add_local_dependence {latentFactoR}

Adds Local Dependence to simulate_factors Data

Description

Adds local dependence to simulated data from simulate_factors. See examples to get started

Usage

add_local_dependence(
  lf_object,
  method = c("correlate_residuals", "minor_factors", "threshold_shifts"),
  proportion_LD,
  proportion_LD_range = NULL,
  add_residuals = NULL,
  add_residuals_range = NULL,
  allow_multiple = FALSE
)

Arguments

lf_object

Data object from simulate_factors

method

Character (length = 1). Method to generate local dependence between variables. Only "correlate_residuals" at the moment. Future developments will include minor factor and threshold-shift methods. Description of methods:

• "correlate_residuals" — Adds residuals directly to the population correlation matrix prior to data generation (uses population correlation matrix from simulate_factors)

• "minor_factors" — Coming soon...

• "threshold_shifts" — Coming soon...

proportion_LD

Numeric (length = 1 or factors). Proportion of variables that should be locally dependent across all or each factor. Accepts number of locally dependent values as well

proportion_LD_range

Numeric (length = 2). Range of proportion of variables that are randomly selected from a random uniform distribution. Accepts number of locally dependent values as well. Defaults to NULL

add_residuals

Numeric (length = 1, factors, or total number of locally dependent variables). Amount of residual to add to the population correlation matrix between two variables. Only used when method = "correlated_residuals". Magnitudes are drawn from a random uniform distribution using +/- 0.05 of value input. Can also be specified directly (same length as total number of locally dependent variables). General effect sizes range from small (0.20), moderate (0.30), to large (0.40)

add_residuals_range

Numeric (length = 2). Range of the residuals to add to the correlation matrix are randomly selected from a random uniform distribution. Defaults to NULL

allow_multiple

Boolean. Whether a variable should be allowed to be locally dependent with more than one other variable. Defaults to FALSE. Set to TRUE for more complex locally dependence patterns

Value

Returns a list containing:

data

Simulated data from the specified factor model

population_correlation

Population correlation matrix with local dependence added

original_correlation

Original population correlation matrix before local dependence was added

correlated_residuals

A data frame with the first two columns specifying the variables that are locally dependent and the third column specifying the magnitude of the added residual for each locally dependent pair

original_results

Original lf_object input into function

Author(s)

Alexander P. Christensen <[e-mail]>, Hudson Golino <[e-mail]>, Luis Eduardo Garrido <[e-mail]>

References

Christensen, A. P., Garrido, L. E., & Golino, H. (2023). Unique variable analysis: A network psychometrics method to detect local dependence. Multivariate Behavioral Research, 1–18.

Examples

# Generate factor data
two_factor <- simulate_factors(
  factors = 2, # factors = 2
  variables = 6, # variables per factor = 6
  loadings = 0.55, # loadings between = 0.45 to 0.65
  cross_loadings = 0.05, # cross-loadings N(0, 0.05)
  correlations = 0.30, # correlation between factors = 0.30
  sample_size = 1000 # number of cases = 1000
)

# Add local dependence
two_factor_LD <- add_local_dependence(
  lf_object = two_factor,
  proportion_LD = 0.25,
  add_residuals = 0.20,
  allow_multiple = FALSE
)

# Randomly vary proportions
two_factor_LD <- add_local_dependence(
  lf_object = two_factor,
  proportion_LD_range = c(0.10, 0.50),
  add_residuals = 0.20,
  allow_multiple = FALSE
)

# Randomly vary residuals
two_factor_LD <- add_local_dependence(
  lf_object = two_factor,
  proportion_LD = 0.25,
  add_residuals_range = c(0.20, 0.40),
  allow_multiple = FALSE
)

# Randomly vary proportions, residuals, and allow multiple
two_factor_LD <- add_local_dependence(
  lf_object = two_factor,
  proportion_LD_range = c(0.10, 0.50),
  add_residuals_range = c(0.20, 0.40),
  allow_multiple = TRUE
)