproducible. Adds SetSeed {} under Options and seeds each
model's bootstrap and local tests, so results do not depend on which models
were refitted in a run. seed/setSeed were already in .procGetDependencies().
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… changes bootstrapSamples was in no dependency list, so changing Replications kept the previous bootstrap. Adds it to .procGetDependencies(), next to errorCalculationMethod and bootstrapCiType. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
Author
|
About the failing All four jobs fail in "Set up the test environment" while restoring dependencies, before any test runs. This is not caused by this PR. Master's scheduled run fails at the same step (every week since at least 2026-08-31), and so do the other modules (jaspRegression, jaspSem, jaspAnova, ...). Fixes are underway in jasp-stats/jasp-actions#89, jasp-stats/jaspTools#88 and jasp-stats/jasp-actions#78. Local run of the full test suite on b101254 (macOS arm64, R 4.5.2, renv lockfile,
The two failures ("model number 10 - continuous" and "model number 10 - factor", 9 expectations) fail in the same way on master, so they are unrelated to this PR. They are MCMC reference values that differ on this machine, and that file is skipped on CI. The 228 warnings are also the same as on master. Once those CI fixes are in, jaspProcess will likely need a separate lockfile update: |
FBartos
approved these changes
Nov 6, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
predicted vs residuals and q-q plot
for Bayesian version (same style as BANOVA)
