Why do the category mappings matter?
Multi-value QCA compares combinations of categories. The mapping from an observed label to a category must be explicit before the combinations are analysed. Numeric codes act as names, not as equally spaced measurements.
The independent reference works through the defined category combinations and their classifications. That allows it to check the possible solutions throughout a small example, rather than checking just one chosen solution.
Which versions belong to these checks?
The table identifies the QuickPLS development and installer versions associated with the saved evidence. The independent calculation and the later run of the QCA reference script are described separately below, so you can see which software each step actually checked.
On a small screen, swipe the table sideways to see every column.
| Record | QuickPLS build or result checked | Reference | Comparison outcome |
|---|---|---|---|
| Multi-value QCA evidence | 1.03.00 / Installer D1.03.02 | See which mathematical and app-workflow checks were performed | Mathematical reference plus checks of the specified calculations and app workflows |
Every classification in the defined category example
The independent reference checks all 729 classifications of a 3x2 universe of category combinations. The universe combines a three-category condition with a two-category condition. Within it, it works through the possible terms and solutions and compares the complete conservative and parsimonious alternatives with the QuickPLS calculation. Conservative solutions avoid simplifying with unobserved configurations; parsimonious solutions permit those simplifying assumptions more broadly.
A separate published single-set example gives consistency 1, PRI 1 and coverage .625. Consistency describes how closely the selected category is contained in the outcome. PRI, proportional reduction in inconsistency, checks for conflicting sufficiency relationships with the outcome and its absence. Coverage describes how much of the outcome is covered by the selected category. The perfect consistency in this example therefore does not mean that it covers the whole outcome.
- Before the 4 November reference attempt, the original review had not run the optional R script for QCA. The completed reference run is described below.
- The original review, before the 4 November reference attempt, used published DEV category counts. It did not invent individual multivariate observations to reconstruct a dataset.
| What was checked | Value / scope |
|---|---|
| Classifications | 729 |
What complete coverage means in this example
The exhaustive result is useful because it checks every classification in the defined category universe. Its scope is that universe. It does not establish that all larger models, datasets or category mappings have been checked in the same way.
The published DEV example answers a narrower question using category counts. Its set-fit measures and the logical enumeration are different checks. Keep them separate when reading the report, rather than treating the number of classifications as a count of independent studies.
- These checks include logical alternatives, numerical measures and workflow rules. There is no single permitted decimal error for the page. Read the expected result for each example.
What the reference run on 4 November 2026 returned
During website preparation, the optional reference script ran with R 4.6.1 and QCA 3.25. For the specified example, it returned the conservative solution A[2]*B[1] and the parsimonious alternatives A[2] or B[1]. Here the bracketed value identifies a category, and the asterisk joins conditions with AND.
A separate example using DEV[2] counts returned coverage .625. This later work confirms the reference calculation only. Neither QuickPLS nor the Store app was run again, so the reference run should not be described as a fresh app comparison.
How to inspect the multi-value QCA report
The JSON summary covers the small-universe check, the published DEV count example and the later reference run. It identifies the versions and source documents. Read the solution examples above alongside this summary.
Repeating those checks requires their original inputs, category mappings, assumptions, settings and software versions. The download documents completed work; the original datasets and app source code are not included.
Download the comparison results
See the values we checked, the differences we found and the software versions used. The download also identifies the original test files.
About these tests
These are QuickPLS development comparisons. Each record identifies the reference software or mathematical calculation, the tested build and the settings used.
How to understand these comparisons · Find checks for another method