Visualization of Moderation effects by processR

The First Example

library(processR)
library(lavaan)
fit=lm(justify ~ frame*skeptic,data=disaster)

Conceptual Model

labels=list(X="frame",W="skeptic",Y="justify")
pmacroModel(1,labels=labels)

Statistical Model

statisticalDiagram(1,labels=labels)

Model Summary

modelsSummaryTable(list(fit),labels=labels)

	Consequent
	Y(justify)
Antecedent	Coef	SE	t	p
X(frame)	-0.562	0.218	-2.581	.011
W(skeptic)	0.105	0.038	2.756	.006
X:W(frame:skeptic)	0.201	0.055	3.640	<.001
Constant	2.452	0.149	16.449	<.001
Observations	211
R2	0.246
Adjusted R2	0.235
Residual SE	0.813 ( df = 207)
F statistic	F(3,207) = 22.543, p < .001

Visualization of Moderation Effect

condPlot(fit,rangemode=2,xpos=0.7,labels=c("Climate change(X=1)","Natural causes(X=0)"))

condPlot(fit,mode=2,xpos=0.6)

condPlot(fit,mode=3,rangemode=2,xpos=0.5,ypos=c(0,2))

Johnson-Neyman Plot

jnPlot(fit,plot=FALSE)

JOHNSON-NEYMAN INTERVAL 

When skeptic is OUTSIDE the interval [1.171, 3.934], the slope of
frame is p < .05.

Note: The range of observed values of skeptic is [1.000, 9.000]

Statistical Diagram

moderator=list(name="skeptic",site=list("c"))
model=tripleEquation(labels=labels,moderator=moderator)
cat(model)
semfit=sem(model=model,data=disaster)
summary(semfit)

justify~c1*frame+c2*skeptic+c3*frame:skeptic
skeptic ~ skeptic.mean*1
skeptic ~~ skeptic.var*skeptic
direct :=c1+c3*skeptic.mean
direct.below:=c1+c3*(skeptic.mean-sqrt(skeptic.var))
direct.above:=c1+c3*(skeptic.mean+sqrt(skeptic.var))
lavaan 0.6-3 ended normally after 24 iterations

  Optimization method                           NLMINB
  Number of free parameters                         12

  Number of observations                           211

  Estimator                                         ML
  Model Fit Test Statistic                     136.428
  Degrees of freedom                                 2
  P-value (Chi-square)                           0.000

Parameter Estimates:

  Information                                 Expected
  Information saturated (h1) model          Structured
  Standard Errors                             Standard

Regressions:
                   Estimate  Std.Err  z-value  P(>|z|)
  justify ~                                           
    frame     (c1)   -0.562    0.175   -3.211    0.001
    skeptic   (c2)    0.105    0.027    3.844    0.000
    frm:skptc (c3)    0.201    0.040    5.077    0.000

Covariances:
                   Estimate  Std.Err  z-value  P(>|z|)
  frame ~~                                            
    frame:skeptic     0.854    0.096    8.890    0.000

Intercepts:
                   Estimate  Std.Err  z-value  P(>|z|)
    skeptic (skp.)    3.378    0.140   24.196    0.000
   .justify           2.452    0.120   20.416    0.000
    frame             0.479    0.034   13.919    0.000
    frm:skp           1.637    0.152   10.771    0.000

Variances:
                   Estimate  Std.Err  z-value  P(>|z|)
    skeptic (skp.)    4.113    0.400   10.271    0.000
   .justify           0.648    0.063   10.271    0.000
    frame             0.250    0.024   10.271    0.000
    frm:skp           4.877    0.475   10.271    0.000

Defined Parameters:
                   Estimate  Std.Err  z-value  P(>|z|)
    direct            0.117    0.114    1.023    0.306
    direct.below     -0.291    0.142   -2.045    0.041
    direct.above      0.525    0.140    3.743    0.000

statisticalDiagram(1,labels=labels,fit=semfit,whatLabel = "est")

The Second Example

labels=list(X="negemot",Y="govact",W="age",C1="posemot",C2="ideology",C3="sex")
moderator=list(name="age",site=list("c"))
covar=list(name=c("posemot","ideology","sex"),site=list("Y","Y","Y"))

Conceptual Model

pmacroModel(1,labels=labels,covar=covar)

Statistical Model

statisticalDiagram(1,labels=labels,covar=covar)

Model Summary

fit=lm(govact~negemot*age+ideology+sex,data=glbwarm)
modelsSummaryTable(list(fit),labels=labels)

	Consequent
	Y(govact)
Antecedent	Coef	SE	t	p
X(negemot)	0.114	0.082	1.388	.165
W(age)	-0.024	0.006	-4.044	<.001
C2(ideology)	-0.211	0.027	-7.883	<.001
C3(sex)	-0.016	0.076	-0.212	.832
X:W(negemot:age)	0.006	0.002	4.145	<.001
Constant	5.132	0.334	15.371	<.001
Observations	815
R2	0.400
Adjusted R2	0.397
Residual SE	1.057 ( df = 809)
F statistic	F(5,809) = 108.033, p < .001

Visualization of Moderation

condPlot(fit,xmode=2,pred.values=c(30,70),hjust=c(-0.1,-0.1,1.1))

condPlot(fit,xmode=2,mode=2,pred.values=c(30,50,70),xpos=0.2)

condPlot(fit,xmode=2,modx.values=2:5,mode=3,xpos=0.6,hjust=c(-0.1,-0.1,-0.1,1.1))

Johnson-Neyman Plot

jnPlot(fit,plot=FALSE)

JOHNSON-NEYMAN INTERVAL 

When age is OUTSIDE the interval [-80.462, 5.112], the slope of
negemot is p < .05.

Note: The range of observed values of age is [17.000, 87.000]

Make lavaan syntax

model=tripleEquation(labels=labels,moderator=moderator,covar=covar)
cat(model)

govact ~ c1*negemot+c2*age+c3*negemot:age + g1*posemot + g2*ideology + g3*sex
age ~ age.mean*1
age ~~ age.var*age
direct :=c1+c3*age.mean
direct.below:=c1+c3*(age.mean-sqrt(age.var))
direct.above:=c1+c3*(age.mean+sqrt(age.var))

Fit Structural Equation Models

semfit=sem(model=model,data=glbwarm)
summary(semfit)

lavaan 0.6-3 ended normally after 24 iterations

  Optimization method                           NLMINB
  Number of free parameters                         30

  Number of observations                           815

  Estimator                                         ML
  Model Fit Test Statistic                    1580.004
  Degrees of freedom                                 5
  P-value (Chi-square)                           0.000

Parameter Estimates:

  Information                                 Expected
  Information saturated (h1) model          Structured
  Standard Errors                             Standard

Regressions:
                   Estimate  Std.Err  z-value  P(>|z|)
  govact ~                                            
    negemot   (c1)    0.120    0.041    2.942    0.003
    age       (c2)   -0.024    0.002  -10.572    0.000
    negemot:g (c3)    0.006    0.001   10.498    0.000
    posemot   (g1)   -0.021    0.028   -0.772    0.440
    ideology  (g2)   -0.212    0.026   -7.987    0.000
    sex       (g3)   -0.011    0.076   -0.148    0.883

Covariances:
                   Estimate  Std.Err  z-value  P(>|z|)
  negemot ~~                                          
    negemot:age     113.943    6.562   17.363    0.000
    posemot           0.263    0.073    3.622    0.000
    ideology         -0.805    0.086   -9.402    0.000
    sex              -0.090    0.027   -3.327    0.001
  negemot:age ~~                                      
    posemot          15.197    4.620    3.290    0.001
    ideology        -24.533    5.223   -4.697    0.000
    sex               0.163    1.705    0.096    0.924
  posemot ~~                                          
    ideology         -0.060    0.071   -0.838    0.402
    sex               0.050    0.024    2.115    0.034
  ideology ~~                                         
    sex               0.100    0.027    3.761    0.000

Intercepts:
                   Estimate  Std.Err  z-value  P(>|z|)
    age     (ag.m)   49.536    0.572   86.649    0.000
   .govact            5.174    0.217   23.807    0.000
    negemot           3.558    0.054   66.497    0.000
    negmt:g         174.831    3.410   51.268    0.000
    posemot           3.132    0.047   66.446    0.000
    ideolgy           4.083    0.053   77.159    0.000
    sex               0.488    0.018   27.890    0.000

Variances:
                   Estimate  Std.Err  z-value  P(>|z|)
    age     (ag.v)  266.367   13.195   20.187    0.000
   .govact            1.108    0.055   20.187    0.000
    negemot           2.333    0.116   20.187    0.000
    negmt:g        9477.578  469.498   20.187    0.000
    posemot           1.811    0.090   20.187    0.000
    ideolgy           2.283    0.113   20.187    0.000
    sex               0.250    0.012   20.187    0.000

Defined Parameters:
                   Estimate  Std.Err  z-value  P(>|z|)
    direct            0.433    0.026   16.446    0.000
    direct.below      0.330    0.029   11.496    0.000
    direct.above      0.537    0.028   19.320    0.000

Estimates Table

estimatesTable2(semfit,vanilla=TRUE)

Variables	Predictors	B	SE	z	p	β
govact	negemot	0.12	0.04	2.94	0.003	0.13
govact	age	-0.02	0.00	-10.57	< 0.001	-0.27
govact	negemot:age	0.01	0.00	10.50	< 0.001	0.43
govact	posemot	-0.02	0.03	-0.77	0.440	-0.02
govact	ideology	-0.21	0.03	-7.99	< 0.001	-0.22
govact	sex	-0.01	0.08	-0.15	0.883	-0.00

Correlation

corTable2(semfit)

rowname	govact	negemot	age	negemot.age	posemot	ideology	sex
govact	1
negemot	0.58***	1
age	-0.10**	-0.06	1
negemot.age	0.44***	0.77***	0.55***	1
posemot	0.04	0.13***	0.04	0.12**	1
ideology	-0.42***	-0.35***	0.21***	-0.17***	-0.03	1
sex	-0.10**	-0.12**	0.17***	0.00	0.07*	0.13***	1

Correlation Plot

corPlot(semfit)

Model Fit Table

modelFitTable2(semfit,vanilla=TRUE)

chisq	df	x2df	p	CFI	GFI	AGFI	TLI	RMR	SRMR	RMSEA(95% CI)	AIC	BIC
1580.00	5.00	316.00	0.00	0.21	0.99	0.90	-0.74	147.37	0.11	0.62(0.6-0.65)	28160.94	28302.03

Model Fit Guide

modelFitGuideTable2(vanilla=TRUE)

x2df	p	CFI	GFI	AGFI	TLI	RMR	SRMR	RMESA	AIC	BIC
< 3	> 0.05	> 0.9	> 0.9	> 0.9	> 0.9	< 0.05	< 0.05	< 0.1(< 0.05)	the lower, the better	the lower, the better

Statistical Diagram

statisticalDiagram(1,labels=labels,covar=covar,fit=semfit,whatLabel = "est")

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