Multiple Mediation Models LEEP
ALI
6/30/2021
library(lavaan)## Warning: package 'lavaan' was built under R version 4.0.4## This is lavaan 0.6-7## lavaan is BETA software! Please report any bugs.library(dplyr)##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union# Libraries
#library(rlang)
library(tidyverse)## Warning: package 'tidyverse' was built under R version 4.0.5## -- Attaching packages --------------------------------------- tidyverse 1.3.0 --## v ggplot2 3.3.2 v purrr 0.3.4
## v tibble 3.0.3 v stringr 1.4.0
## v tidyr 1.1.2 v forcats 0.5.0
## v readr 1.4.0## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x dplyr::filter() masks stats::filter()
## x dplyr::lag() masks stats::lag()library(dplyr)
library(knitr)
library(lavaan)
library(psych)##
## Attaching package: 'psych'## The following objects are masked from 'package:ggplot2':
##
## %+%, alpha## The following object is masked from 'package:lavaan':
##
## cor2covlibrary(MBESS)## Warning: package 'MBESS' was built under R version 4.0.4##
## Attaching package: 'MBESS'## The following object is masked from 'package:psych':
##
## cor2cov## The following object is masked from 'package:lavaan':
##
## cor2covlibrary(stargazer)##
## Please cite as:## Hlavac, Marek (2018). stargazer: Well-Formatted Regression and Summary Statistics Tables.## R package version 5.2.2. https://CRAN.R-project.org/package=stargazerlibrary(mediation)## Warning: package 'mediation' was built under R version 4.0.4## Loading required package: MASS##
## Attaching package: 'MASS'## The following object is masked from 'package:dplyr':
##
## select## Loading required package: Matrix##
## Attaching package: 'Matrix'## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpack## Loading required package: mvtnorm## Loading required package: sandwich## mediation: Causal Mediation Analysis
## Version: 4.5.0##
## Attaching package: 'mediation'## The following object is masked from 'package:psych':
##
## mediatelibrary(multilevel)## Warning: package 'multilevel' was built under R version 4.0.4## Loading required package: nlme##
## Attaching package: 'nlme'## The following object is masked from 'package:dplyr':
##
## collapselibrary(car)## Loading required package: carData## Registered S3 methods overwritten by 'car':
## method from
## influence.merMod lme4
## cooks.distance.influence.merMod lme4
## dfbeta.influence.merMod lme4
## dfbetas.influence.merMod lme4##
## Attaching package: 'car'## The following object is masked from 'package:psych':
##
## logit## The following object is masked from 'package:purrr':
##
## some## The following object is masked from 'package:dplyr':
##
## recodelibrary(gtsummary)## Warning: package 'gtsummary' was built under R version 4.0.4##
## Attaching package: 'gtsummary'## The following object is masked from 'package:MASS':
##
## selectlibrary(flextable)## Warning: package 'flextable' was built under R version 4.0.4##
## Attaching package: 'flextable'## The following object is masked from 'package:gtsummary':
##
## as_flextable## The following object is masked from 'package:purrr':
##
## composelibrary(skimr)## Warning: package 'skimr' was built under R version 4.0.5library(DiagrammeR )
library(processR)## Warning: package 'processR' was built under R version 4.0.4## This version of bslib is designed to work with shiny version 1.5.0.9007 or higher.##
## Attaching package: 'processR'## The following objects are masked from 'package:car':
##
## densityPlot, qqPlot## The following object is masked from 'package:psych':
##
## corPlotlibrary(lavaan)
library(lavaan)
library(semPlot)## Warning: package 'semPlot' was built under R version 4.0.4library(OpenMx)## Warning: package 'OpenMx' was built under R version 4.0.4##
## Attaching package: 'OpenMx'## The following objects are masked from 'package:Matrix':
##
## %&%, expm## The following object is masked from 'package:psych':
##
## tr## The following object is masked from 'package:lavaan':
##
## vechlibrary(tidyverse)
library(knitr)
library(kableExtra)##
## Attaching package: 'kableExtra'## The following objects are masked from 'package:flextable':
##
## as_image, footnote## The following object is masked from 'package:dplyr':
##
## group_rowslibrary(GGally)## Registered S3 method overwritten by 'GGally':
## method from
## +.gg ggplot2https://www.rpubs.com/cardiomoon/481347 https://ademos.people.uic.edu/Chapter15.html#12_mediation
LEEP$seizure<-Recode(LEEP$Seizure_Frequency, recodes="'I have not had a seizure in the past 12 months'=0;else=1",as.numeric=T)
LEEP$PHQ_2<-Recode(LEEP$PHQ_2, recodes="'Depression'=1;'No Depression'=0",as.numeric=T)
LEEP$severe<-as.numeric(LEEP$severe)LEEP<-LEEP %>%
rename(QOLIE=means)
# including a contrast in the model
multipleMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 26 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## QOLIE ~
## severe (b1) 0.155 0.024 6.335 0.000
## seizure (b2) 0.557 0.079 7.017 0.000
## PHQ_2 (c) 0.341 0.071 4.771 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .QOLIE 0.297 0.026 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 0.129 0.035 3.676 0.000
## indirect2 0.145 0.038 3.824 0.000
## total 0.615 0.084 7.301 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 8)
constrainedMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 1684.5 1716.5 0.0000
## constrFit 1 1682.6 1711.1 0.1032 0.10317 1 0.7481contrastsMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 26 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## QOLIE ~
## severe (b1) 0.155 0.024 6.335 0.000
## seizure (b2) 0.557 0.079 7.017 0.000
## PHQ_2 (c) 0.341 0.071 4.771 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .QOLIE 0.297 0.026 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 0.129 0.035 3.676 0.000
## indirect2 0.145 0.038 3.824 0.000
## contrast 0.015 0.048 0.321 0.748
## total 0.615 0.084 7.301 0.000# including a contrast in the model
multipleMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 42 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## PCS_m ~
## severe (b1) -2.321 0.387 -5.998 0.000
## seizure (b2) -1.582 1.257 -1.258 0.208
## PHQ_2 (c) -4.973 1.131 -4.396 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .PCS_m 74.440 6.516 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -1.939 0.538 -3.607 0.000
## indirect2 -0.411 0.339 -1.213 0.225
## total -7.323 1.171 -6.256 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 8)
constrainedMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 3126.2 3158.2 0.0000
## constrFit 1 3130.2 3158.7 6.0528 6.0528 1 0.01388 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1contrastsMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 42 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## PCS_m ~
## severe (b1) -2.321 0.387 -5.998 0.000
## seizure (b2) -1.582 1.257 -1.258 0.208
## PHQ_2 (c) -4.973 1.131 -4.396 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .PCS_m 74.440 6.516 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -1.939 0.538 -3.607 0.000
## indirect2 -0.411 0.339 -1.213 0.225
## contrast 1.528 0.674 2.267 0.023
## total -7.323 1.171 -6.256 0.000# including a contrast in the model
multipleMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 44 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## MCS_m ~
## severe (b1) -1.274 0.397 -3.210 0.001
## seizure (b2) -1.180 1.290 -0.915 0.360
## PHQ_2 (c) -9.665 1.160 -8.329 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .MCS_m 78.321 6.856 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -1.064 0.407 -2.616 0.009
## indirect2 -0.306 0.342 -0.897 0.370
## total -11.036 1.131 -9.754 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 10)
constrainedMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 3139.4 3171.5 0.0000
## constrFit 1 3139.1 3167.6 1.7096 1.7096 1 0.191contrastsMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * PHQ_2
severe ~ a1 * PHQ_2
seizure ~ a2 * PHQ_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 44 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## MCS_m ~
## severe (b1) -1.274 0.397 -3.210 0.001
## seizure (b2) -1.180 1.290 -0.915 0.360
## PHQ_2 (c) -9.665 1.160 -8.329 0.000
## severe ~
## PHQ_2 (a1) 0.835 0.185 4.514 0.000
## seizure ~
## PHQ_2 (a2) 0.260 0.057 4.560 0.000
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.258 0.045 5.699 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .MCS_m 78.321 6.856 11.424 0.000
## .severe 2.220 0.194 11.424 0.000
## .seizure 0.210 0.018 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -1.064 0.407 -2.616 0.009
## indirect2 -0.306 0.342 -0.897 0.370
## contrast 0.758 0.595 1.274 0.203
## total -11.036 1.131 -9.754 0.000# including a contrast in the model
multipleMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 24 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## QOLIE ~
## severe (b1) 0.150 0.025 6.101 0.000
## seizure (b2) 0.588 0.078 7.500 0.000
## GAD2_2 (c) 0.343 0.071 4.844 0.000
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .QOLIE 0.296 0.026 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 0.129 0.035 3.686 0.000
## indirect2 0.115 0.037 3.060 0.002
## total 0.587 0.085 6.874 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 8)
estimatesTable(fit)## Variables Predictors label B SE z p ß
## 1 QOLIE severe b1 0.15 0.02 6.10 < 0.001 0.31
## 2 QOLIE seizure b2 0.59 0.08 7.50 < 0.001 0.38
## 3 QOLIE GAD2_2 c 0.34 0.07 4.84 < 0.001 0.23
## 4 severe GAD2_2 a1 0.86 0.19 4.63 < 0.001 0.28
## 5 seizure GAD2_2 a2 0.19 0.06 3.35 0.001 0.20constrainedMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 1689.0 1721.0 0.0000
## constrFit 1 1687.1 1715.6 0.0921 0.092055 1 0.7616contrastsMediation <- '
QOLIE ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 24 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## QOLIE ~
## severe (b1) 0.150 0.025 6.101 0.000
## seizure (b2) 0.588 0.078 7.500 0.000
## GAD2_2 (c) 0.343 0.071 4.844 0.000
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .QOLIE 0.296 0.026 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 0.129 0.035 3.686 0.000
## indirect2 0.115 0.037 3.060 0.002
## contrast -0.014 0.046 -0.304 0.761
## total 0.587 0.085 6.874 0.000# including a contrast in the model
multipleMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 42 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## PCS_m ~
## severe (b1) -2.337 0.396 -5.910 0.000
## seizure (b2) -2.158 1.261 -1.712 0.087
## GAD2_2 (c) -3.849 1.140 -3.377 0.001
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .PCS_m 76.603 6.706 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -2.003 0.550 -3.642 0.000
## indirect2 -0.420 0.276 -1.524 0.127
## total -6.272 1.198 -5.238 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 8)
constrainedMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 3138.8 3170.9 0.0000
## constrFit 1 3144.2 3172.7 7.4527 7.4527 1 0.006334 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1contrastsMediation <- '
PCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 42 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## PCS_m ~
## severe (b1) -2.337 0.396 -5.910 0.000
## seizure (b2) -2.158 1.261 -1.712 0.087
## GAD2_2 (c) -3.849 1.140 -3.377 0.001
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .PCS_m 76.603 6.706 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -2.003 0.550 -3.642 0.000
## indirect2 -0.420 0.276 -1.524 0.127
## contrast 1.582 0.633 2.499 0.012
## total -6.272 1.198 -5.238 0.000# including a contrast in the model
multipleMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = multipleMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 44 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## MCS_m ~
## severe (b1) -1.114 0.395 -2.822 0.005
## seizure (b2) -2.013 1.258 -1.601 0.109
## GAD2_2 (c) -10.068 1.137 -8.854 0.000
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .MCS_m 76.238 6.674 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -0.954 0.396 -2.409 0.016
## indirect2 -0.392 0.272 -1.445 0.149
## total -11.414 1.122 -10.176 0.000boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")
semPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'tree')
ly<-matrix(c(1, 1,
0.33, 2,
0.33, 0,
-1, 1,
-1, -1), ncol=2, byrow=TRUE)
semPaths(fit, whatLabels = 'std', nCharNodes = 0, layout = ly,
edge.label.cex = 1.2, residuals=FALSE, sizeMan=8)
semPaths(fit, what = "std",layout=ly, residuals = FALSE,
nCharNodes = 0, edge.label.cex = 1.2, sizeMan = 10)
constrainedMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
total := c + (a1 * b1) + (a2 * b2)
# covariances
severe ~~ seizure
# constrain
indirect1 == indirect2
'
noConstrFit <- sem(model = multipleMediation, data = LEEP)
constrFit <- sem(model = constrainedMediation, data = LEEP)
anova(noConstrFit, constrFit)## Chi-Squared Difference Test
##
## Df AIC BIC Chisq Chisq diff Df diff Pr(>Chisq)
## noConstrFit 0 3137.5 3169.6 0.0000
## constrFit 1 3136.7 3165.2 1.1665 1.1665 1 0.2801contrastsMediation <- '
MCS_m ~ b1 * severe + b2 * seizure + c * GAD2_2
severe ~ a1 * GAD2_2
seizure ~ a2 * GAD2_2
indirect1 := a1 * b1
indirect2 := a2 * b2
contrast := indirect2-indirect1
total := c + (a1 * b1) + (a2 * b2)
severe ~~ seizure
'
fit <- sem(model = contrastsMediation, data = LEEP)
summary(fit)## lavaan 0.6-7 ended normally after 44 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 9
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Standard
## Information Expected
## Information saturated (h1) model Structured
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## MCS_m ~
## severe (b1) -1.114 0.395 -2.822 0.005
## seizure (b2) -2.013 1.258 -1.601 0.109
## GAD2_2 (c) -10.068 1.137 -8.854 0.000
## severe ~
## GAD2_2 (a1) 0.857 0.185 4.626 0.000
## seizure ~
## GAD2_2 (a2) 0.195 0.058 3.352 0.001
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.270 0.046 5.863 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .MCS_m 76.238 6.674 11.424 0.000
## .severe 2.212 0.194 11.424 0.000
## .seizure 0.218 0.019 11.424 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -0.954 0.396 -2.409 0.016
## indirect2 -0.392 0.272 -1.445 0.149
## contrast 0.562 0.526 1.069 0.285
## total -11.414 1.122 -10.176 0.000LEEP$seizure<-Recode(LEEP$Seizure_Frequency, recodes="'I have not had a seizure in the past 12 months'=0;else=1",as.numeric=T)
myModel <- '
PCS_m ~ b1 * severe + b2 * seizure + c1 * PHQ_2 + c2 * GAD2_2 + c3 * RSES
severe ~ a1 * PHQ_2 + a2 * GAD2_2 + a5 * RSES
seizure ~ a3 * PHQ_2 + a4 * GAD2_2 + a6 * RSES
#indirect effects
indirect1 := a1 * b1
indirect2 := a3 * b2
indirect3 := a2 * b1
indirect4 := a4 * b2
# contrasts
con1 := a1 * b1 - a3 * b2
con2 := a2 * b1 - a4 * b2
con3 := (a1-a2) * b1
con4 := (a3-a4) * b2
# total effect
total1 := c1 + (a1 * b1) + (a3 * b2)
total2 := c2 + (a2 * b1) + (a4 * b2)
total3 := c3 + (a5 * b1) + (a6 * b2)
# covariates
severe ~~ seizure
'
fit <- sem(myModel,
data=LEEP,
se = "bootstrap",
bootstrap = 50)
summary(fit)## lavaan 0.6-7 ended normally after 61 iterations
##
## Estimator ML
## Optimization method NLMINB
## Number of free parameters 15
##
## Number of observations 261
##
## Model Test User Model:
##
## Test statistic 0.000
## Degrees of freedom 0
##
## Parameter Estimates:
##
## Standard errors Bootstrap
## Number of requested bootstrap draws 50
## Number of successful bootstrap draws 50
##
## Regressions:
## Estimate Std.Err z-value P(>|z|)
## PCS_m ~
## severe (b1) -2.224 0.392 -5.680 0.000
## seizure (b2) -1.574 1.394 -1.129 0.259
## PHQ_2 (c1) -4.064 1.057 -3.844 0.000
## GAD2_2 (c2) -1.990 1.286 -1.548 0.122
## RSES (c3) -0.024 0.197 -0.120 0.905
## severe ~
## PHQ_2 (a1) 0.577 0.222 2.600 0.009
## GAD2_2 (a2) 0.622 0.226 2.755 0.006
## RSES (a5) 0.045 0.030 1.520 0.129
## seizure ~
## PHQ_2 (a3) 0.207 0.065 3.178 0.001
## GAD2_2 (a4) 0.072 0.079 0.907 0.365
## RSES (a6) -0.011 0.009 -1.249 0.212
##
## Covariances:
## Estimate Std.Err z-value P(>|z|)
## .severe ~~
## .seizure 0.254 0.039 6.499 0.000
##
## Variances:
## Estimate Std.Err z-value P(>|z|)
## .PCS_m 73.754 7.837 9.411 0.000
## .severe 2.138 0.141 15.132 0.000
## .seizure 0.208 0.011 18.989 0.000
##
## Defined Parameters:
## Estimate Std.Err z-value P(>|z|)
## indirect1 -1.283 0.569 -2.255 0.024
## indirect2 -0.327 0.341 -0.959 0.338
## indirect3 -1.383 0.551 -2.510 0.012
## indirect4 -0.113 0.158 -0.714 0.475
## con1 -0.956 0.678 -1.410 0.159
## con2 -1.270 0.559 -2.271 0.023
## con3 0.100 0.836 0.119 0.905
## con4 -0.214 0.373 -0.574 0.566
## total1 -5.674 1.217 -4.662 0.000
## total2 -3.485 1.407 -2.478 0.013
## total3 -0.106 0.180 -0.587 0.557boot.fit <- parameterEstimates(fit, boot.ci.type="bca.simple")## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, adj.alpha): extreme order statistics used as endpointssemPaths(fit, what='std', nCharNodes=6, sizeMan=10,
edge.label.cex=1.25, curvePivot = F, fade=FALSE, layout = 'circle')