Performs a grid search over thresholds of multiple X variables. For each combination of thresholds in sweep_list, the outcome Y and all X variables are binarized, and a crisp-set QCA is executed.

ctSweepM(
  dat,
  outcome = NULL,
  conditions = NULL,
  sweep_list,
  thrY,
  thrX_default = NULL,
  pre_calibrated = NULL,
  dir.exp = NULL,
  include = "",
  incl.cut = 0.8,
  n.cut = 1,
  pri.cut = 0,
  extract_mode = c("first", "all", "essential"),
  return_details = TRUE,
  Yvar = NULL,
  Xvars = NULL
)

Arguments

dat

Data frame containing the outcome and condition variables.

outcome

Character. Outcome variable name. Supports negation with tilde prefix (e.g., "~Y") following QCA package conventions.

conditions

Character vector. Names of condition variables.

sweep_list

Named list. Each element is a numeric vector of candidate thresholds for the corresponding X. Names must match conditions. Variables listed in pre_calibrated do not need a sweep_list entry. Every condition must receive a threshold from one of sweep_list, pre_calibrated, or thrX_default; a condition covered by none of these raises an error. To hold a condition at a fixed value while sweeping others, give it a length-one entry (e.g. list(SPV = 6:8, PRD = 7)).

thrY

Numeric. Threshold for Y (fixed).

thrX_default

Numeric or NULL. Threshold used to binarize any condition that appears in neither sweep_list nor pre_calibrated, mirroring the argument of the same name in ctSweepS. Default is NULL, in which case such a condition is an error rather than being silently dropped.

pre_calibrated

Character vector or NULL. Names of condition variables that have been pre-calibrated (e.g., via QCA::calibrate()) and should be passed through to QCA::truthTable() without binarization. These variables must contain values in the [0, 1] range. Variables not listed here will be binarized using sweep_list thresholds as usual. Default is NULL (all variables binarized). Variables listed here are never swept: they are used as they are. To sweep a variable that holds membership scores, leave it out of this argument; it is then binarized at the thresholds you supply (for example 0.3, 0.5, 0.7), like any other numeric variable.

dir.exp

Directional expectations for minimize. If NULL (default), no directional expectations are applied. To compute the intermediate solution, specify a numeric vector (1, 0, or -1 for each condition). Example: dir.exp = c(1, 1, 1) for three conditions all expected to contribute positively.

include

Inclusion rule for minimize. "" (default, QCA compatible) computes the complex solution without logical remainders. Use "?" to include logical remainders for parsimonious (with dir.exp = NULL) or intermediate solutions (with dir.exp specified).

incl.cut

Consistency cutoff for truthTable.

n.cut

Frequency cutoff for truthTable.

pri.cut

PRI cutoff for minimize.

extract_mode

Character. How to handle multiple solutions: "first" (default), "all", or "essential". See qca_extract for details.

return_details

Logical. If TRUE (default), returns both summary and detailed objects for use with generate_report().

Yvar

Deprecated. Use outcome instead.

Xvars

Deprecated. Use conditions instead.

Value

If return_details = FALSE, a data frame with columns:

  • combo_id — index of the threshold combination

  • threshold — character string summarizing thresholds, e.g. "X1=6, X2=7, X3=7"

  • expression — minimized solution expression

  • inclS — solution consistency

  • covS — solution coverage

  • (additional columns depending on extract_mode)

If return_details = TRUE, a list with:

  • summary — the data frame above

  • details — per-combination list of combo_id, thrX_vec, truth_table, solution

Note that return_details changes the type of the returned object, not just its contents: with TRUE the summary table is at result$summary, whereas with FALSE the summary table is the returned object and result$summary is NULL. Code intended to work under both settings should branch on inherits(result, "data.frame") (or simply always pass return_details = TRUE) rather than assuming result$summary exists.

Examples

# Load sample data
data(sample_data)

# === Three Types of QCA Solutions ===

# Quick demonstration with 2 conditions
sweep_list <- list(X1 = 7, X2 = 7)

# 1. Complex Solution (default, QCA compatible)
result_comp <- ctSweepM(
  dat = sample_data,
  outcome = "Y",
  conditions = c("X1", "X2"),
  sweep_list = sweep_list,
  thrY = 7
  # include = "" (default), dir.exp = NULL (default)
)
head(result_comp$summary)
#>   combo_id  threshold expression inclS      covS n_solutions
#> 1        1 X1=7, X2=7      X1*X2     1 0.3030303           1

# 2. Parsimonious Solution (include = "?")
result_pars <- ctSweepM(
  dat = sample_data,
  outcome = "Y",
  conditions = c("X1", "X2"),
  sweep_list = sweep_list,
  thrY = 7,
  include = "?"  # Include logical remainders
)
head(result_pars$summary)
#>   combo_id  threshold expression inclS      covS n_solutions
#> 1        1 X1=7, X2=7      X1*X2     1 0.3030303           1

# 3. Intermediate Solution (include = "?" + dir.exp)
result_int <- ctSweepM(
  dat = sample_data,
  outcome = "Y",
  conditions = c("X1", "X2"),
  sweep_list = sweep_list,
  thrY = 7,
  include = "?",
  dir.exp = c(1, 1)  # Positive expectations
)
head(result_int$summary)
#>   combo_id  threshold expression inclS      covS n_solutions
#> 1        1 X1=7, X2=7      X1*X2     1 0.3030303           1

# === Threshold Sweep Example ===

# Using 2 conditions and 2 threshold levels
sweep_list <- list(
  X1 = 6:7,
  X2 = 6:7
)

# Run multiple condition threshold sweep (complex solutions by default)
result_quick <- ctSweepM(
  dat = sample_data,
  outcome = "Y",
  conditions = c("X1", "X2"),
  sweep_list = sweep_list,
  thrY = 7
)
head(result_quick$summary)
#>   combo_id  threshold  expression     inclS      covS n_solutions
#> 1        1 X1=6, X2=6       X1*X2 0.8333333 0.4545455           1
#> 2        2 X1=7, X2=6       X1*X2 1.0000000 0.3939394           1
#> 3        3 X1=6, X2=7 No solution        NA        NA           0
#> 4        4 X1=7, X2=7       X1*X2 1.0000000 0.3030303           1

# Run with negated outcome (~Y)
result_neg <- ctSweepM(
  dat = sample_data,
  outcome = "~Y",
  conditions = c("X1", "X2"),
  sweep_list = sweep_list,
  thrY = 7
)
head(result_neg$summary)
#>   combo_id  threshold  expression inclS covS n_solutions
#> 1        1 X1=6, X2=6 No solution    NA   NA           0
#> 2        2 X1=7, X2=6 No solution    NA   NA           0
#> 3        3 X1=6, X2=7 No solution    NA   NA           0
#> 4        4 X1=7, X2=7 No solution    NA   NA           0

# \donttest{
# Full multi-condition analysis (27 combinations)
sweep_list_full <- list(
  X1 = 6:8,
  X2 = 6:8,
  X3 = 6:8
)

result_full <- ctSweepM(
  dat = sample_data,
  outcome = "Y",
  conditions = c("X1", "X2", "X3"),
  sweep_list = sweep_list_full,
  thrY = 7
)
head(result_full$summary)
#>   combo_id        threshold expression     inclS      covS n_solutions
#> 1        1 X1=6, X2=6, X3=6      X1*X2 0.8333333 0.4545455           1
#> 2        2 X1=7, X2=6, X3=6      X1*X2 1.0000000 0.3939394           1
#> 3        3 X1=8, X2=6, X3=6      X1*X2 1.0000000 0.3030303           1
#> 4        4 X1=6, X2=7, X3=6  X1*X2*~X3 0.8181818 0.2727273           1
#> 5        5 X1=7, X2=7, X3=6      X1*X2 1.0000000 0.3030303           1
#> 6        6 X1=8, X2=7, X3=6      X1*X2 1.0000000 0.2121212           1
# }