sim.questionnaire.multi: Simulate multiple questionnaires with custom linear model...

In ccamp83/mu: R package for general statistical analysis of univariate and multivariate data

Simulate multiple questionnaires with custom linear model relationships

Description

Simulate multiple questionnaires with custom linear model relationships

Usage

sim.questionnaire.multi(
  N = 100,
  questionnaires,
  linear_model = NULL,
  linear_models = NULL,
  correlations = NULL,
  seed = NULL,
  debug = FALSE,
  max_retries = 20,
  verbose = TRUE
)

Arguments

N	Number of participants
questionnaires	List of questionnaire specifications. Each element should be a list with: - n_items: number of items - likert: vector of possible Likert scale values - total_mean: mean of total score distribution - total_sd: standard deviation of desired total score distribution - total_min: minimum possible total score (optional, will be calculated if not provided) - total_max: maximum possible total score (optional, will be calculated if not provided)
linear_model	(For backward compatibility or single models) A list specifying a single linear model.
linear_models	(Recommended) A list of lists, where each inner list specifies a linear model to be applied sequentially. Use this for mediation or other path models.
correlations	Optional correlation matrix between questionnaire total scores
seed	Random seed for reproducibility
debug	If TRUE, returns additional diagnostic information about the linear model
max_retries	The maximum number of attempts to generate a dataset that matches the desired coefficient signs.
verbose	If TRUE, prints messages about retry attempts for matching coefficient signs.

Value

Data frame with item-level responses and total scores for all questionnaires

Examples

# Example 1: Single moderation model (backward compatible)
set.seed(456)
simdat_single_mod <- sim.questionnaire.multi(
   N = 200,
   questionnaires = list(
     Anxiety = list(n_items = 10, likert = 1:5, total_mean = 30, total_sd = 6),
     Depression = list(n_items = 12, likert = 1:5, total_mean = 28, total_sd = 7),
     Stress = list(n_items = 8, likert = 1:7, total_mean = 15, total_sd = 4)
   ),
   linear_model = list(
     formula = "Stress ~ Anxiety + Depression + Anxiety*Depression",
     desired_coefficient_signs = list(Anxiety = 1, Depression = -1, "Anxiety:Depression" = -1)
   )
)
summary(lm(Stress_total ~ Anxiety_total * Depression_total, data = simdat_single_mod))

# Example 2: Mediation model (IV -> M -> DV)
set.seed(111)
simdat_mediation <- sim.questionnaire.multi(
  N = 150,
  questionnaires = list(
    Workload = list(n_items = 10, likert = 1:5, total_mean = 30, total_sd = 5),
    Burnout = list(n_items = 8, likert = 1:7, total_mean = 35, total_sd = 6),
    Satisfaction = list(n_items = 6, likert = 0:10, total_mean = 40, total_sd = 8)
  ),
  linear_models = list(
    list(formula = "Burnout ~ Workload", desired_coefficient_signs = list(Workload = 1)),
    list(formula = "Satisfaction ~ Burnout + Workload", desired_coefficient_signs = list(Burnout = -1, Workload = -1))
  ),
  max_retries = 10
)
summary(lm(Burnout_total ~ Workload_total, data = simdat_mediation))
summary(lm(Satisfaction_total ~ Burnout_total + Workload_total, data = simdat_mediation))

ccamp83/mu documentation built on July 4, 2025, 6:20 p.m.