parameterEstimates_boot

library(semboottools)
library(lavaan)

Function Syntax

parameterEstimates_boot(object,
                        level = .95,
                        standardized = FALSE,
                        boot_org_ratio = FALSE,
                        boot_ci_type = c("perc", "bc", "bca.simple"),
                        save_boot_est = TRUE,
                        boot_pvalue = TRUE,
                        boot_pvalue_min_size = 1000,
                        ...)

Arguments

Argument Description
object A model fitted by lavaan.
level Confidence level for the confidence intervals. For example, .95 gives 95% confidence intervals.
standardized Whether to return standardized estimates. Same as in lavaan::parameterEstimates(). You can use "std.all", "std.lv", etc. For detailed standardized results with CIs, use standardizedSolution_boot() instead.
boot_org_ratio Whether to calculate how wide the bootstrap confidence interval is compared to the original confidence interval (from delta method). Useful to compare the two methods.
boot_ci_type Method for forming bootstrap confidence intervals. "perc" gives percentile intervals; "bc" and "bca.simple" give bias-corrected intervals.
save_boot_est Whether to save the bootstrap estimates in the result. Saved in attributes boot_est_ustd (free parameters) and boot_def (user-defined parameters) if TRUE.
boot_pvalue Whether to compute asymmetric p-values based on bootstrap results. Only available when percentile confidence intervals are used.
boot_pvalue_min_size Minimum number of valid bootstrap samples needed to compute asymmetric p-values. If fewer samples are available, p-values will not be computed and will be shown as NA.
... Additional arguments passed to lavaan::parameterEstimates().

Example

Data and Model

# Set seed for reproducibility
set.seed(1234)

# Generate data
n <- 1000
x <- runif(n) - 0.5
m <- 0.20 * x + rnorm(n)
y <- 0.17 * m + rnorm(n)
dat <- data.frame(x, y, m)

# Specify mediation model in lavaan syntax
mod <- '
  m ~ a * x
  y ~ b * m + cp * x
  ab := a * b
  total := a * b + cp
'

Basic usage: default settings

# Ensure bootstrap estimates are stored
fit <- sem(mod, data = dat, fixed.x = FALSE)
fit <- store_boot(fit) 
est_boot <- parameterEstimates_boot(fit)
print(est_boot)
#> 
#> Bootstrapping:
#>                                     
#>  Valid Bootstrap Samples: 1000      
#>  Level of Confidence:     95.0%     
#>  CI Type:                 Percentile
#>  P-Value:                 Asymmetric
#> 
#> Parameter Estimates Settings:
#>                                              
#>  Standard errors:                  Standard  
#>  Information:                      Expected  
#>  Information saturated (h1) model: Structured
#> 
#> Regressions:
#>                Estimate    SE     p  CI.Lo CI.Up   bSE    bp bCI.Lo bCI.Up
#>  m ~                                                                      
#>   x (a)           0.089 0.103 0.386 -0.113 0.291 0.108 0.446 -0.121  0.287
#>  y ~                                                                      
#>   m (b)           0.192 0.034 0.000  0.125 0.260 0.037 0.000  0.121  0.265
#>   x (cp)         -0.018 0.112 0.871 -0.238 0.202 0.112 0.868 -0.240  0.214
#> 
#> Variances:
#>                Estimate    SE     p  CI.Lo CI.Up   bSE    bp bCI.Lo bCI.Up
#>   .m              0.898 0.040 0.000  0.819 0.977 0.041 0.000  0.820  0.983
#>   .y              1.065 0.048 0.000  0.972 1.159 0.045 0.000  0.973  1.151
#>    x              0.085 0.004 0.000  0.077 0.092 0.002 0.000  0.080  0.089
#> 
#> Defined Parameters:
#>                Estimate    SE     p  CI.Lo CI.Up   bSE    bp bCI.Lo bCI.Up
#>  ab (ab)          0.017 0.020 0.392 -0.022 0.056 0.021 0.446 -0.025  0.059
#>  total (total)   -0.001 0.114 0.993 -0.224 0.222 0.115 0.994 -0.229  0.225
#> 
#> Footnote:
#> - SE: Original standard errors.
#> - p: Original p-values.
#> - CI.Lo, CI.Up: Original confidence intervals.
#> - bSE: Bootstrap standard errors.
#> - bCI.Lo, bCI.Up: Bootstrap confidence intervals.
#> - bp: Bootstrap p-values.

parameterEstimates_boot(): Different Options


# Change confidence level to 99%
est_boot <- parameterEstimates_boot(fit, level = 0.99)
# Use bias-corrected (BC) bootstrap confidence intervals
est_boot <- parameterEstimates_boot(fit, boot_ci_type = "bc")
# Turn off asymmetric bootstrap p-values
est_boot <- parameterEstimates_boot(fit, boot_pvalue = FALSE)
# Do not save bootstrap estimates (for memory saving)
est_boot <- parameterEstimates_boot(fit, save_boot_est = FALSE)
# Compute and display bootstrap-to-original CI ratio
est_boot <- parameterEstimates_boot(fit, boot_org_ratio = TRUE)
# Combine options: BC CI, 99% level, no p-values
est_boot <- parameterEstimates_boot(fit,
                                    level = 0.99,
                                    boot_ci_type = "bc",
                                    boot_pvalue = FALSE)