NEWS.md

In ThSQCA: Threshold-Sweep QCA

ThSQCA 2.0.8

This release corrects compute_fiss_core(). Core/peripheral classifications obtained with earlier versions should be recomputed; see the first item.

Changes in results

• compute_fiss_core() now classifies core and peripheral conditions configuration by configuration, as in the solution tables of Fiss (2011): the core conditions of a term of the intermediate solution are the conditions of the parsimonious term(s) contained in it, and its other conditions are peripheral. Up to 2.0.7 a condition was core whenever it appeared, with the same status, in any term of the parsimonious solution. That rule does not reproduce Fiss's tables, where the same condition can be core in one configuration and peripheral in another. When the parsimonious solution is unique, the change can only turn core conditions into peripheral ones; in a simulation with random crisp-set data, about 3% of such data sets had at least one condition reclassified. The help page has a new section, "Relation to Fiss (2011)".

• compute_fiss_core() now follows how QCA::minimize() derived the intermediate solution. With dir.exp, QCA stores one entry in sol$i.sol for each pair of a complex and a parsimonious minimal solution (C1P1, C1P2, ...), holding the parsimonious model in $p.sol and the intermediate model(s) obtained from it in $solution. Up to 2.0.7 this record was not used, which caused two errors when the parsimonious solution had tied minimal solutions:

• The terms of all intermediate models were pooled, so the Fiss chart and print_fiss_summary() could show terms that are not part of the reported solution (M1). Only M1 is now classified, matching the sweep summary.
• M1 was compared with every tied parsimonious solution, including ones it was not obtained from. M1 is now compared only with its source solution(s).

Because M1 is no longer compared with tied solutions it was not obtained from, conditions can also change from peripheral to core in this case. When QCA derives the same M1 from several tied parsimonious solutions, a condition is core only if it is core relative to each of them (the polarity safeguard introduced in 2.0.5), and the warning names the source solutions. When a term of the intermediate solution contains no parsimonious term, its conditions are classified as peripheral and a warning is given.

Bug fixes

• compute_fiss_core() returned an empty result without any message for results of ctSweepM() and dtSweep(). It now stops with an error saying that only otSweep() and ctSweepS() results are supported.
• Condition names containing a dot (such as high.trust, as produced by read.csv()) were matched incorrectly in configuration charts and in compute_fiss_core(): the dot matched any character, so A.B was found in the term A*B, and a condition A was found in A.B. Names are now matched literally.
• For ctSweepS() results, Fiss charts, print_fiss_summary() and the Fiss charts in generate_report() labeled the swept condition threshold as thrY. They now show the swept condition (for example X3=6).

Output changes

• compute_fiss_core(): parsim_expression shows the parsimonious solution(s) M1 was compared with, labeled with QCA's names ("P1: ...; P2: ...") when there are several, instead of the terms of all tied solutions joined into one expression. New fields interm_n_solutions and parsim_sources record the number of intermediate minimal solutions and the i.sol entries M1 was obtained from. These fields are also present (as NA) for thresholds without a truth table or solution.

Documentation

• Descriptions of core conditions in generate_fiss_chart(), generate_report() and the package description now state the configuration-level rule.
• The Fiss section of the tutorial vignette explains the configuration-level rule and the handling of tied parsimonious solutions. The example output in section 12 of the reproducible-code vignette now matches the actual output. The README points to compute_fiss_core().

ThSQCA 2.0.7

Documentation

• citation("ThSQCA") now also lists the accompanying preprint (Toyoda, 2026, SocArXiv, ), which is also added to the README and to the reference list of the tutorial vignette.
• Help page titles now use the ThS-QCA labels ("OTS", "CTS (single)", "CTS (multiple)", "DTS") instead of "OTS-QCA", "CTS-QCA", "MCTS-QCA" and "DTS-QCA", and the package overview page describes results in terms of sufficiency structures and threshold transitions rather than causal structures.
• ThSQCA_Tutorial_EN was reorganized and extended: a quick start, data preparation (binary variables, pre_calibrated, sweeping membership scores), a section on choosing between complex, parsimonious and intermediate solutions and one on multiple minimal solutions (both with a second simulated data set that has logical remainders), guidance on reporting, and answers to common questions. Version-history headings were removed.
• Unified terminology with the companion methodology paper: the framework is now called Threshold-Sweep QCA (ThS-QCA), and the sweeps are referred to as CTS (ctSweepS() for one condition, ctSweepM() for several conditions), OTS (otSweep()), and DTS (dtSweep()). The former labels "TS-QCA" and "MCTS-QCA" were removed from the README and the vignettes. No function names, arguments, or results changed.
• README: added links to the two browser-based helpers (Sweep Builder and QCA Quickstart) and to the tutorial vignettes, added install.packages("ThSQCA") as the primary installation route, and made the "Basic Setup" example use the bundled sample_data so that it runs as written. Fixed two bullet lists that were rendered as running text on CRAN.
• Vignettes: example output files in ThSQCA_Reproducible_EN are now named ThSQCA_* instead of TSQCA_*; TSQCA_MCTS_results.csv became ThSQCA_CTS_multi_results.csv. Added a pointer to Sweep Builder in ThSQCA_Tutorial_EN.

Bug fixes and output changes

• Sweeps no longer print one stray blank line for every "No solution" cell. The calls to QCA::truthTable() and QCA::minimize() inside the sweep loops now go through an internal helper, quiet_try(), which behaves like try(..., silent = TRUE) but also discards console output written by the QCA call itself. Warnings are passed on unchanged (they are collected while the output is diverted and signalled again afterwards). Results are unchanged.
• A pre_calibrated variable that contains memberships of exactly 0.5 now produces one warning per sweep, naming the variable and the number of cases. Previously QCA's own warning ("Fuzzy causal conditions should not have values of 0.5 in the data") was repeated for every cell of the sweep (twice per cell) without saying which variable was affected; that per-cell warning is no longer repeated.
• Printed and summarised results now use the ThS-QCA labels (OTS, DTS, CTS (single), CTS (multiple)) instead of "OTS-QCA", "DTS-QCA", "CTS-QCA" and "MCTS-QCA". The S3 class names (otSweep_result, tsqca_result, and so on) are unchanged, so existing code that dispatches on them keeps working.
• The Necessity Analysis table in full reports no longer shows a leading space in some condition names (for example " ~TRU").
• The verification code at the end of a report now reproduces the solution type that was actually used (minimize(tt), minimize(tt, include = "?") or minimize(tt, include = "?", dir.exp = c(...)) with your values) instead of a placeholder.
• generate_report() now shows the outcome name you supplied (for example LOY, or ~LOY for a negated outcome) in solution formulas, in the captured QCA console output, and in the "Solutions Overview" headings, where it previously showed the internal column name Y. The verification snippet at the end of each report now names your outcome and conditions and uses your incl.cut. Internally the outcome column is still called Y.
• The documentation no longer recommends sweeping variables only on their raw scale. Membership scores can be swept like any other numeric variable (leave them out of pre_calibrated and give thresholds on the 0 to 1 scale); the tutorial has a new subsection on this. The warning for a variable that is both in pre_calibrated and in a sweep list now says how to sweep it (remove it from pre_calibrated) and no longer points to a vignette section that does not exist. The rule itself is unchanged: a pre_calibrated variable is used as it is and its sweep thresholds are ignored.
• The warning issued when a sweep finds more than one equivalent solution now reads "Multiple equivalent solutions exist ..." (it used to say "intermediate solutions" even for complex and parsimonious runs, where it also applies).

ThSQCA 2.0.6

Bug fixes

• get_n_solutions() no longer double-counts models when the intermediate solution spans multiple prime implicant charts. When dir.exp produces more than one prime implicant chart (QCA's own chart indexing, visible as "From C1P1, C2P1:" in print(), as opposed to multiple parsimonious paths within a single chart, "From C1P1, C1P2:"), sol$i.sol holds one list entry per chart/path combination. Each entry's own $solution field is computed independently by QCA::minimize() and is not guaranteed to be chart-exclusive: depending on the data, it may already enumerate the full, cross-chart set of tied minimal models. Summing length($solution) across every chart name, as get_n_solutions() previously did, could therefore double- (or N-fold) count identical models. This was visible as an inflated n_solutions in res$summary and in the Summary Table of generate_report()'s Markdown output; the displayed solution formula itself was already correct. A new internal helper, collect_unique_i_sol(), performs the same enumeration and then deduplicates by comparing each model's term set (order-independent), so structurally identical models are counted once regardless of which chart(s) produced them. generate_report()'s two sol_list-construction blocks (used for the "Full Solutions" listing and as input to identify_epi()) were updated to use the same helper; the identify_epi() EPI/SPI content itself was unaffected by the duplication (set operations are insensitive to duplicate entries), only the displayed solution count was wrong. Cells with a single prime implicant chart, which are the common case, are unaffected.

• The standalone identify_epi() function itself was audited and confirmed correct: it is a pure function over its solutions argument and does not read sol$solution internally. All three internal call sites already built their input via the sol$i.sol traversal above rather than a naive sol$solution read; the double-counting above was the only defect found in that path. Passing sol$solution (top level) directly to identify_epi(), as opposed to a sol_list built via sol$i.sol, is still not meaningful when dir.exp is specified and should be avoided: that slot holds the parsimonious solution's terms in that case, not the intermediate solution's.

• qca_extract(extract_mode = "all" | "essential") now summarizes across the same set of solutions that n_solutions counts. These two modes derived their term sets from sol$i.sol$C1P1$solution, that is, from the first prime implicant chart only. That slot is the right basis for extract_mode = "first" (its first entry is the displayed M1), but whether it enumerates every chart's model or only its own is an implementation detail of QCA::minimize()'s internal getSolution() call rather than a documented guarantee. Both modes now build their model list with collect_unique_i_sol(), the same enumeration get_n_solutions() counts, so the reported EPI/SPI terms and the reported solution count can no longer describe different sets. On data where the first chart already enumerated every model, which includes all cases checked here, the output is unchanged.

• The same duplication defect was found and fixed in two further places. A follow-up audit of every sol$i.sol traversal in the package turned up two more sites built on the same chart-by-chart concatenation: extract_solution_list() (used by generate_config_chart()), where the inflated count made the chart announce too many equivalent solutions and emit one identical table per duplicate; and extract_sol_terms_by_model() (used by compute_fiss_core()), where it inflated the reported parsim_n_solutions and the tie warning derived from it. The core/peripheral classification itself was not affected, because it counts a status only where every minimal solution agrees, which is insensitive to duplicates. Both now deduplicate by term-set content. extract_all_metrics() reads the first chart's IC deliberately, to match the displayed M1, and was left as it is.

API consistency and usability

• ctSweepM() gained a thrX_default argument and no longer fails silently. A condition present in conditions but absent from both sweep_list and pre_calibrated previously had no column created at all, which made QCA::truthTable() fail and returned "No solution" for every cell with no error or warning. Such a condition is now either binarized at thrX_default (mirroring ctSweepS()) or, if thrX_default is not supplied, reported as an error naming the uncovered conditions and showing both ways to fix the call. Calls that already listed every condition in sweep_list are unaffected and produce identical results.
• print_fiss_summary() no longer requires thr_key. Omitting it raised R's generic "argument is missing" error, and discovering the valid keys meant inspecting names(result$fiss_core) by hand. thr_key now defaults to NULL, which summarizes every threshold level in turn. Passing an explicit thr_key behaves as before, including the existing error that lists the available keys.
• Product terms in configuration charts are now labeled T1, T2, ... instead of M1, M2, ... M1/M2 denote whole alternative minimal solutions throughout the package (as in print(sol)), so reusing M for the product terms within one solution made the two levels indistinguishable. This affects the column headers produced by build_config_matrix() (used by config_chart_multi_solutions() and related chart functions), the term columns of the cross-threshold Fiss chart, and the [Term ...] headings printed by print_fiss_summary(). Charts now read, for example, "Solution M1" with term columns T1, T2. This is a change to displayed output only; no computed value is affected.
• The @return documentation of all four sweep functions (otSweep(), ctSweepS(), ctSweepM(), dtSweep()) now states explicitly that return_details changes the type of the returned object, not only its contents: with TRUE the summary table is at result$summary, while with FALSE the summary table is the returned object and result$summary is NULL. Code meant to work under both settings should branch on inherits(result, "data.frame") or always pass return_details = TRUE. The behavior itself is unchanged, since altering the return type would break existing scripts.

ThSQCA 2.0.5

Release date: 2026-07-23

Bug fixes

• Fit measures are reported again when the intermediate solution itself has several minimal solutions. This is a regression introduced in 2.0.4. When dir.exp is used and the intermediate solution has several tied minimal solutions, QCA::minimize() splits the intermediate fit into $i.sol$C1P1$IC$individual[[k]] and $overall rather than storing a flat $sol.incl.cov, exactly as it does for the parsimonious solution. Version 2.0.4 read only the flat slot, so those cells reported inclS and covS as NA. The intermediate path now follows the same cascade as the parsimonious path: the displayed solution's own fit for extract_mode = "first", and the aggregate for "all" and "essential". Cells that reported a fit in 2.0.4 are unchanged.
• compute_fiss_core() no longer biases conditions toward "core" when the parsimonious solution has several tied minimal solutions. The terms of all minimal solutions were pooled before the condition-status map was built, so a condition present in one minimal solution and absent in another ended up holding both statuses. An intermediate term then matched whichever polarity it used, and the condition was classified core either way. A status now counts only where every minimal parsimonious solution agrees, following the same principle as essential prime implicants: what cannot be asserted regardless of which minimal solution is selected is not treated as core. Cells with a single minimal parsimonious solution, which are the common case, are unaffected.
• compute_fiss_core() now records parsim_n_solutions for each threshold and warns when the parsimonious solution has tied minimal solutions, so the ambiguity behind a classification is visible rather than hidden.
• The internal chart helpers extract_solution_metrics_for_chart() and extract_path_metrics_for_chart() had the same gap. There it produced not NA but a silent fall-through to the parsimonious values, so configuration charts could show the parsimonious fit and the parsimonious per-term table for a displayed intermediate solution with several minimal solutions.

Other changes

• If a solution is found but no fit measures can be located for it, the sweep functions now emit a warning explaining that inclS and covS are NA for that reason, instead of returning an unexplained NA.
• Added regression tests for the multiple-intermediate-solution case, covering qca_extract(), both chart helpers, generate_report(), and the sweep functions end to end.

ThSQCA 2.0.4

Release date: 2026-07-22

Bug fixes

• Fit measures for the intermediate solution now come from the intermediate solution itself. When dir.exp is specified, the sweep functions display the intermediate solution (sol$i.sol$C1P1$solution), but qca_extract() read inclS/covS from sol$IC, which describes the parsimonious solution. The displayed intermediate formula was therefore paired with the parsimonious fit: for a single parsimonious solution via sol$IC$sol.incl.cov, and for several (tied) parsimonious solutions via sol$IC$overall (the disjunction of the parsimonious solutions). The extractor now branches on the displayed solution and reads sol$i.sol$C1P1$IC$sol.incl.cov for intermediate solutions, so the fit matches the displayed expression and QCA::minimize()'s own print(). This affects the intermediate-solution output of otSweep(), ctSweepS(), ctSweepM(), and dtSweep() in every extract_mode, whether the cell has one or several minimal solutions.
• The parsimonious/complex fix from 2.0.3 (first-solution fit read from sol$IC$individual[[1]] rather than sol$IC$overall) is retained. Complex and parsimonious solutions, crisp-set analyses, single-solution cells, and no-solution cells are unchanged.
• The report and chart fit extractors received the same correction. The same fit-source confusion existed independently in extract_all_metrics() (used by generate_report()) and in the internal chart helpers extract_solution_metrics_for_chart() and extract_path_metrics_for_chart(): with multiple minimal solutions they reported the overall aggregate rather than the displayed solution's own fit, and for intermediate solutions they could report the parsimonious fit. They now prefer the intermediate solution (when dir.exp is used) and then the displayed solution's individual[[k]] fit, falling back to overall only when the per-solution fit is unavailable. generate_report() and the configuration charts now show fit measures consistent with the sweep tables.
• generate_report() now reports the intermediate fit in every section. Beyond the extractor fix above, the report's "Solution Fit" and "Cross-Threshold Comparison" sections passed sol$IC (the parsimonious solution) to the extractor even when the displayed solution was the intermediate one, so those sections showed the parsimonious inclS/covS (and a parsimonious per-term row) under an intermediate formula, contradicting the report's own Summary Table and its embedded QCA verification output. Both call sites now select sol$i.sol$C1P1$IC for intermediate solutions, matching the Summary Table. Non-intermediate reports are byte-for-byte unchanged.

ThSQCA 2.0.3

Release date: 2026-07-22

Bug fixes

• Fit measures now match the displayed solution when multiple minimal solutions exist (fuzzy data). When QCA::minimize() returns more than one minimal solution, it stores the fit of each solution in sol$IC$individual[[k]]$sol.incl.cov and the fit of the disjunction of all solutions in sol$IC$overall$sol.incl.cov. With extract_mode = "first" (the default) the sweep functions display the first solution (M1), but the internal extractor qca_extract() reported the overall aggregate for inclS/covS, so the printed formula (M1) and the printed fit did not correspond. The extractor now reads sol$IC$individual[[1]]$sol.incl.cov for the first solution. This affects otSweep(), ctSweepS(), ctSweepM(), and dtSweep(), which share qca_extract().
• Impact is confined to cells with multiple minimal solutions on fuzzy data, where overall (the union coverage) is greater than any single solution's coverage; the previously reported covS was biased upward. covS is the materially affected measure; inclS differs only marginally.
• Crisp-set (csQCA) results are unchanged: for crisp data the per-solution and overall fit coincide, so single-solution and no-solution cells, and all crisp analyses, return identical values to previous versions.
• extract_mode = "all" and extract_mode = "essential" still report the overall aggregate, which is appropriate because those modes summarize across all solutions rather than displaying one.

ThSQCA 2.0.2

Release date: 2026-07-15

Documentation and metadata

• Corrected an author attribution in DESCRIPTION and the package-level help. The robustness protocol at is by Oana and Schneider (2024), not "Rubinson et al. (2019)". The vignettes and README already cited this work correctly; this aligns the package metadata with them.
• inst/CITATION now derives the package version dynamically via paste("R package version", meta$Version) instead of hard-coding it.

ThSQCA 2.0.1

Release date: 2026-07-xx

Bug fixes

• Guard against QCA::truthTable() type issues before minimization. With QCA 3.25 / admisc 0.40, truthTable() can, for sparse truth tables, return the incl/PRI columns as character and represent logical-remainder rows (observed n = 0) with the string "-". Passing such a truth table to QCA::minimize() can, for some truth table structures, cause the minimization to hang or return misleading fit values. All sweep functions (otSweep(), ctSweepS(), ctSweepM(), dtSweep()) and the Fiss parsimonious step now coerce these columns to numeric and set remainder rows to 0 via an internal sanitize_truthtable() helper before calling minimize(). The OUT column is left untouched, so remainder handling is unaffected.

Notes

• This fix concerns robustness on sparse (typically crisp) truth tables that can arise across a sweep. It does not change results that already ran to completion: the guarded conditions either halt a run or leave completed output unchanged, so previously obtained, completed sweep results are not silently affected. Re-running is only necessary if a prior sweep hung or failed to return a solution on data that produced sparse truth tables.

ThSQCA 2.0.0

Release date: 2026-05-XX

Package Renamed

The package has been renamed from TSQCA to ThSQCA (Threshold-Sweep QCA).

Background

Reviewers noted that TSQCA risks being misread as Time-Series QCA, since ts is the base R class for time-series objects. The new name ThSQCA unambiguously reflects the "Threshold-Sweep" methodology described in Toyoda (2026b, Quality & Quantity).

Migration

Users of TSQCA should update their code as follows:

# Old
install.packages("TSQCA")
library(TSQCA)

# New
install.packages("ThSQCA")
library(ThSQCA)

All function names (otSweep, ctSweepS, ctSweepM, dtSweep, etc.) and internal logic remain unchanged. TSQCA will remain on CRAN for backward compatibility but will not receive further updates.

Reference

Toyoda, Y. (2026b). Threshold-Sweep QCA (ThS-QCA): Systematic exploration of threshold dependency in Qualitative Comparative Analysis. Quality & Quantity (forthcoming).

ThSQCA 1.3.2

Release date: 2026-03-14

New Features

Fiss (2011) Core/Peripheral Condition Classification

Three new functions implement the core/peripheral distinction introduced by Fiss (2011, Academy of Management Journal), which distinguishes between conditions that are central to a causal configuration and those that merely supplement it:

• compute_fiss_core(result, conditions) — Augments any sweep result object with core/peripheral classification. For each threshold, it automatically re-runs QCA::minimize() with dir.exp = NULL to obtain the parsimonious solution, then compares it to the already-stored intermediate solution. Conditions appearing in both solutions are classified as core; conditions appearing only in the intermediate solution are classified as peripheral.

• generate_fiss_chart(result, conditions, symbol_set, language) — Generates a Markdown-formatted cross-threshold configuration chart using four distinct symbols:

| Symbol (unicode) | Symbol (latex) | Symbol (ascii) | Meaning | |-----------------|----------------|----------------|---------| | ● | $\bullet$ | O | Core condition present | | ⊗ | $\otimes$ | X | Core condition absent | | ⊙ | $\odot$ | o | Peripheral condition present | | ⊘ | $\oslash$ | x | Peripheral condition absent |

• print_fiss_summary(result, thr_key, language) — Prints a human-readable summary of the core/peripheral classification for a specific threshold value.

Usage:

# Step 1: Run intermediate sweep (include = "?" + dir.exp required)
res <- otSweep(
  dat = sample_data, outcome = "Y",
  conditions = c("X1", "X2", "X3"),
  sweep_range = 6:8, thrX = c(X1 = 7, X2 = 7, X3 = 7),
  include = "?", dir.exp = c(1, 1, 1),
  return_details = TRUE
)

# Step 2: Compute Fiss classification
res_fiss <- compute_fiss_core(res, conditions = c("X1", "X2", "X3"))

# Step 3: Generate Fiss-style four-symbol chart
cat(generate_fiss_chart(res_fiss, symbol_set = "unicode"))
cat(generate_fiss_chart(res_fiss, symbol_set = "latex"))   # for papers

# Step 4: Inspect a specific threshold
print_fiss_summary(res_fiss, thr_key = "7")

Prerequisites: compute_fiss_core() requires return_details = TRUE, include = "?", and dir.exp to be specified in the original sweep. If any of these are missing, an informative error is raised.

Fiss Integration in generate_report()

generate_report() gains a new include_fiss_core argument (default FALSE). When set to TRUE on a result augmented by compute_fiss_core(), all configuration charts in the report use the four-symbol Fiss notation and the Notes section includes the Fiss (2011) reference and symbol legend.

res_fiss <- compute_fiss_core(res)
generate_report(res_fiss, "report.md",
                include_fiss_core = TRUE,
                chart_symbol_set  = "unicode")

Updated Label Dictionary (get_config_labels)

Both English and Japanese label dictionaries in get_config_labels() now include Fiss-specific entries: fiss_core, fiss_peripheral, fiss_parsim, fiss_interm, fiss_note.

Internal Changes

• Added SYMBOL_SETS_FISS constant (analogous to SYMBOL_SETS) with four-symbol sets for unicode, latex, and ascii output formats.
• Added internal helpers: extract_cond_status_map(), classify_term_conditions(), run_parsimonious(), extract_sol_terms(), build_fiss_matrix().
• write_full_report() and write_simple_report() accept a new use_fiss logical parameter propagated from generate_report().

Documentation

• Vignette TSQCA_Tutorial_EN.Rmd gains a new section "Fiss (2011) Core/Peripheral Classification (New in v1.3.2)" with full workflow and interpretation guidance.
• Vignette TSQCA_Reproducible_EN.Rmd gains a new Section 12 with complete reproducible Fiss workflow code.
• New Rd documentation files for compute_fiss_core, generate_fiss_chart, print_fiss_summary, and SYMBOL_SETS_FISS.

Reference

Fiss, P. C. (2011). Building better causal theories: A fuzzy set approach to typologies in organization research. Academy of Management Journal, 54(2), 393–420.

ThSQCA 1.3.1

Release date: 2026-02-18

Documentation

• Removed non-reproducible code example from the pre_calibrated vignette section. The example referenced variables not included in the bundled sample_data. The section now provides a prose description of the feature; a worked example will be added when a suitable public dataset is available.
• Corrected @param documentation for thrX, thrX_default, sweep_list, and sweep_list_X: pre-calibrated variables do not require a threshold entry (the previous documentation incorrectly stated otherwise).

Note

This is a documentation-only patch. No changes to code logic or behavior.

ThSQCA 1.3.0

Release date: 2026-02-18

New Features

Pre-Calibrated Variable Pass-Through (PCVP)

All four sweep functions (otSweep, dtSweep, ctSweepS, ctSweepM) now support a pre_calibrated argument. Variables listed in pre_calibrated are passed through to QCA::truthTable() without binarization, enabling mixed crisp/fuzzy analyses where some conditions are pre-calibrated via QCA::calibrate() while others are binarized by threshold sweep.

Usage:

# AGE is pre-calibrated as a fuzzy set; other conditions are binarized
result <- otSweep(
  dat            = dat,
  outcome        = "INT",
  conditions     = c("CHT", "PRC", "UNQ", "AGE", "GEN"),
  sweep_range    = 6:9,
  thrX           = c(CHT = 7, PRC = 7, UNQ = 7, AGE = 0.5, GEN = 1),
  pre_calibrated = c("AGE"),
  include        = "?",
  dir.exp        = c(1, 1, 1, "-", "-"),
  incl.cut       = 0.80,
  n.cut          = 2,
  pri.cut        = 0.50
)

Validation: The function raises an error if a pre-calibrated variable is not found in conditions, or if its values fall outside the [0, 1] range. A warning is issued if a pre-calibrated variable is also listed as a sweep target (in sweep_list_X / sweep_list), since threshold sweeping has no effect on fixed fuzzy values.

Backward compatibility: When pre_calibrated = NULL (the default), all functions produce exactly the same output as v1.2.0.

Internal Changes

• New internal helper prepare_dat_bin() in tsqca_core.R centralizes data preparation logic, replacing the inline binarization code that was duplicated across all four sweep functions.
• New internal helper validate_pre_calibrated() in tsqca_core.R performs input validation for the pre_calibrated parameter.
• pre_calibrated is now stored in the params object returned by all sweep functions (when return_details = TRUE).
• generate_report() now displays pre-calibrated conditions in the Analysis Overview section.

ThSQCA 1.2.0

Release date: 2026-01-19

Bug Fixes

CRITICAL: Fixed Intermediate Solution Extraction

Problem: When dir.exp is specified for intermediate solutions, the QCA package stores: - sol$solution — Contains the Parsimonious solution - sol$i.sol$C1P1$solution — Contains the true Intermediate solution

Previous versions of TSQCA incorrectly prioritized sol$solution, causing Parsimonious solutions to be extracted and displayed when Intermediate solutions were expected.

Fix: All solution extraction functions now correctly prioritize sol$i.sol when available:

• get_n_solutions() — Now checks i.sol first
• qca_extract() — Now checks i.sol first
• extract_solution_list() — Now checks i.sol first
• write_full_report() — Fixed 3 locations
• write_simple_report() — Fixed 1 location

Impact: Users who specified dir.exp for intermediate solutions may have received incorrect results in: - Report generation (generate_report()) - Configuration charts - Solution expression extraction

Verification: The print(sol) output was always correct because the QCA package's print method handles this correctly. Only programmatic extraction was affected.

New Features

Solution Type Display

Reports now explicitly display the solution type in the Analysis Overview section:

| Include | dir.exp | Solution Type | |---------|---------|-------------------| | "" | any | Complex (Conservative) | | "?" | NULL | Parsimonious | | "?" | specified | Intermediate |

QCA Package Output for Verification

Added optional raw QCA output section to reports for verification purposes:

generate_report(result, "report.md", include_raw_output = TRUE)  # default
generate_report(result, "report.md", include_raw_output = FALSE) # disable

When enabled, each threshold's detailed results include:

#### QCA Package Output (for verification)

DEVURBLITSTB + DEVLIT~INDSTB -> SURV

This allows researchers to verify that TSQCA's extraction matches the QCA package's native output.

Migration Guide

If you used intermediate solutions (with dir.exp) in previous versions, we recommend re-running analyses to ensure correct results. Compare your new results with print(minimize(...)) output for verification.

ThSQCA 1.1.0

Release date: 2026-01-17

Bug Fixes

CRITICAL: Fixed dir.exp = NULL Behavior

Problem: In v1.0.0, when dir.exp = NULL (the default), the package incorrectly converted it to c(1, 1, ...), which forced intermediate solution calculation regardless of user intent.

Fix: dir.exp = NULL is now correctly passed to QCA::minimize() without modification.

Breaking Changes

Default Arguments Now Match QCA Package

To ensure consistency with the QCA package, default argument values have been changed:

| Argument | v1.0.0 Default | v1.1.0 Default | Effect | |----------|---------------|---------------|--------| | include | "?" | "" | Complex solution (no logical remainders) | | dir.exp | NULL → c(1,1,...) (bug) | NULL | No directional expectations |

Result: TSQCA now produces complex solutions by default, matching QCA::minimize() default behavior.

Solution Type Summary

| Solution Type | How to Compute | |--------------|----------------| | Complex (default) | include = "", dir.exp = NULL | | Parsimonious | include = "?", dir.exp = NULL | | Intermediate | include = "?", dir.exp = c(1, 1, ...) |

Migration Guide

# v1.0.0 (incorrect: intermediate solution by default due to bug)
result <- otSweep(dat, "Y", c("X1", "X2", "X3"), sweep_range = 7, thrX = thrX)

# v1.1.0: Complex solution (new default, QCA compatible)
result_comp <- otSweep(dat, "Y", c("X1", "X2", "X3"), sweep_range = 7, thrX = thrX)

# v1.1.0: Parsimonious solution (include = "?")
result_pars <- otSweep(dat, "Y", c("X1", "X2", "X3"), sweep_range = 7, thrX = thrX,
                       include = "?")

# v1.1.0: Intermediate solution (include = "?" + dir.exp)
result_int <- otSweep(dat, "Y", c("X1", "X2", "X3"), sweep_range = 7, thrX = thrX,
                      include = "?",
                      dir.exp = c(1, 1, 1))

Documentation Improvements

Enhanced Examples for All Sweep Functions

All four sweep functions (otSweep, dtSweep, ctSweepS, ctSweepM) now include examples demonstrating:

1. Complex Solution — default (QCA compatible)
2. Parsimonious Solution — using include = "?"
3. Intermediate Solution — using include = "?" with dir.exp

Updated Parameter Documentation

The @param dir.exp and @param include documentation now clearly explains: - Default behavior produces complex solutions (QCA compatible) - include = "?" enables logical remainders for parsimonious/intermediate - dir.exp specifies directional expectations for intermediate solutions

ThSQCA 1.0.0

Release date: 2026-01-06

Changes

Default Chart Level Changed to "term" (Fiss-style)

The default value for chart_level parameter has been changed from "summary" to "term".

Rationale: The solution-term level format (Fiss, 2011 notation) is the standard for academic publications, where each column represents one prime implicant (configuration). The previous default ("summary") aggregated all configurations at each threshold into a single column, which obscured the distinction between different sufficient paths.

Column header format updated: - Old format: thrY=6_M1 - New format: thrY = 6 (M1) (consistent with the paper format)

Affected functions: - generate_report() — default chart_level is now "term" - generate_cross_threshold_chart() — default chart_level is now "term"

Migration: If you prefer the previous behavior (threshold-level summary), explicitly specify chart_level = "summary":

generate_report(result, "report.md", chart_level = "summary")
generate_cross_threshold_chart(result, conditions, chart_level = "summary")

ThSQCA 0.5.3

Release date: 2026-01-03

New Features

Solution-Term Level Configuration Charts (Fiss-style)

Added support for solution-term level configuration charts following Fiss (2011) notation. This feature allows generating charts where each column represents a single prime implicant (configuration), which is the standard format for academic publications.

New parameter for generate_report():

• chart_level — Character. Either "summary" (default) or "term".
• "summary": Threshold-level summaries where each column represents one threshold, showing all conditions that appear in any configuration at that threshold.
• "term": Solution-term level (Fiss-style) where each column represents one prime implicant (sufficient configuration). Recommended for academic publications.

New functions:

• generate_cross_threshold_chart() — Generate configuration charts from sweep results with chart_level option.
• parse_solution_terms() — Internal function to parse solution expressions into individual terms.
• get_condition_status() — Internal function to determine condition presence/absence in a term.
• generate_term_level_chart() — Internal function for term-level chart generation.
• generate_threshold_level_chart() — Internal function for threshold-level chart generation.

Example:

# Threshold-level summary (default)
generate_report(result, "report.md", chart_level = "summary")

# Solution-term level (Fiss-style, recommended for publications)
generate_report(result, "report.md", chart_level = "term")

When the solution is X3 + X1*X2, the term-level chart will show two separate columns (thrY=7_M1 for X3 and thrY=7_M2 for X1*X2), while the summary-level chart shows one column (thrY=7) with all three conditions marked.

ThSQCA 0.5.2

Release date: 2026-01-03

Documentation Fixes

Vignette Examples Corrected

Fixed non-working code examples in all vignettes (Tutorial and Reproducible, both EN/JA):

• Updated argument names from deprecated Yvar/Xvars to outcome/conditions
• Fixed ctSweepM() examples to use sweep_list parameter instead of old sweep_vars/sweep_range
• Added required dat parameter to all generate_report() examples
• Changed output from head(res$summary) to summary(res) for consistency with S3 methods
• Reduced sweep ranges (e.g., 6:9 → 6:8, 6:8 → 6:7) for faster example execution

README Updates

• Updated all code examples to use new argument names (outcome, conditions)
• Added dat parameter to generate_report() examples
• Consistent sweep ranges across all examples

Test Scripts Added

• test_quick.R — Minimal verification script (6 tests)
• test_tutorial_code.R — Comprehensive verification script (11 tests)

ThSQCA 0.5.1

Release date: 2026-01-01

New Features

Multiple Solutions Note in Configuration Charts

When multiple logically equivalent solutions (M1, M2, M3...) exist, configuration charts now automatically include a note explaining that M1 is displayed.

New parameters for generate_report():

• solution_note — Logical. If TRUE (default), adds note when multiple solutions exist
• solution_note_style — "simple" (default) or "detailed" (includes EPIs)
• solution_note_lang — "en" (default) or "ja" for Japanese

New parameters for config_chart_from_paths():

• n_sol — Number of equivalent solutions (triggers note if > 1)
• solution_note — Logical. Whether to add solution note
• solution_note_style — "simple" or "detailed"
• epi_list — Character vector of EPIs for detailed notes

New exported functions:

• generate_solution_note() — Generate solution note text
• identify_epi() — Identify Essential Prime Implicants from multiple solutions

Example output (simple):

*Note: 2 logically equivalent solutions were identified. This table presents configurations based on M1.*

Example output (detailed with EPIs):

*Note: 3 logically equivalent solutions were identified (M1-M3). This table presents configurations based on M1. All solutions share the essential prime implicants: A·B and C.*

Example (Japanese):

*注: 論理的に等価な2つの解が得られた。本表はM1に基づく構成を示す。*

ThSQCA 0.5.0

Release date: 2025-12-31

New Features

Configuration Chart Integration in Reports

Configuration charts are now automatically included in reports generated by generate_report().

New parameters for generate_report():

• include_chart — Logical. If TRUE (default), includes Fiss-style configuration charts
• chart_symbol_set — Symbol set: "unicode" (default), "ascii", or "latex"

Example:

# Generate report with configuration charts (default)
generate_report(result, "my_report.md", format = "full")

# Generate report without charts
generate_report(result, "my_report.md", include_chart = FALSE)

# Generate report with LaTeX symbols (for PDF/academic papers)
generate_report(result, "my_report.md", chart_symbol_set = "latex")

Standalone Configuration Chart Functions

• generate_config_chart() — Generate chart from QCA solution object
• config_chart_from_paths() — Generate chart from path strings (e.g., "AB~C")
• config_chart_multi_solutions() — Generate separate charts for multiple solutions

Features:

• Three symbol sets: "unicode" (● / ⊗), "ascii" (O / X), "latex" ($\bullet$ / $\otimes$)
• Automatic condition extraction from solution paths
• Bilingual labels (English / Japanese)
• Markdown table output

Breaking Changes (from 0.4.x)

Terminology Correction

Fixed incorrect use of "Core Conditions" terminology:

| Old (incorrect) | New (correct) | Meaning | |-----------------|---------------|---------| | extract_mode = "core" | extract_mode = "essential" | Mode for extracting shared terms | | core_terms | essential_terms | Terms in ALL solutions | | peripheral_terms | selective_terms | Terms in SOME solutions |

Migration: Change extract_mode = "core" to extract_mode = "essential".

Documentation

• Added docs/TSQCA_Terminology_Guide_EN.md — English terminology guide
• Added docs/TSQCA_Terminology_Guide_JA.md — Japanese terminology guide
• Added README_JP.md — Japanese README

ThSQCA 0.4.2

Terminology Correction (Breaking Change)

Corrected QCA Terminology for Multiple Solutions

Fixed incorrect use of "Core Conditions" terminology. The terms that appear in ALL equivalent solutions (M1, M2, M3...) are now correctly called Essential Prime Implicants (EPI), following standard Boolean minimization terminology.

Changed terms:

| Old (incorrect) | New (correct) | Meaning | |-----------------|---------------|---------| | extract_mode = "core" | extract_mode = "essential" | Mode for extracting shared terms | | core_terms | essential_terms | Terms in ALL solutions | | peripheral_terms | selective_terms | Terms in SOME solutions | | "Core Conditions" | "Essential Prime Implicants (EPI)" | Report labels | | "Peripheral Terms" | "Selective Prime Implicants (SPI)" | Report labels |

Why this matters:

The term "Core Conditions" in QCA literature (Fiss, 2011) refers to conditions appearing in both parsimonious AND intermediate solutions—a comparison between solution types. This is distinct from terms shared across multiple equivalent solutions of the same type, which are properly called "Essential Prime Implicants" in Boolean algebra terminology.

Migration:

If you used extract_mode = "core" in previous versions, change to extract_mode = "essential". The output structure is identical; only the names have changed for methodological accuracy.

New Features

Configuration Chart Generator

Added functions for generating Fiss-style configuration charts (Table 5 format) commonly used in QCA publications.

New functions:

• generate_config_chart() — Generate configuration chart from QCA solution object
• config_chart_from_paths() — Generate chart from path strings (e.g., "AB~C")
• config_chart_multi_solutions() — Generate separate charts for multiple solutions

Features:

• Three symbol sets: "unicode" (● / ⊗), "ascii" (O / X), "latex" ($\bullet$ / $\otimes$)
• Automatic condition extraction from solution paths
• Optional metrics rows (Consistency, Coverage, Unique Coverage)
• Bilingual labels (English / Japanese)
• Markdown table output for easy integration with reports

Example:

# From QCA solution object
chart <- generate_config_chart(sol, symbol_set = "unicode")
cat(chart)

# From path strings
paths <- c("A*B*~C", "A*D")
chart <- config_chart_from_paths(paths)
cat(chart)

ThSQCA 0.4.1

Bug Fixes

Report Generation Improvements

• Fixed empty "Detailed Results" and "Cross-Threshold Comparison" sections for large threshold sweeps
• Added threshold limit (27 combinations) for detailed output in reports
• When threshold combinations exceed 27, detailed per-threshold results are omitted with explanatory message
• Users are directed to access details programmatically via result$details
• Affected functions: generate_report() for dtSweep and ctSweepM results

ThSQCA 0.4.0

New Features

S3 Methods

• Added S3 class system for all sweep function results
• Class hierarchy: otSweep_result, dtSweep_result, ctSweepS_result, ctSweepM_result inherit from tsqca_result
• Added print() methods for all result types
• Displays analysis overview: outcome, conditions, thresholds swept
• Shows summary statistics: valid solutions, no solution, multiple solutions
• Added summary() methods for all result types
• Displays analysis parameters and full results table
• Notes when multiple solutions exist

Changes

Backward Compatibility

• All existing workflows continue to work unchanged
• Direct access to $summary, $details, and $params components still works
• When return_details = FALSE, returns plain data.frame without S3 class

ThSQCA 0.3.0

New Features

QCA-Compatible Argument Names

• Renamed Yvar to outcome and Xvars to conditions in all sweep functions
• Follows QCA package naming conventions for consistency
• Old argument names (Yvar, Xvars) are still supported with deprecation warnings

Negated Outcome Support

• Added support for negated outcomes using tilde prefix (e.g., outcome = "~Y")
• Analyzes conditions sufficient for the absence of the outcome (Y < threshold)
• Follows QCA package's truthTable() convention for negation
• Works with all sweep functions: otSweep(), dtSweep(), ctSweepS(), ctSweepM()

Enhanced Report Generation

• generate_report() now displays "(negated)" indicator when analyzing negated outcomes
• Supports both old and new parameter names for backward compatibility

Changes

Argument Names (Backward Compatible)

• Yvar → outcome (recommended)
• Xvars → conditions (recommended)
• Using old argument names will trigger a deprecation warning but will continue to work

Parameter Storage

• $params now includes:
• outcome: New argument name (also stores ~Y notation if negated)
• conditions: New argument name
• negate_outcome: Boolean indicating if outcome was negated

Migration Guide

# Old syntax (still works, but shows deprecation warning)
result <- otSweep(dat, Yvar = "Y", Xvars = c("X1", "X2"), ...)

# New syntax (recommended)
result <- otSweep(dat, outcome = "Y", conditions = c("X1", "X2"), ...)

# Negated outcome (new feature)
result <- otSweep(dat, outcome = "~Y", conditions = c("X1", "X2"), ...)

ThSQCA 0.2.0

New Features

Multiple Solution Handling

• Added extract_mode parameter to all sweep functions (otSweep(), dtSweep(), ctSweepS(), ctSweepM()) with three options:
• "first" (default): Returns only the first solution (M1), maintaining backward compatibility
• "all": Returns all intermediate solutions concatenated (e.g., "M1: AB; M2: AC")

• "essential": Returns essential prime implicants common to all solutions, plus peripheral and unique terms

• Added get_n_solutions() helper function to count the number of intermediate solutions

Report Generation

• Added generate_report() function for automatic markdown report generation with two formats:
• "full": Comprehensive report including all analysis details, solution formulas, and fit measures
• "simple": Condensed format designed for journal manuscript supplementary materials

Reproducibility

• All sweep functions now return analysis parameters in $params for full reproducibility
• Parameters include: variable names, thresholds, QCA settings (incl.cut, n.cut, pri.cut, dir.exp, include)

Changes

Default Value Updates

• Changed return_details default from FALSE to TRUE for better integration with generate_report()
• Changed n.cut default from 2 to 1 to align with QCA package conventions
• Changed pri.cut default from 0.5 to 0 to align with QCA package conventions

Output Structure

• When return_details = TRUE, results are now accessed via $summary (e.g., result$summary$expression)
• Added n_solutions column when using extract_mode = "all" or "essential"
• Added selective_terms and unique_terms columns when using extract_mode = "essential"

Documentation

• Updated README with new features section and usage examples
• Added new vignette sections:
• "Handling Multiple Solutions" explaining essential vs. selective prime implicants
• "Generating Reports" with workflow examples
• "Best Practices" including computational complexity guidance
• Updated all code examples to reflect new default values and output structure

ThSQCA 0.1.2

• Initial release for paper submission
• Implemented four threshold sweep methods:
• ctSweepS(): Single-condition X sweep (CTS-QCA)
• ctSweepM(): Multi-condition X sweep (MCTS-QCA)
• otSweep(): Outcome Y sweep (OTS-QCA)
• dtSweep(): Two-dimensional X and Y sweep (DTS-QCA)
• Core QCA functions: qca_bin(), qca_extract()
• Sample dataset included

ThSQCA 0.1.1

• Bug fixes and documentation improvements

ThSQCA 0.1.0

• Initial development version

ThSQCA documentation built on Sept. 26, 2026, 5:07 p.m.