Animal Naming Test - Psychological Aggression
Model 1
ANlog1 <- glm(PsychAggress ~ FluencyWord + WRATtotal, data=Allvars,family = "binomial")
summary(ANlog1)
##
## Call:
## glm(formula = PsychAggress ~ FluencyWord + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.1467 0.4854 0.5222 0.5561 0.7324
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 0.78911 0.66312 1.190 0.234
## FluencyWord 0.01647 0.02323 0.709 0.478
## WRATtotal 0.01757 0.01569 1.120 0.263
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 505.40 on 640 degrees of freedom
## Residual deviance: 502.72 on 638 degrees of freedom
## AIC: 508.72
##
## Number of Fisher Scoring iterations: 4
confint(ANlog1)
## 2.5 % 97.5 %
## (Intercept) -0.48613879 2.12175343
## FluencyWord -0.02858561 0.06261612
## WRATtotal -0.01377810 0.04792455
exp(cbind(OR = coef(ANlog1), confint(ANlog1)))
## OR 2.5 % 97.5 %
## (Intercept) 2.201430 0.6149964 8.345758
## FluencyWord 1.016608 0.9718191 1.064618
## WRATtotal 1.017724 0.9863164 1.049092
#Wald chi-square Test
Anova(ANlog1, type="II", test="Wald")
## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## FluencyWord 1 0.5028 0.4783
## WRATtotal 1 1.2533 0.2629
########Compare to null model
#Difference in Deviance
with(ANlog1,null.deviance - deviance)
## [1] 2.676568
#Degrees of freedom for the difference between two models
with(ANlog1,df.null - df.residual)
## [1] 2
#p-value
with(ANlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))
## [1] 0.2622953
#Pseudo R-Squared
nagelkerke(ANlog1)
## $Models
##
## Model: "glm, PsychAggress ~ FluencyWord + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00529593
## Cox and Snell (ML) 0.00416691
## Nagelkerke (Cragg and Uhler) 0.00763933
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -1.3383 2.6766 0.2623
##
## $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
ANlog3 <- glm(PsychAggress ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(ANlog3)
##
## Call:
## glm(formula = PsychAggress ~ (FluencyWord + Sex + PovStat)^3 +
## Age + WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.3284 0.4143 0.4835 0.5639 1.0062
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) 2.008377 1.153266 1.741 0.0816 .
## FluencyWord 0.011820 0.038971 0.303 0.7617
## SexMen -1.424781 1.027353 -1.387 0.1655
## PovStatBelow -0.019326 1.426951 -0.014 0.9892
## Age -0.020953 0.013261 -1.580 0.1141
## WRATtotal 0.019146 0.015830 1.209 0.2265
## FluencyWord:SexMen 0.043940 0.052092 0.844 0.3989
## FluencyWord:PovStatBelow 0.007169 0.077334 0.093 0.9261
## SexMen:PovStatBelow 2.731876 2.084764 1.310 0.1901
## FluencyWord:SexMen:PovStatBelow -0.119742 0.105978 -1.130 0.2585
## ---
## 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: 490.61 on 631 degrees of freedom
## AIC: 510.61
##
## Number of Fisher Scoring iterations: 5
confint(ANlog3)
## 2.5 % 97.5 %
## (Intercept) -0.23228008 4.29821884
## FluencyWord -0.06215661 0.09117398
## SexMen -3.44849510 0.58663361
## PovStatBelow -2.81316385 2.80853892
## Age -0.04715221 0.00494707
## WRATtotal -0.01238725 0.04984477
## FluencyWord:SexMen -0.05919918 0.14563277
## FluencyWord:PovStatBelow -0.14070102 0.16460308
## SexMen:PovStatBelow -1.31542020 6.89273664
## FluencyWord:SexMen:PovStatBelow -0.33129450 0.08587307
exp(cbind(OR = coef(ANlog3), confint(ANlog3)))
## OR 2.5 % 97.5 %
## (Intercept) 7.4512169 0.79272407 73.568640
## FluencyWord 1.0118901 0.93973570 1.095460
## SexMen 0.2405612 0.03179345 1.797926
## PovStatBelow 0.9808595 0.06001481 16.585667
## Age 0.9792650 0.95394219 1.004959
## WRATtotal 1.0193306 0.98768916 1.051108
## FluencyWord:SexMen 1.0449201 0.94251902 1.156771
## FluencyWord:PovStatBelow 1.0071944 0.86874901 1.178925
## SexMen:PovStatBelow 15.3616755 0.26836153 985.093576
## FluencyWord:SexMen:PovStatBelow 0.8871492 0.71799369 1.089668
#Wald chi-square Test
Anova(ANlog3, type="II", test="Wald")
## Analysis of Deviance Table (Type II tests)
##
## Response: PsychAggress
## Df Chisq Pr(>Chisq)
## FluencyWord 1 0.8003 0.37101
## Sex 1 3.8177 0.05072 .
## PovStat 1 0.9826 0.32155
## Age 1 2.4964 0.11410
## WRATtotal 1 1.4628 0.22649
## FluencyWord:Sex 1 0.1096 0.74061
## FluencyWord:PovStat 1 1.1462 0.28436
## Sex:PovStat 1 0.6926 0.40529
## FluencyWord:Sex:PovStat 1 1.2766 0.25853
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
########Compare to null model
#Difference in Deviance
with(ANlog3,null.deviance - deviance)
## [1] 14.78833
#Degrees of freedom for the difference between two models
with(ANlog3,df.null - df.residual)
## [1] 9
#p-value
with(ANlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))
## [1] 0.09691678
#Pseudo R-Squared
nagelkerke(ANlog3)
## $Models
##
## Model: "glm, PsychAggress ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PsychAggress ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0292606
## Cox and Snell (ML) 0.0228066
## Nagelkerke (Cragg and Uhler) 0.0418121
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -7.3942 14.788 0.096917
##
## $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(ANlog1,ANlog3,test = "LR")
## Analysis of Deviance Table
##
## Model 1: PsychAggress ~ FluencyWord + WRATtotal
## Model 2: PsychAggress ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 502.72
## 2 631 490.61 7 12.112 0.09694 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Animal Naming Test - Physical Assault
Model 1
ANlog1 <- glm(PhysAssault ~ FluencyWord + WRATtotal, data=Allvars,family = "binomial")
summary(ANlog1)
##
## Call:
## glm(formula = PhysAssault ~ FluencyWord + WRATtotal, family = "binomial",
## data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.6213 -0.5361 -0.5098 -0.4756 2.1770
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.36821 0.75469 -3.138 0.0017 **
## FluencyWord -0.02147 0.02364 -0.908 0.3638
## WRATtotal 0.01914 0.01760 1.088 0.2768
## ---
## 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: 481.02 on 638 degrees of freedom
## AIC: 487.02
##
## Number of Fisher Scoring iterations: 4
confint(ANlog1)
## 2.5 % 97.5 %
## (Intercept) -3.90449219 -0.93822551
## FluencyWord -0.06846401 0.02433492
## WRATtotal -0.01443322 0.05470170
exp(cbind(OR = coef(ANlog1), confint(ANlog1)))
## OR 2.5 % 97.5 %
## (Intercept) 0.09364814 0.02015118 0.3913216
## FluencyWord 0.97876343 0.93382707 1.0246334
## WRATtotal 1.01932175 0.98567044 1.0562255
#Wald chi-square Test
Anova(ANlog1, type="II", test="Wald")
## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## FluencyWord 1 0.8248 0.3638
## WRATtotal 1 1.1827 0.2768
########Compare to null model
#Difference in Deviance
with(ANlog1,null.deviance - deviance)
## [1] 1.515757
#Degrees of freedom for the difference between two models
with(ANlog1,df.null - df.residual)
## [1] 2
#p-value
with(ANlog1,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))
## [1] 0.4686597
#Pseudo R-Squared
nagelkerke(ANlog1)
## $Models
##
## Model: "glm, PhysAssault ~ FluencyWord + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.00314125
## Cox and Snell (ML) 0.00236188
## Nagelkerke (Cragg and Uhler) 0.00446527
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -2 -0.75788 1.5158 0.46866
##
## $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
ANlog3 <- glm(PhysAssault ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal, data = Allvars, family = "binomial")
summary(ANlog3)
##
## Call:
## glm(formula = PhysAssault ~ (FluencyWord + Sex + PovStat)^3 +
## Age + WRATtotal, family = "binomial", data = Allvars)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.8525 -0.5493 -0.4681 -0.3801 2.5158
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -1.600446 1.212631 -1.320 0.1869
## FluencyWord 0.003624 0.036624 0.099 0.9212
## SexMen 0.975863 1.161915 0.840 0.4010
## PovStatBelow 1.495936 1.229964 1.216 0.2239
## Age -0.034897 0.013937 -2.504 0.0123 *
## WRATtotal 0.024426 0.018371 1.330 0.1837
## FluencyWord:SexMen -0.056565 0.056061 -1.009 0.3130
## FluencyWord:PovStatBelow -0.057081 0.064744 -0.882 0.3780
## SexMen:PovStatBelow -1.816965 1.938739 -0.937 0.3487
## FluencyWord:SexMen:PovStatBelow 0.098700 0.097335 1.014 0.3106
## ---
## 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: 466.97 on 631 degrees of freedom
## AIC: 486.97
##
## Number of Fisher Scoring iterations: 5
confint(ANlog3)
## 2.5 % 97.5 %
## (Intercept) -4.01506723 0.747960669
## FluencyWord -0.06998106 0.074381636
## SexMen -1.31267722 3.255391216
## PovStatBelow -0.90815521 3.931751771
## Age -0.06267745 -0.007917641
## WRATtotal -0.01061634 0.061555509
## FluencyWord:SexMen -0.16772931 0.052793528
## FluencyWord:PovStatBelow -0.18772678 0.067587070
## SexMen:PovStatBelow -5.66030108 1.961694248
## FluencyWord:SexMen:PovStatBelow -0.09128825 0.291518237
exp(cbind(OR = coef(ANlog3), confint(ANlog3)))
## OR 2.5 % 97.5 %
## (Intercept) 0.2018065 0.018041742 2.1126872
## FluencyWord 1.0036301 0.932411483 1.0772178
## SexMen 2.6534555 0.269098655 25.9297567
## PovStatBelow 4.4635102 0.403267482 50.9962332
## Age 0.9657049 0.939246381 0.9921136
## WRATtotal 1.0247266 0.989439812 1.0634895
## FluencyWord:SexMen 0.9450048 0.845582691 1.0542120
## FluencyWord:PovStatBelow 0.9445178 0.828841128 1.0699234
## SexMen:PovStatBelow 0.1625182 0.003481469 7.1113653
## FluencyWord:SexMen:PovStatBelow 1.1037354 0.912754575 1.3384580
#Wald chi-square Test
Anova(ANlog3, type="II", test="Wald")
## Analysis of Deviance Table (Type II tests)
##
## Response: PhysAssault
## Df Chisq Pr(>Chisq)
## FluencyWord 1 0.9577 0.32777
## Sex 1 0.2036 0.65182
## PovStat 1 3.5181 0.06070 .
## Age 1 6.2696 0.01228 *
## WRATtotal 1 1.7678 0.18365
## FluencyWord:Sex 1 0.2705 0.60301
## FluencyWord:PovStat 1 0.0770 0.78134
## Sex:PovStat 1 0.0245 0.87554
## FluencyWord:Sex:PovStat 1 1.0283 0.31057
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
########Compare to null model
#Difference in Deviance
with(ANlog3,null.deviance - deviance)
## [1] 15.56602
#Degrees of freedom for the difference between two models
with(ANlog3,df.null - df.residual)
## [1] 9
#p-value
with(ANlog3,pchisq(null.deviance-deviance,df.null-df.residual,lower.tail=FALSE))
## [1] 0.07651624
#Pseudo R-Squared
nagelkerke(ANlog3)
## $Models
##
## Model: "glm, PhysAssault ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal, binomial, Allvars"
## Null: "glm, PhysAssault ~ 1, binomial, Allvars"
##
## $Pseudo.R.squared.for.model.vs.null
## Pseudo.R.squared
## McFadden 0.0322590
## Cox and Snell (ML) 0.0239915
## Nagelkerke (Cragg and Uhler) 0.0453572
##
## $Likelihood.ratio.test
## Df.diff LogLik.diff Chisq p.value
## -9 -7.783 15.566 0.076516
##
## $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(ANlog1,ANlog3,test = "LR")
## Analysis of Deviance Table
##
## Model 1: PhysAssault ~ FluencyWord + WRATtotal
## Model 2: PhysAssault ~ (FluencyWord + Sex + PovStat)^3 + Age + WRATtotal
## Resid. Df Resid. Dev Df Deviance Pr(>Chi)
## 1 638 481.02
## 2 631 466.97 7 14.05 0.05029 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1