CVLT_2model
Dissertation Analyses
load(file="/Users/meganwilliams/Desktop/Dissertation/StroopMixed.rdata")
load(file="/Users/meganwilliams/Desktop/Dissertation/Allvars.rdata")
library(effects)
library(interactions)
library(car)
library(rcompanion)California Verbal Learning Test (Total Correct Trial A) - Psychological Aggression
Model 1
CVLTlog1 <- glm(PsychAggress ~ CVLtca + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PsychAggress ~ CVLtca + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.2246 0.4582 0.5152 0.5595 0.7484
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.79553 0.63686 1.249 0.2116
## CVLtca 0.02714 0.01515 1.792 0.0732 .
## WRATtotal 0.01285 0.01537 0.836 0.4031
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 500.04 on 638 degrees of freedom
## AIC: 506.04
##
## Number of Fisher Scoring iterations: 4confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -0.421676866 2.08358015
## CVLtca -0.002635186 0.05685499
## WRATtotal -0.017830910 0.04259585exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 2.215620 0.6559460 8.033177
## CVLtca 1.027513 0.9973683 1.058502
## WRATtotal 1.012933 0.9823271 1.043516#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLtca 1 3.2102 0.07318 .
## WRATtotal 1 0.6990 0.40313
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plots
plot(allEffects(CVLTlog1))
########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 5.363783#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.0684336#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PsychAggress ~ CVLtca + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.01061290
## Cox and Snell (ML) 0.00833292
## Nagelkerke (Cragg and Uhler) 0.01527700
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -2.6819 5.3638 0.068434
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PsychAggress ~ (CVLtca + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PsychAggress ~ (CVLtca + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.3337 0.4129 0.4814 0.5690 0.9980
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.507921 1.041509 1.448 0.148
## CVLtca 0.032917 0.027323 1.205 0.228
## SexMen -0.490835 0.717735 -0.684 0.494
## PovStatBelow 1.064531 1.054751 1.009 0.313
## Age -0.018731 0.013525 -1.385 0.166
## WRATtotal 0.017976 0.015501 1.160 0.246
## CVLtca:SexMen 0.002083 0.035676 0.058 0.953
## CVLtca:PovStatBelow -0.047366 0.048002 -0.987 0.324
## SexMen:PovStatBelow 0.365349 1.389997 0.263 0.793
## CVLtca:SexMen:PovStatBelow -0.009220 0.068861 -0.134 0.893
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 490.18 on 631 degrees of freedom
## AIC: 510.18
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -0.49988121 3.595142572
## CVLtca -0.02103625 0.086641647
## SexMen -1.93022742 0.897900681
## PovStatBelow -0.91709632 3.263924014
## Age -0.04543934 0.007698432
## WRATtotal -0.01291089 0.048037702
## CVLtca:SexMen -0.06781452 0.072447800
## CVLtca:PovStatBelow -0.14316689 0.046005026
## SexMen:PovStatBelow -2.39711923 3.102579238
## CVLtca:SexMen:PovStatBelow -0.14502772 0.125832049exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 4.5173308 0.60660271 36.420892
## CVLtca 1.0334647 0.97918347 1.090506
## SexMen 0.6121152 0.14511519 2.454445
## PovStatBelow 2.8994777 0.39967789 26.151957
## Age 0.9814433 0.95557757 1.007728
## WRATtotal 1.0181383 0.98717210 1.049210
## CVLtca:SexMen 1.0020853 0.93443378 1.075137
## CVLtca:PovStatBelow 0.9537384 0.86660943 1.047080
## SexMen:PovStatBelow 1.4410172 0.09097967 22.255279
## CVLtca:SexMen:PovStatBelow 0.9908219 0.86499831 1.134092#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLtca 1 1.5459 0.21374
## Sex 1 2.7357 0.09813 .
## PovStat 1 1.2112 0.27109
## Age 1 1.9180 0.16607
## WRATtotal 1 1.3448 0.24619
## CVLtca:Sex 1 0.0002 0.98973
## CVLtca:PovStat 1 2.2706 0.13185
## Sex:PovStat 1 0.1247 0.72398
## CVLtca:Sex:PovStat 1 0.0179 0.89348
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 15.22463#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.08494771#Pseudo R-Squared
nagelkerke(CVLTlog3)## $Models
##
## Model: "glm, PsychAggress ~ (CVLtca + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0301238
## Cox and Snell (ML) 0.0234715
## Nagelkerke (Cragg and Uhler) 0.0430311
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -7.6123 15.225 0.084948
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PsychAggress ~ CVLtca + WRATtotal
## Model 2: PsychAggress ~ (CVLtca + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 500.04
## 2 631 490.18 7 9.8608 0.1966California Verbal Learning Test (Total Correct Trial A) - Physical Assault
Model 1
CVLTlog1 <- glm(PhysAssault ~ CVLtca + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PhysAssault ~ CVLtca + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.6025 -0.5336 -0.5109 -0.4793 2.1725
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.593942 0.732865 -3.539 0.000401 ***
## CVLtca 0.012518 0.015859 0.789 0.429923
## WRATtotal 0.009005 0.017158 0.525 0.599698
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 481.22 on 638 degrees of freedom
## AIC: 487.22
##
## Number of Fisher Scoring iterations: 4confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -4.09481803 -1.21492624
## CVLtca -0.01830031 0.04395313
## WRATtotal -0.02387068 0.04354938exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 0.07472492 0.01665878 0.2967319
## CVLtca 1.01259635 0.98186612 1.0449334
## WRATtotal 1.00904603 0.97641197 1.0445116#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLtca 1 0.6230 0.4299
## WRATtotal 1 0.2755 0.5997#Plots
plot(allEffects(CVLTlog1))
########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 1.3093#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.5196239#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PhysAssault ~ CVLtca + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00271339
## Cox and Snell (ML) 0.00204050
## Nagelkerke (Cragg and Uhler) 0.00385769
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -0.65465 1.3093 0.51962
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PhysAssault ~ (CVLtca + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PhysAssault ~ (CVLtca + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8807 -0.5521 -0.4699 -0.3868 2.3692
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.56944 1.17685 -1.334 0.1823
## CVLtca 0.01507 0.02827 0.533 0.5939
## SexMen 0.31859 0.90423 0.352 0.7246
## PovStatBelow 1.19628 0.93802 1.275 0.2022
## Age -0.03285 0.01415 -2.321 0.0203 *
## WRATtotal 0.01566 0.01802 0.869 0.3847
## CVLtca:SexMen -0.02338 0.04019 -0.582 0.5608
## CVLtca:PovStatBelow -0.03456 0.04091 -0.845 0.3982
## SexMen:PovStatBelow -1.59191 1.38680 -1.148 0.2510
## CVLtca:SexMen:PovStatBelow 0.08675 0.06569 1.321 0.1867
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 467.48 on 631 degrees of freedom
## AIC: 487.48
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -3.94186703 0.682851594
## CVLtca -0.03911156 0.072130045
## SexMen -1.45623021 2.114702092
## PovStatBelow -0.64898891 3.058400756
## Age -0.06103090 -0.005420145
## WRATtotal -0.01877807 0.052007524
## CVLtca:SexMen -0.10277124 0.055227806
## CVLtca:PovStatBelow -0.11546733 0.045572506
## SexMen:PovStatBelow -4.37034666 1.095833308
## CVLtca:SexMen:PovStatBelow -0.04073430 0.217556779exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 0.2081609 0.01941194 1.9795145
## CVLtca 1.0151877 0.96164343 1.0747951
## SexMen 1.3751903 0.23311341 8.2871166
## PovStatBelow 3.3077874 0.52257388 21.2934765
## Age 0.9676869 0.94079417 0.9945945
## WRATtotal 1.0157839 0.98139714 1.0533837
## CVLtca:SexMen 0.9768927 0.90233337 1.0567813
## CVLtca:PovStatBelow 0.9660316 0.89094968 1.0466269
## SexMen:PovStatBelow 0.2035359 0.01264686 2.9916746
## CVLtca:SexMen:PovStatBelow 1.0906242 0.96008419 1.2430360#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLtca 1 0.0398 0.84197
## Sex 1 0.3009 0.58329
## PovStat 1 3.9476 0.04694 *
## Age 1 5.3860 0.02030 *
## WRATtotal 1 0.7556 0.38471
## CVLtca:Sex 1 0.0824 0.77412
## CVLtca:PovStat 1 0.0010 0.97491
## Sex:PovStat 1 0.0462 0.82981
## CVLtca:Sex:PovStat 1 1.7437 0.18666
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 15.04792#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.08962819#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PhysAssault ~ CVLtca + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00271339
## Cox and Snell (ML) 0.00204050
## Nagelkerke (Cragg and Uhler) 0.00385769
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -0.65465 1.3093 0.51962
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PhysAssault ~ CVLtca + WRATtotal
## Model 2: PhysAssault ~ (CVLtca + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 481.22
## 2 631 467.48 7 13.739 0.05603 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1California Verbal Learning Test (Long Delayed Free Recall) - Psychological Aggression
Model 1
CVLTlog1 <- glm(PsychAggress ~ CVLfrl + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PsychAggress ~ CVLfrl + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.4487 0.4048 0.4990 0.5795 0.7633
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.872285 0.636356 1.371 0.17045
## CVLfrl 0.114613 0.037672 3.042 0.00235 **
## WRATtotal 0.009216 0.015087 0.611 0.54131
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 493.64 on 638 degrees of freedom
## AIC: 499.64
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -0.34280180 2.16049574
## CVLfrl 0.04168684 0.18965208
## WRATtotal -0.02093494 0.03838319exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 2.392371 0.7097789 8.675437
## CVLfrl 1.121439 1.0425679 1.208829
## WRATtotal 1.009258 0.9792827 1.039129#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLfrl 1 9.2560 0.002347 **
## WRATtotal 1 0.3731 0.541310
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plots
plot(allEffects(CVLTlog1))
########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 11.76059#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.002793964#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PsychAggress ~ CVLfrl + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0232698
## Cox and Snell (ML) 0.0181800
## Nagelkerke (Cragg and Uhler) 0.0333299
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -5.8803 11.761 0.002794
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PsychAggress ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PsychAggress ~ (CVLfrl + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.4092 0.3891 0.4796 0.5819 0.9066
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.594364 0.995644 1.601 0.109
## CVLfrl 0.081954 0.064865 1.263 0.206
## SexMen -0.539601 0.519371 -1.039 0.299
## PovStatBelow 0.034576 0.728639 0.047 0.962
## Age -0.013555 0.013553 -1.000 0.317
## WRATtotal 0.014485 0.015217 0.952 0.341
## CVLfrl:SexMen 0.020872 0.087237 0.239 0.811
## CVLfrl:PovStatBelow 0.019247 0.118391 0.163 0.871
## SexMen:PovStatBelow 0.403466 0.951778 0.424 0.672
## CVLfrl:SexMen:PovStatBelow 0.007243 0.175255 0.041 0.967
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.4 on 640 degrees of freedom
## Residual deviance: 487.7 on 631 degrees of freedom
## AIC: 507.7
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -0.33012860 3.58191681
## CVLfrl -0.04333937 0.21245834
## SexMen -1.58108983 0.46411696
## PovStatBelow -1.35788726 1.53077967
## Age -0.04029900 0.01294495
## WRATtotal -0.01586669 0.04396449
## CVLfrl:SexMen -0.15057415 0.19264865
## CVLfrl:PovStatBelow -0.21063952 0.25722477
## SexMen:PovStatBelow -1.48466750 2.27285707
## CVLfrl:SexMen:PovStatBelow -0.33464295 0.35634530exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 4.9251943 0.7188313 35.942369
## CVLfrl 1.0854060 0.9575864 1.236715
## SexMen 0.5829807 0.2057507 1.590609
## PovStatBelow 1.0351811 0.2572036 4.621779
## Age 0.9865361 0.9605022 1.013029
## WRATtotal 1.0145904 0.9842585 1.044945
## CVLfrl:SexMen 1.0210918 0.8602139 1.212457
## CVLfrl:PovStatBelow 1.0194336 0.8100660 1.293336
## SexMen:PovStatBelow 1.4970042 0.2265777 9.707095
## CVLfrl:SexMen:PovStatBelow 1.0072689 0.7155935 1.428101#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLfrl 1 6.0502 0.0139 *
## Sex 1 1.7947 0.1804
## PovStat 1 1.5312 0.2159
## Age 1 1.0004 0.3172
## WRATtotal 1 0.9061 0.3411
## CVLfrl:Sex 1 0.0897 0.7645
## CVLfrl:PovStat 1 0.0666 0.7964
## Sex:PovStat 1 0.6148 0.4330
## CVLfrl:Sex:PovStat 1 0.0017 0.9670
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 17.70202#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.03879225#Pseudo R-Squared
nagelkerke(CVLTlog3)## $Models
##
## Model: "glm, PsychAggress ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0350257
## Cox and Snell (ML) 0.0272384
## Nagelkerke (Cragg and Uhler) 0.0499371
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -8.851 17.702 0.038792
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PsychAggress ~ CVLfrl + WRATtotal
## Model 2: PsychAggress ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 493.64
## 2 631 487.70 7 5.9414 0.5466California Verbal Learning Test (Long Delayed Free Recall) - Physical Assault
Model 1
CVLTlog1 <- glm(PhysAssault ~ CVLfrl + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PhysAssault ~ CVLfrl + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.6087 -0.5337 -0.5068 -0.4807 2.1707
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.53235 0.72754 -3.481 0.0005 ***
## CVLfrl 0.03102 0.03648 0.850 0.3952
## WRATtotal 0.00920 0.01700 0.541 0.5883
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 481.13 on 638 degrees of freedom
## AIC: 487.13
##
## Number of Fisher Scoring iterations: 4confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -4.02329131 -1.16424365
## CVLfrl -0.04046725 0.10283424
## WRATtotal -0.02336295 0.04342529exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 0.07947179 0.01789397 0.3121587
## CVLfrl 1.03150469 0.96034061 1.1083077
## WRATtotal 1.00924276 0.97690785 1.0443820#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLfrl 1 0.723 0.3952
## WRATtotal 1 0.293 0.5883########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 1.405392#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.4952483#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PhysAssault ~ CVLfrl + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00291253
## Cox and Snell (ML) 0.00219010
## Nagelkerke (Cragg and Uhler) 0.00414050
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -0.7027 1.4054 0.49525
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PhysAssault ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PhysAssault ~ (CVLfrl + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.9608 -0.5456 -0.4611 -0.3756 2.3880
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.54116 1.11091 -1.387 0.1654
## CVLfrl 0.05847 0.06311 0.926 0.3542
## SexMen 0.16453 0.66352 0.248 0.8042
## PovStatBelow 1.74987 0.69446 2.520 0.0117 *
## Age -0.03334 0.01429 -2.333 0.0196 *
## WRATtotal 0.01402 0.01794 0.782 0.4345
## CVLfrl:SexMen -0.04768 0.09093 -0.524 0.6000
## CVLfrl:PovStatBelow -0.21185 0.10025 -2.113 0.0346 *
## SexMen:PovStatBelow -1.68383 1.02211 -1.647 0.0995 .
## CVLfrl:SexMen:PovStatBelow 0.30140 0.15435 1.953 0.0509 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 463.48 on 631 degrees of freedom
## AIC: 483.48
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -3.768138754 0.596098027
## CVLfrl -0.063914094 0.184956139
## SexMen -1.134460897 1.487606633
## PovStatBelow 0.394379812 3.136825838
## Age -0.061778689 -0.005627172
## WRATtotal -0.020308030 0.050188403
## CVLfrl:SexMen -0.228410112 0.129351707
## CVLfrl:PovStatBelow -0.412916523 -0.018096870
## SexMen:PovStatBelow -3.734693589 0.293286423
## CVLfrl:SexMen:PovStatBelow 0.001638781 0.608694441exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 0.2141328 0.02309501 1.8150228
## CVLfrl 1.0602112 0.93808558 1.2031657
## SexMen 1.1788423 0.32159545 4.4264886
## PovStatBelow 5.7538469 1.48346388 23.0306477
## Age 0.9672068 0.94009092 0.9943886
## WRATtotal 1.0141210 0.97989679 1.0514692
## CVLfrl:SexMen 0.9534429 0.79579783 1.1380903
## CVLfrl:PovStatBelow 0.8090857 0.66171752 0.9820659
## SexMen:PovStatBelow 0.1856610 0.02388049 1.3408268
## CVLfrl:SexMen:PovStatBelow 1.3517439 1.00164012 1.8380302#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLfrl 1 0.0146 0.90370
## Sex 1 0.3583 0.54947
## PovStat 1 4.0028 0.04542 *
## Age 1 5.4437 0.01964 *
## WRATtotal 1 0.6109 0.43447
## CVLfrl:Sex 1 0.6043 0.43693
## CVLfrl:PovStat 1 1.2288 0.26764
## Sex:PovStat 1 0.0075 0.93115
## CVLfrl:Sex:PovStat 1 3.8128 0.05086 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plots
interact_plot(model = CVLTlog3, pred = CVLfrl, modx = PovStat)
sim_slopes(CVLTlog3, pred = CVLfrl, modx = PovStat, centered = "all",jnplot = TRUE)## Warning: Johnson-Neyman intervals are not available for factor moderators.## SIMPLE SLOPES ANALYSIS
##
## Slope of CVLfrl when PovStat = Below:
##
## Est. S.E. z val. p
## ------- ------ -------- ------
## -0.03 0.06 -0.48 0.63
##
## Slope of CVLfrl when PovStat = Above:
##
## Est. S.E. z val. p
## ------ ------ -------- ------
## 0.04 0.05 0.74 0.46########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 19.05303#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.02474572#Pseudo R-Squared
nagelkerke(CVLTlog3)## $Models
##
## Model: "glm, PhysAssault ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0394855
## Cox and Snell (ML) 0.0292865
## Nagelkerke (Cragg and Uhler) 0.0553677
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -9.5265 19.053 0.024746
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PhysAssault ~ CVLfrl + WRATtotal
## Model 2: PhysAssault ~ (CVLfrl + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 481.13
## 2 631 463.48 7 17.648 0.01367 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1California Verbal Learning Test (Short Delayed Free Recall) - Psychological Aggression
Model 1
CVLTlog1 <- glm(PsychAggress ~ CVLfrs + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PsychAggress ~ CVLfrs + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.3416 0.4408 0.5101 0.5760 0.7676
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.90534 0.63284 1.431 0.1525
## CVLfrs 0.08323 0.03743 2.223 0.0262 *
## WRATtotal 0.01185 0.01515 0.782 0.4343
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 498.21 on 638 degrees of freedom
## AIC: 504.21
##
## Number of Fisher Scoring iterations: 4confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -0.30325412 2.18649124
## CVLfrs 0.01039890 0.15743330
## WRATtotal -0.01839075 0.04119211exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 2.472761 0.7384114 8.903916
## CVLfrs 1.086792 1.0104532 1.170503
## WRATtotal 1.011919 0.9817773 1.042052#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLfrs 1 4.9439 0.02618 *
## WRATtotal 1 0.6113 0.43432
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plots
plot(allEffects(CVLTlog1))
########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 7.193642#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.02741072#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PsychAggress ~ CVLfrs + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0142335
## Cox and Snell (ML) 0.0111598
## Nagelkerke (Cragg and Uhler) 0.0204596
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -3.5968 7.1936 0.027411
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PsychAggress ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PsychAggress ~ (CVLfrs + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.4191 0.4137 0.4819 0.5750 0.8827
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 1.746423 0.985751 1.772 0.0764 .
## CVLfrs 0.061861 0.064502 0.959 0.3375
## SexMen -0.521850 0.526470 -0.991 0.3216
## PovStatBelow 0.451159 0.777362 0.580 0.5617
## Age -0.016645 0.013563 -1.227 0.2197
## WRATtotal 0.017110 0.015252 1.122 0.2619
## CVLfrs:SexMen 0.009237 0.085199 0.108 0.9137
## CVLfrs:PovStatBelow -0.058364 0.117321 -0.497 0.6189
## SexMen:PovStatBelow -0.178900 1.012604 -0.177 0.8598
## CVLfrs:SexMen:PovStatBelow 0.114186 0.174686 0.654 0.5133
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 490.65 on 631 degrees of freedom
## AIC: 510.65
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -0.15644107 3.717405147
## CVLfrs -0.06328045 0.191168570
## SexMen -1.57745880 0.496345881
## PovStatBelow -1.01346251 2.072883547
## Age -0.04341788 0.009867831
## WRATtotal -0.01327224 0.046698314
## CVLfrs:SexMen -0.15853628 0.176616872
## CVLfrs:PovStatBelow -0.29008092 0.173430402
## SexMen:PovStatBelow -2.20399826 1.799787040
## CVLfrs:SexMen:PovStatBelow -0.22596721 0.462209001exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 5.7340536 0.8551819 41.157458
## CVLfrs 1.0638144 0.9386802 1.210664
## SexMen 0.5934216 0.2064992 1.642708
## PovStatBelow 1.5701307 0.3629600 7.947708
## Age 0.9834927 0.9575112 1.009917
## WRATtotal 1.0172574 0.9868154 1.047806
## CVLfrs:SexMen 1.0092796 0.8533920 1.193174
## CVLfrs:PovStatBelow 0.9433069 0.7482030 1.189378
## SexMen:PovStatBelow 0.8361892 0.1103610 6.048359
## CVLfrs:SexMen:PovStatBelow 1.1209605 0.7977443 1.587577#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## CVLfrs 1 2.6509 0.1035
## Sex 1 2.5011 0.1138
## PovStat 1 1.2946 0.2552
## Age 1 1.5061 0.2197
## WRATtotal 1 1.2585 0.2619
## CVLfrs:Sex 1 0.2400 0.6242
## CVLfrs:PovStat 1 0.0061 0.9375
## Sex:PovStat 1 0.4491 0.5028
## CVLfrs:Sex:PovStat 1 0.4273 0.5133########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 14.75029#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.09802732#Pseudo R-Squared
nagelkerke(CVLTlog3)## $Models
##
## Model: "glm, PsychAggress ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0291853
## Cox and Snell (ML) 0.0227486
## Nagelkerke (Cragg and Uhler) 0.0417058
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -7.3751 14.75 0.098027
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PsychAggress ~ CVLfrs + WRATtotal
## Model 2: PsychAggress ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 498.21
## 2 631 490.65 7 7.5567 0.3733California Verbal Learning Test (Short Delayed Free Recall) - Physical Assault
Model 1
CVLTlog1 <- glm(PhysAssault ~ CVLfrs + WRATtotal, data=Allvars,family = "binomial")
summary(CVLTlog1)##
## Call:
## glm(formula = PhysAssault ~ CVLfrs + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.5732 -0.5300 -0.5146 -0.4875 2.1798
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.53166 0.72810 -3.477 0.000507 ***
## CVLfrs 0.01456 0.03723 0.391 0.695685
## WRATtotal 0.01138 0.01700 0.669 0.503398
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 481.70 on 638 degrees of freedom
## AIC: 487.7
##
## Number of Fisher Scoring iterations: 4confint(CVLTlog1)## 2.5 % 97.5 %
## (Intercept) -4.02371675 -1.16251937
## CVLfrs -0.05841422 0.08781414
## WRATtotal -0.02119180 0.04562049exp(cbind(OR = coef(CVLTlog1), confint(CVLTlog1)))## OR 2.5 % 97.5 %
## (Intercept) 0.07952681 0.01788636 0.3126974
## CVLfrs 1.01466740 0.94325915 1.0917852
## WRATtotal 1.01144314 0.97903117 1.0466771#Wald chi-square Test
Anova(CVLTlog1, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLfrs 1 0.1530 0.6957
## WRATtotal 1 0.4478 0.5034#Plots
plot(predictorEffect("CVLfrs",CVLTlog1))
########Compare to null model
#Difference in Deviance
with(CVLTlog1,null.deviance - deviance)## [1] 0.8346165#Degrees of freedom for the difference between two models
with(CVLTlog1,df.null - df.residual)## [1] 2#p-value
with(CVLTlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.6588178#Pseudo R-Squared
nagelkerke(CVLTlog1)## $Models
##
## Model: "glm, PhysAssault ~ CVLfrs + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00172966
## Cox and Snell (ML) 0.00130121
## Nagelkerke (Cragg and Uhler) 0.00246000
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -0.41731 0.83462 0.65882
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Model 3
CVLTlog3 <- glm(PhysAssault ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(CVLTlog3)##
## Call:
## glm(formula = PhysAssault ~ (CVLfrs + Sex + PovStat)^3 + Age +
## WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -1.0298 -0.5418 -0.4576 -0.3739 2.5096
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.75276 1.12368 -1.560 0.11880
## CVLfrs 0.07310 0.06396 1.143 0.25305
## SexMen 0.42285 0.67236 0.629 0.52942
## PovStatBelow 2.04485 0.69990 2.922 0.00348 **
## Age -0.03437 0.01427 -2.408 0.01602 *
## WRATtotal 0.01754 0.01814 0.967 0.33363
## CVLfrs:SexMen -0.09053 0.09203 -0.984 0.32529
## CVLfrs:PovStatBelow -0.26603 0.10348 -2.571 0.01014 *
## SexMen:PovStatBelow -1.90367 1.03608 -1.837 0.06615 .
## CVLfrs:SexMen:PovStatBelow 0.33770 0.15904 2.123 0.03372 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 482.53 on 640 degrees of freedom
## Residual deviance: 461.85 on 631 degrees of freedom
## AIC: 481.85
##
## Number of Fisher Scoring iterations: 5confint(CVLTlog3)## 2.5 % 97.5 %
## (Intercept) -4.00877842 0.40460004
## CVLfrs -0.05036067 0.20195206
## SexMen -0.89138897 1.76734247
## PovStatBelow 0.68639107 3.45020200
## Age -0.06278525 -0.00671841
## WRATtotal -0.01713924 0.05412889
## CVLfrs:SexMen -0.27352409 0.08858167
## CVLfrs:PovStatBelow -0.47447559 -0.06691616
## SexMen:PovStatBelow -3.98657186 0.09702919
## CVLfrs:SexMen:PovStatBelow 0.02909837 0.65460576exp(cbind(OR = coef(CVLTlog3), confint(CVLTlog3)))## OR 2.5 % 97.5 %
## (Intercept) 0.1732950 0.01815556 1.4987030
## CVLfrs 1.0758389 0.95088640 1.2237893
## SexMen 1.5262979 0.41008576 5.8552721
## PovStatBelow 7.7279951 1.98653331 31.5067561
## Age 0.9662183 0.93914513 0.9933041
## WRATtotal 1.0176916 0.98300680 1.0556207
## CVLfrs:SexMen 0.9134514 0.76069401 1.0926235
## CVLfrs:PovStatBelow 0.7664193 0.62221126 0.9352736
## SexMen:PovStatBelow 0.1490209 0.01856324 1.1018925
## CVLfrs:SexMen:PovStatBelow 1.4017152 1.02952586 1.9243837#Wald chi-square Test
Anova(CVLTlog3, type="II", test="Wald")## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## CVLfrs 1 0.0765 0.78210
## Sex 1 0.4477 0.50345
## PovStat 1 3.8651 0.04930 *
## Age 1 5.8001 0.01602 *
## WRATtotal 1 0.9348 0.33363
## CVLfrs:Sex 1 0.0923 0.76128
## CVLfrs:PovStat 1 2.4495 0.11756
## Sex:PovStat 1 0.0001 0.99069
## CVLfrs:Sex:PovStat 1 4.5087 0.03372 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#Plots
interact_plot(model = CVLTlog3, pred = CVLfrs, modx = Sex,mod2 = PovStat)
sim_slopes(CVLTlog3, pred = CVLfrs, modx = Sex, mod2 = PovStat, centered = "all",jnplot = TRUE)## Warning: Johnson-Neyman intervals are not available for factor moderators.## █████████████████████ While PovStat (2nd moderator) = Above ████████████████████
##
## SIMPLE SLOPES ANALYSIS
##
## Slope of CVLfrs when Sex = Women:
##
## Est. S.E. z val. p
## ------ ------ -------- ------
## 0.07 0.06 1.14 0.25
##
## Slope of CVLfrs when Sex = Men:
##
## Est. S.E. z val. p
## ------- ------ -------- ------
## -0.02 0.07 -0.25 0.80
##
## █████████████████████ While PovStat (2nd moderator) = Below ████████████████████
##
## SIMPLE SLOPES ANALYSIS
##
## Slope of CVLfrs when Sex = Women:
##
## Est. S.E. z val. p
## ------- ------ -------- ------
## -0.19 0.08 -2.30 0.02
##
## Slope of CVLfrs when Sex = Men:
##
## Est. S.E. z val. p
## ------ ------ -------- ------
## 0.05 0.10 0.53 0.59
##
## NULL########Compare to null model
#Difference in Deviance
with(CVLTlog3,null.deviance - deviance)## [1] 20.67999#Degrees of freedom for the difference between two models
with(CVLTlog3,df.null - df.residual)## [1] 9#p-value
with(CVLTlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))## [1] 0.01414967#Pseudo R-Squared
nagelkerke(CVLTlog3)## $Models
##
## Model: "glm, PhysAssault ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0428572
## Cox and Snell (ML) 0.0317472
## Nagelkerke (Cragg and Uhler) 0.0600198
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -10.34 20.68 0.01415
##
## $Number.of.observations
##
## Model: 641
## Null: 641
##
## $Messages
## [1] "Note: For models fit with REML, these statistics are based on refitting with ML"
##
## $Warnings
## [1] "None"Compare Models 1 & 3
anova(CVLTlog1,CVLTlog3,test = "LR")## Analysis of Deviance Table
##
## Model 1: PhysAssault ~ CVLfrs + WRATtotal
## Model 2: PhysAssault ~ (CVLfrs + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 481.70
## 2 631 461.85 7 19.845 0.005914 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1