Add Group Sequential Design and Bayes Factor (Sequential) Design by FBartos · Pull Request #87 · jasp-stats/jaspPower

FBartos · 2026-05-25T11:22:07Z

adds the following analyses:

Group Sequential Design (via the gsDesign R package)
Bayes Factor Design (via the bfpwr R package)
Bayes Factor Sequential Design (via the bfpwr R package)

Introduce a new Evidence analysis: export Evidence in NAMESPACE; add a stub R handler R/Evidence.R that creates an interface panel with a placeholder message (calculations to be implemented later); register the analysis in inst/Description.qml; and add the full UI definition inst/qml/Evidence.qml (QML form with parameters, priors, design options, plots, and advanced settings).

Replace placeholder Evidence interface with a full implementation: R/Evidence.R now prepares settings, computes evidence probabilities and sample-size searches, builds results/prior/design tables, explanatory HTML, and multiple plots (by N, effect size, and priors). Uses bfpwr for core Bayesian calculations and includes robust error handling and plotting via ggplot2/jaspGraphs. DESCRIPTION updated to import and reference the SamCH93/bfpwr package. UI/QML updated to add a Bayesian group and include a new powerBayes SVG icon (inst/icons/powerBayes.svg) for the Evidence analysis.

github-actions · 2026-05-25T11:22:17Z

👋 Friendly reminder: It looks like NEWS.md wasn't updated in this Pull Request.

If your changes include bug fixes, new features, or UI tweaks, please consider adding a quick note to the # jaspYourModule (development version) section so our users know about your awesome work! (If this is just a minor typo fix, feel free to ignore this message.)

FBartos · 2026-05-25T11:26:07Z

build builder bot, build for glory!

github-actions · 2026-05-25T11:31:42Z

✅ Build Complete!

Your JASP module bundles have been successfully built for macOS (ARM) and Windows.

👉 Click here to download the artifacts

(Scroll to the bottom of the page to find the "Artifacts" section).

…ntain permissions' Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>

vandenman · 2026-06-05T12:56:05Z

+Form
+{
+	info: qsTr("Bayes Factor sequential design allows you to design sequential experiments for conclusive evidence.\n\n" + 
+	"See [this tutorial](TODO) for a detailed introduction to the module.")


TODO is intended still?

lets comment this line out for now, and put it back in once the tutorial is preprinted

vandenman · 2026-06-05T12:56:18Z

+Form
+{
+	info: qsTr("Bayes Factor design allows you to design fixed sample size experiments for conclusive evidence.\n\n" + 
+	"See [this tutorial](TODO) for a detailed introduction to the module.")


TODO is intended still?

vandenman · 2026-06-16T14:26:30Z

+  if (settings[["isIndependentSamples"]] && settings[["sampleSizeRatio"]] != 1)
+    stop(gettext("Sample-size search for unequal group sizes is handled internally."))
+
+  if (settings[["drangeMode"]] == "custom")
+    stop(gettext("Sample-size search for custom t search ranges is handled internally."))
+
+  if (!isTRUE(all.equal(settings[["nullValue"]], 0)))
+    stop(gettext("Sample-size search for nonzero null values is handled internally."))


Are these errors meant for users? When/ how can they occur? If they occur only when an analyst has done something wrong, they shouldn't be translated so no gettext. Please add a comment describing what these errors are for.

these errors looks odd and should be never reached

vandenman · 2026-06-16T14:36:34Z

+.bfdReportTestLabel <- function(settings) {
+  .bfdRequireKnownOption(gettext("test"), settings[["test"]], .bfdKnownTests)
+
+  switch(settings[["test"]],


Same switch statement as .bfdTestLabel, avoid duplication.

vandenman · 2026-06-16T14:39:44Z

+.bfdPriorSpikeRow <- function(x, prior, height) {
+  return(data.frame(
+    x       = x,
+    height  = height,
+    prior   = prior,
+    stringsAsFactors = FALSE
+  ))
+}
+
+.bfdPriorSpikeRows <- function(rows) {
+  if (length(rows) == 0) {
+    return(data.frame(
+      x       = numeric(0),
+      height  = numeric(0),
+      prior   = character(0),
+      stringsAsFactors = FALSE
+    ))
+  }
+
+  return(do.call(rbind, rows))
+}


Suggested change

.bfdPriorSpikeRow <- function(x, prior, height) {

return(data.frame(

x = x,

height = height,

prior = prior,

stringsAsFactors = FALSE

))

}

.bfdPriorSpikeRows <- function(rows) {

if (length(rows) == 0) {

return(data.frame(

x = numeric(0),

height = numeric(0),

prior = character(0),

stringsAsFactors = FALSE

))

}

return(do.call(rbind, rows))

}

.bfdPriorSpikeRow <- function(x, prior, height) {

return(data.frame(

x = x,

height = height,

prior = prior,

stringsAsFactors = FALSE

))

}

.bfdPriorSpikeRows <- function(rows) {

if (length(rows) == 0) {

return(.bfdPriorSpikeRow(numeric(), numeric(), character()))

}

return(do.call(rbind, rows))

}

too much duplication again

vandenman · 2026-06-16T14:49:53Z

+    isZTest              = grepl("ZTest", test, fixed = TRUE) || identical(test, "generalZApproximation"),
+    isBinomial           = FALSE,
+    calculation          = options[["calculationTarget"]],
+    bf10Threshold        = options[["conclusiveEvidenceThresholdH1"]],


This is confusion and renames some options. Why not just add options as a whole to the settings? That way someone can understand which name maps onto which option in QML.

vandenman · 2026-06-16T14:51:27Z

+.bfdCreatePlot <- function(parent, key, title, position, dependencies, width, height) {
+  if (!is.null(parent[[key]]))
+    return(NULL)
+
+  plot <- createJaspPlot(title = title, width = width, height = height)
+  plot$dependOn(dependencies)
+  plot$position <- position
+  parent[[key]] <- plot
+
+  return(plot)
+}


This feels overengineered.

vandenman · 2026-06-16T14:52:08Z

+.bfdCachedPlotData <- function(jaspResults, stateKey, dependencies, dataKey, compute) {
+  state <- jaspResults[[stateKey]]
+  if (is.null(state)) {
+    state <- createJaspState()
+    state$dependOn(dependencies)
+    jaspResults[[stateKey]] <- state
+  }
+
+  cache <- state$object
+  if (is.null(cache))
+    cache <- list()
+
+  if (!is.null(cache[[dataKey]]))
+    return(cache[[dataKey]])
+
+  cache[[dataKey]] <- try(compute(), silent = TRUE)
+  state$object <- cache
+
+  return(cache[[dataKey]])
+}


This reimplements %setOrRetrieve%, no? See https://github.com/jasp-stats/jaspBase/blob/master/R/setOrRetrieve.R

vandenman · 2026-06-16T14:52:46Z

+  html <- .bfdCreateHtml(
+    parent       = parent,
+    key          = key,
+    title        = title,
+    position     = position,
+    dependencies = dependencies
+  )


bfdCreateHtml is likely also overengineered

fully agree

vandenman · 2026-06-16T14:53:17Z

+    testType               = testType,
+    designTypeLabel        = .csdDesignTypeLabel(options[["designType"]]),
+    numberOfLooks          = options[["numberOfLooks"]],
+    timingMode             = options[["timingMode"]],


same remark as before, just append options.

FBartos and others added 18 commits March 5, 2026 16:26

agent instructions

7658cb8

update

9a5d92c

feat: update bpower designs

c596a20

fix naming

7841c99

improve org

70b2d9f

update GUI

06ebf08

update instructions

72a666b

update module

d551675

update results

3f81daf

update results

bac87ea

update + tests

f389ac6

update notes

a5d9e85

update Bayes

3455dcd

add references

f7ead24

update frequentist GSD

8f51de4

Merge remote-tracking branch 'origin/bpower' into resolve/pr86-on-master

3febd2d

FBartos added 3 commits May 25, 2026 13:36

add news and readme

5dc0587

Create test-coverage.yml

77eb191

update deps

24c383b

github-advanced-security AI found potential problems May 25, 2026

View reviewed changes

Comment thread .github/workflows/test-coverage.yml Fixed

Potential fix for pull request finding 'CodeQL / Workflow does not co…

ad06170

…ntain permissions' Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>

FBartos requested review from juliuspfadt and removed request for juliuspfadt May 25, 2026 16:43

FBartos added 2 commits June 1, 2026 16:42

update calculations

ce63fa1

cleanup

9101d09

FBartos requested a review from vandenman June 1, 2026 14:56

vandenman requested changes Jun 16, 2026

View reviewed changes

Conversation

FBartos commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions Bot commented May 25, 2026

Uh oh!

FBartos commented May 25, 2026

Uh oh!

github-actions Bot commented May 25, 2026

✅ Build Complete!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

FBartos commented May 25, 2026 •

edited

Loading