#5.4 The Students data file at the text website shows responses of a class of social science graduate students at the University of Florida to a questionnaire that asked about gender (1 = female, 0 = male), age, hsgpa = high school GPA (on a four-point scale), cogpa = college GPA, dhome = distance (in miles) of the campus from your home town, dres = distance (in miles) of the classroom from your current residence, tv=average number of hours perweek that youwatch TV, sport=average number of hours per week that you participate in sports or have other physical exercise, news = number of times a week you read a newspaper, aids = number of people you know who have died from AIDS or who are HIV+, veg = whether you are a vegetarian (1 = yes, 0 = no), affil = political affiliation (1 = Democrat, 2 = Republican, 3 = Independent), ideol = political ideology (1 = very liberal, 2 = liberal, 3 = slightly liberal, 4 = moderate, 5 = slightly conservative, 6 = conservative, 7 = very conservative), relig = how often you attend religious services (0 = never, 1 = occasionally, 2 = most weeks, 3 = every week), abor = opinion about whether abortion should be legal in the first three months of pregnancy (1 = yes, 0 = no), affirm = support affirmative action (1 = yes, 0 = no), and life = belief in life after death (1 = yes, 2 = no, 3 = undecided).
## subject gender age hsgpa cogpa dhome dres tv sport news aids veg affil
## 1 1 0 32 2.2 3.5 0 5.00 3.0 5 0 0 0 2
## 2 2 1 23 2.1 3.5 1200 0.30 15.0 7 5 6 1 1
## 3 3 1 27 3.3 3.0 1300 1.50 0.0 4 3 0 1 1
## 4 4 1 35 3.5 3.2 1500 8.00 5.0 5 6 3 0 3
## 5 5 0 23 3.1 3.5 1600 10.00 6.0 6 3 0 0 3
## 6 6 0 39 3.5 3.5 350 3.00 4.0 5 7 0 1 1
## 7 7 0 24 3.6 3.7 0 0.20 5.0 12 4 2 0 3
## 8 8 1 31 3.0 3.0 5000 1.50 5.0 3 3 1 0 3
## 9 9 0 34 3.0 3.0 5000 2.00 7.0 5 3 0 0 3
## 10 10 0 28 4.0 3.1 900 2.00 1.0 1 2 1 1 3
## 11 11 0 23 2.3 2.6 253 1.50 10.0 15 1 1 0 2
## 12 12 1 27 3.5 3.6 190 3.00 14.0 3 7 0 0 1
## 13 13 0 36 3.3 3.5 245 1.50 6.0 15 12 5 0 1
## 14 14 0 28 3.2 3.2 500 6.00 3.0 10 1 2 0 3
## 15 15 1 28 3.0 3.5 3500 1.00 4.0 3 1 0 0 1
## 16 16 1 25 3.8 3.3 210 10.00 7.0 6 1 0 1 3
## 17 17 1 41 4.0 3.0 1000 15.00 6.0 7 3 10 0 3
## 18 18 0 50 3.8 3.8 0 3.00 5.0 9 6 10 0 1
## 19 19 0 71 4.0 3.5 5000 3.00 6.0 12 2 2 0 3
## 20 20 1 28 3.0 3.8 120 1.00 25.0 0 0 2 1 1
## 21 21 1 26 3.7 3.7 8000 8.00 4.0 4 4 1 0 3
## 22 22 1 27 4.0 3.7 2 2.50 4.0 2 7 0 0 3
## 23 23 0 31 2.7 3.5 1700 5.00 7.0 7 2 0 0 2
## 24 24 1 23 3.7 3.7 2 2.00 7.0 4 2 0 0 3
## 25 25 0 23 3.2 3.8 450 4.00 0.0 7 7 3 0 3
## 26 26 1 44 3.0 3.0 0 2.00 2.0 3 2 3 1 3
## 27 27 0 26 3.7 3.0 1000 3.00 8.0 2 7 0 0 1
## 28 28 1 31 3.7 3.8 850 10.00 10.0 3 7 0 0 2
## 29 29 0 24 3.3 3.1 420 2.00 10.0 6 5 0 0 1
## 30 30 1 26 3.3 3.3 1200 0.75 10.0 0 3 0 0 2
## 31 31 0 26 3.3 3.5 1000 1.50 0.0 3 3 3 1 1
## 32 32 1 32 3.5 3.9 150 12.00 10.0 2 0 0 0 1
## 33 33 0 26 3.4 3.4 2000 1.50 2.0 7 14 0 0 1
## 34 34 1 22 3.2 2.8 316 2.00 10.0 3 5 2 0 3
## 35 35 1 24 3.5 3.9 900 1.75 8.0 0 0 1 0 1
## 36 36 0 24 3.6 3.3 250 2.00 4.0 6 3 1 0 2
## 37 37 0 23 3.8 3.7 180 0.50 10.0 5 7 0 0 3
## 38 38 0 33 3.4 3.4 6000 1.50 8.0 5 6 2 0 3
## 39 39 0 23 2.8 3.2 950 2.00 37.0 10 5 0 0 2
## 40 40 0 31 3.8 3.5 1100 0.75 0.5 3 5 2 0 2
## 41 41 0 26 3.4 3.4 1300 1.20 0.0 8 2 0 0 3
## 42 42 0 28 2.0 3.0 360 0.25 10.0 8 3 0 0 1
## 43 43 1 24 3.8 3.9 1800 2.00 2.0 5 4 1 0 2
## 44 44 0 23 3.0 3.6 900 15.00 12.0 0 5 0 0 2
## 45 45 1 25 3.0 4.0 5000 5.00 1.5 0 4 0 0 3
## 46 46 1 24 3.0 3.5 300 1.00 10.0 5 5 0 0 1
## 47 47 1 27 3.0 3.8 2000 20.00 28.0 7 14 2 1 2
## 48 48 0 24 3.3 3.8 630 1.30 2.0 3 5 0 0 2
## 49 49 1 26 3.8 4.0 1200 1.00 0.0 4 3 1 0 1
## 50 50 1 27 3.0 4.0 580 2.00 5.0 15 1 2 0 1
## 51 51 0 32 3.0 3.0 2000 5.00 5.0 5 2 1 0 2
## 52 52 1 41 4.0 4.0 0 8.00 8.0 4 2 2 0 2
## 53 53 1 29 3.0 3.9 300 3.70 2.0 5 1 11 0 1
## 54 54 1 50 3.5 3.8 6 6.00 7.0 3 7 0 0 1
## 55 55 1 22 3.4 3.7 80 7.00 10.0 1 2 2 0 3
## 56 56 1 23 3.6 3.2 375 1.50 5.0 10 5 0 0 2
## 57 57 0 26 3.5 3.6 2000 0.30 16.0 8 3 0 0 1
## 58 58 0 30 3.0 3.0 1 1.10 1.0 4 3 0 0 3
## 59 59 1 23 3.0 3.0 112 0.50 15.0 3 3 0 0 3
## 60 60 1 22 3.4 3.0 650 4.00 8.0 16 7 1 0 3
## ideol relig abor affirm life
## 1 6 2 0 0 1
## 2 2 1 1 1 3
## 3 2 2 1 1 3
## 4 4 1 1 1 2
## 5 1 0 1 0 2
## 6 2 1 1 1 3
## 7 2 1 1 1 1
## 8 2 1 1 1 1
## 9 1 1 1 1 3
## 10 3 0 0 1 1
## 11 5 1 0 1 1
## 12 2 1 1 1 3
## 13 1 1 1 1 1
## 14 4 1 1 0 1
## 15 1 0 1 1 1
## 16 2 3 1 1 1
## 17 3 3 0 0 1
## 18 2 0 1 0 2
## 19 2 0 1 0 2
## 20 1 1 1 1 1
## 21 4 1 1 1 1
## 22 2 1 1 1 1
## 23 7 3 0 0 1
## 24 4 0 1 1 1
## 25 1 0 1 1 1
## 26 3 2 1 1 1
## 27 2 1 1 1 3
## 28 5 2 1 0 1
## 29 4 1 1 1 3
## 30 2 1 1 1 3
## 31 2 1 1 1 2
## 32 2 1 0 0 1
## 33 2 0 1 1 2
## 34 2 1 1 1 3
## 35 1 1 1 1 3
## 36 5 3 0 1 1
## 37 2 0 1 0 3
## 38 2 0 1 1 2
## 39 5 2 1 0 1
## 40 6 2 1 0 3
## 41 2 1 0 1 2
## 42 3 0 1 1 3
## 43 6 3 0 1 1
## 44 5 0 1 0 2
## 45 4 1 1 1 2
## 46 2 0 1 1 2
## 47 3 1 1 1 1
## 48 7 3 0 0 1
## 49 2 0 1 1 2
## 50 1 1 1 1 2
## 51 5 3 0 1 1
## 52 4 1 0 0 1
## 53 2 1 1 1 1
## 54 2 1 1 1 3
## 55 2 0 1 1 3
## 56 6 3 0 0 1
## 57 4 1 1 1 3
## 58 3 3 1 0 1
## 59 4 2 1 1 1
## 60 4 1 1 1 1fit<-glm(abor~factor(ideol)+factor(relig)+news+hsgpa+factor(gender),data=dat)
summary(fit)##
## Call:
## glm(formula = abor ~ factor(ideol) + factor(relig) + news + hsgpa +
## factor(gender), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.73147 -0.08132 0.05987 0.16384 0.55097
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 1.37187 0.33706 4.070 0.000179 ***
## factor(ideol)2 -0.10495 0.12715 -0.825 0.413333
## factor(ideol)3 -0.42461 0.18019 -2.357 0.022667 *
## factor(ideol)4 -0.08241 0.14941 -0.552 0.583840
## factor(ideol)5 -0.60965 0.19253 -3.166 0.002709 **
## factor(ideol)6 -0.85515 0.22784 -3.753 0.000479 ***
## factor(ideol)7 -0.91838 0.30165 -3.044 0.003812 **
## factor(relig)1 -0.13306 0.10294 -1.293 0.202462
## factor(relig)2 0.26572 0.16573 1.603 0.115561
## factor(relig)3 -0.17145 0.17603 -0.974 0.335053
## news 0.04207 0.01383 3.043 0.003830 **
## hsgpa -0.14310 0.09843 -1.454 0.152629
## factor(gender)1 0.03654 0.08732 0.418 0.677546
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09088174)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.2714 on 47 degrees of freedom
## AIC: 39.729
##
## Number of Fisher Scoring iterations: 2with(fit, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9204582#No evidence to suggest that at least one explanatory variable has an effect
library(car)## Loading required package: carData#anova(fit)
Anova(fit)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(ideol) 19.5732 6 0.003298 **
## factor(relig) 10.2705 3 0.016402 *
## news 9.2579 1 0.002345 **
## hsgpa 2.1137 1 0.145984
## factor(gender) 0.1751 1 0.675641
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#but we see significant variables while testing them individually##Mod1
mod1<-glm(abor~factor(gender) + age + hsgpa + cogpa + dhome + dres + tv + sport + news+ aids + factor(veg) + factor(affil) + factor(ideol) + relig + factor(affirm) + factor(life), data=dat)
summary(mod1)##
## Call:
## glm(formula = abor ~ factor(gender) + age + hsgpa + cogpa + dhome +
## dres + tv + sport + news + aids + factor(veg) + factor(affil) +
## factor(ideol) + relig + factor(affirm) + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.65734 -0.13528 0.01858 0.15510 0.52924
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.348e-01 7.298e-01 1.007 0.32074
## factor(gender)1 1.391e-01 1.178e-01 1.181 0.24552
## age 1.128e-03 6.346e-03 0.178 0.85992
## hsgpa -1.988e-01 1.217e-01 -1.634 0.11103
## cogpa 1.811e-01 1.767e-01 1.025 0.31208
## dhome 1.527e-05 3.008e-05 0.508 0.61486
## dres -2.249e-02 1.414e-02 -1.590 0.12060
## tv 6.202e-03 8.357e-03 0.742 0.46280
## sport -9.483e-03 1.397e-02 -0.679 0.50158
## news 5.434e-02 1.687e-02 3.221 0.00271 **
## aids 2.752e-03 2.035e-02 0.135 0.89317
## factor(veg)1 1.571e-01 1.442e-01 1.089 0.28326
## factor(affil)2 -2.908e-01 2.314e-01 -1.257 0.21698
## factor(affil)3 1.627e-01 1.285e-01 1.267 0.21343
## factor(ideol)2 -1.483e-01 1.469e-01 -1.010 0.31934
## factor(ideol)3 -2.655e-01 2.196e-01 -1.209 0.23455
## factor(ideol)4 -2.401e-02 1.659e-01 -0.145 0.88574
## factor(ideol)5 1.306e-02 3.122e-01 0.042 0.96685
## factor(ideol)6 -3.680e-01 3.443e-01 -1.069 0.29224
## factor(ideol)7 -6.227e-01 4.144e-01 -1.502 0.14171
## relig -1.159e-02 6.961e-02 -0.167 0.86867
## factor(affirm)1 -1.911e-01 1.461e-01 -1.308 0.19911
## factor(life)2 1.158e-01 1.399e-01 0.828 0.41325
## factor(life)3 2.526e-01 1.346e-01 1.876 0.06872 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1037835)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 3.7362 on 36 degrees of freedom
## AIC: 53.696
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod1, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9997273#No evidence to sugest that at least one explanatory variable has an effect
Anova(mod1)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 1.3937 1 0.237778
## age 0.0316 1 0.858920
## hsgpa 2.6692 1 0.102310
## cogpa 1.0512 1 0.305239
## dhome 0.2576 1 0.611764
## dres 2.5278 1 0.111854
## tv 0.5508 1 0.457989
## sport 0.4608 1 0.497243
## news 10.3750 1 0.001277 **
## aids 0.0183 1 0.892421
## factor(veg) 1.1866 1 0.276026
## factor(affil) 4.4672 2 0.107141
## factor(ideol) 6.6402 6 0.355403
## relig 0.0277 1 0.867740
## factor(affirm) 1.7112 1 0.190823
## factor(life) 3.5270 2 0.171440
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##Mod2 minus hgpa, age, aids
mod2<-glm(abor~factor(gender) + cogpa + dhome + dres + tv + sport + news+ factor(veg) + factor(affil) + factor(ideol) + relig + factor(affirm) + factor(life), data=dat)
summary(mod2)##
## Call:
## glm(formula = abor ~ factor(gender) + cogpa + dhome + dres +
## tv + sport + news + factor(veg) + factor(affil) + factor(ideol) +
## relig + factor(affirm) + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.66603 -0.11953 0.03325 0.14511 0.47705
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 4.345e-01 6.900e-01 0.630 0.53256
## factor(gender)1 1.411e-01 1.171e-01 1.204 0.23566
## cogpa 9.287e-02 1.661e-01 0.559 0.57928
## dhome 1.770e-05 2.827e-05 0.626 0.53484
## dres -2.156e-02 1.402e-02 -1.538 0.13222
## tv 9.825e-03 7.870e-03 1.249 0.21929
## sport -6.040e-03 1.314e-02 -0.460 0.64834
## news 4.635e-02 1.607e-02 2.884 0.00636 **
## factor(veg)1 1.608e-01 1.434e-01 1.121 0.26896
## factor(affil)2 -3.204e-01 2.287e-01 -1.401 0.16904
## factor(affil)3 1.109e-01 1.220e-01 0.909 0.36881
## factor(ideol)2 -1.865e-01 1.430e-01 -1.305 0.19964
## factor(ideol)3 -2.426e-01 2.143e-01 -1.132 0.26448
## factor(ideol)4 -6.670e-02 1.633e-01 -0.408 0.68517
## factor(ideol)5 2.544e-02 3.098e-01 0.082 0.93498
## factor(ideol)6 -3.166e-01 3.376e-01 -0.938 0.35413
## factor(ideol)7 -4.775e-01 3.976e-01 -1.201 0.23710
## relig -3.484e-02 6.766e-02 -0.515 0.60949
## factor(affirm)1 -1.547e-01 1.342e-01 -1.153 0.25574
## factor(life)2 1.349e-01 1.386e-01 0.973 0.33646
## factor(life)3 2.499e-01 1.334e-01 1.874 0.06844 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1031451)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.0227 on 39 degrees of freedom
## AIC: 52.129
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod2, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9986624Anova(mod2)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 1.4508 1 0.228406
## cogpa 0.3126 1 0.576083
## dhome 0.3921 1 0.531194
## dres 2.3642 1 0.124147
## tv 1.5588 1 0.211842
## sport 0.2112 1 0.645792
## news 8.3196 1 0.003922 **
## factor(veg) 1.2576 1 0.262103
## factor(affil) 3.8056 2 0.149153
## factor(ideol) 5.3819 6 0.495842
## relig 0.2652 1 0.606581
## factor(affirm) 1.3305 1 0.248721
## factor(life) 3.5707 2 0.167735
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##Mod3, removing cogpa,dhome ,sport, relig affirm(low LR Chisq coeff)
mod3<-glm(abor~factor(gender)+ dres + tv + news+ factor(veg) + factor(affil) + factor(ideol) + factor(life), data=dat)
summary(mod3)##
## Call:
## glm(formula = abor ~ factor(gender) + dres + tv + news + factor(veg) +
## factor(affil) + factor(ideol) + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.78035 -0.14188 -0.00088 0.17820 0.55734
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.584663 0.164217 3.560 0.000903 ***
## factor(gender)1 0.107326 0.097563 1.100 0.277286
## dres -0.011763 0.011996 -0.981 0.332180
## tv 0.011724 0.007316 1.603 0.116199
## news 0.040648 0.014738 2.758 0.008438 **
## factor(veg)1 0.098617 0.129112 0.764 0.449055
## factor(affil)2 -0.277932 0.221236 -1.256 0.215649
## factor(affil)3 0.113681 0.108564 1.047 0.300759
## factor(ideol)2 -0.186613 0.135002 -1.382 0.173860
## factor(ideol)3 -0.290674 0.189313 -1.535 0.131843
## factor(ideol)4 -0.092767 0.156464 -0.593 0.556286
## factor(ideol)5 -0.096819 0.272116 -0.356 0.723690
## factor(ideol)6 -0.320197 0.295748 -1.083 0.284853
## factor(ideol)7 -0.464702 0.340077 -1.366 0.178737
## factor(life)2 0.201442 0.122984 1.638 0.108565
## factor(life)3 0.255852 0.117379 2.180 0.034674 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09828232)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.3244 on 44 degrees of freedom
## AIC: 46.469
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod3, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9820214Anova(mod3)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 1.2101 1 0.271304
## dres 0.9615 1 0.326819
## tv 2.5680 1 0.109044
## news 7.6068 1 0.005815 **
## factor(veg) 0.5834 1 0.444979
## factor(affil) 3.7925 2 0.150134
## factor(ideol) 5.2323 6 0.514385
## factor(life) 5.3250 2 0.069774 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##Mod4 removing dres, factor(veg)
mod4<-glm(abor~factor(gender)+ tv + news + factor(affil) + factor(ideol) + factor(life), data=dat)
summary(mod4)##
## Call:
## glm(formula = abor ~ factor(gender) + tv + news + factor(affil) +
## factor(ideol) + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.76773 -0.14001 -0.00682 0.19247 0.56645
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.600319 0.161593 3.715 0.000549 ***
## factor(gender)1 0.085374 0.093649 0.912 0.366715
## tv 0.011750 0.007211 1.629 0.110052
## news 0.037584 0.014428 2.605 0.012336 *
## factor(affil)2 -0.339533 0.212011 -1.601 0.116115
## factor(affil)3 0.084365 0.105188 0.802 0.426654
## factor(ideol)2 -0.168005 0.133083 -1.262 0.213164
## factor(ideol)3 -0.282420 0.178785 -1.580 0.121037
## factor(ideol)4 -0.102440 0.155151 -0.660 0.512379
## factor(ideol)5 -0.101155 0.269326 -0.376 0.708952
## factor(ideol)6 -0.280805 0.289055 -0.971 0.336401
## factor(ideol)7 -0.445205 0.336071 -1.325 0.191803
## factor(life)2 0.175884 0.120491 1.460 0.151162
## factor(life)3 0.259779 0.115836 2.243 0.029780 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09717359)
##
## Null deviance: 10.183 on 59 degrees of freedom
## Residual deviance: 4.470 on 46 degrees of freedom
## AIC: 44.455
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod4, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9559661Anova(mod4)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 0.8311 1 0.361962
## tv 2.6550 1 0.103222
## news 6.7857 1 0.009189 **
## factor(affil) 4.1973 2 0.122625
## factor(ideol) 4.7292 6 0.578979
## factor(life) 5.2740 2 0.071577 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1##Mod5 #removing tv and gender
mod5<-glm(abor~ news + factor(affil) + factor(ideol) + factor(life), data=dat)
summary(mod5)##
## Call:
## glm(formula = abor ~ news + factor(affil) + factor(ideol) + factor(life),
## data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.82731 -0.11040 0.01182 0.20212 0.50215
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.75689 0.14014 5.401 2.03e-06 ***
## news 0.03910 0.01435 2.723 0.00898 **
## factor(affil)2 -0.23344 0.20939 -1.115 0.27048
## factor(affil)3 0.05094 0.10592 0.481 0.63272
## factor(ideol)2 -0.17468 0.13383 -1.305 0.19803
## factor(ideol)3 -0.30325 0.18208 -1.665 0.10233
## factor(ideol)4 -0.08751 0.15735 -0.556 0.58072
## factor(ideol)5 -0.19264 0.26294 -0.733 0.46733
## factor(ideol)6 -0.47335 0.27785 -1.704 0.09491 .
## factor(ideol)7 -0.66028 0.32148 -2.054 0.04546 *
## factor(life)2 0.11597 0.11792 0.983 0.33032
## factor(life)3 0.25228 0.11411 2.211 0.03185 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1011964)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.8574 on 48 degrees of freedom
## AIC: 45.442
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod5, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9143618Anova(mod5)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## news 7.4173 1 0.00646 **
## factor(affil) 2.0326 2 0.36193
## factor(ideol) 8.1991 6 0.22388
## factor(life) 4.8876 2 0.08683 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1AIC(mod1);AIC(mod2);AIC(mod3);AIC(mod4); AIC(mod5)## [1] 53.69617## [1] 52.12852## [1] 46.46866## [1] 44.45505## [1] 45.44248#Based on AIC, model 4 performs better2 and 3. Add to the mode any variables that were not included in step one but that are significant when adjusting for the variables in the model after step 2 ND Check fo possible interactions after step3
mod4.1<-glm(abor~factor(gender)+ tv + news + factor(affil) + ideol*relig + factor(life), data=dat)
summary(mod4.1)##
## Call:
## glm(formula = abor ~ factor(gender) + tv + news + factor(affil) +
## ideol * relig + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.80367 -0.12745 0.01797 0.17236 0.54464
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.47409 0.19304 2.456 0.0176 *
## factor(gender)1 0.07305 0.08510 0.858 0.3948
## tv 0.01140 0.00675 1.688 0.0977 .
## news 0.03340 0.01384 2.414 0.0196 *
## factor(affil)2 -0.26725 0.16845 -1.586 0.1191
## factor(affil)3 0.05899 0.10106 0.584 0.5621
## ideol 0.03085 0.05540 0.557 0.5801
## relig 0.07824 0.11150 0.702 0.4862
## factor(life)2 0.14293 0.12254 1.166 0.2491
## factor(life)3 0.20870 0.10791 1.934 0.0589 .
## ideol:relig -0.03954 0.02796 -1.414 0.1636
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09339537)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.5764 on 49 degrees of freedom
## AIC: 39.866
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod4.1, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.8471337Anova(mod4.1)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 0.7370 1 0.39064
## tv 2.8500 1 0.09137 .
## news 5.8269 1 0.01578 *
## factor(affil) 4.5802 2 0.10125
## ideol 0.2785 1 0.59767
## relig 1.0051 1 0.31608
## factor(life) 3.8686 2 0.14453
## ideol:relig 1.9999 1 0.15731
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod4.2<-glm(abor~factor(gender)+ tv*news + factor(affil) + ideol*relig + factor(life), data=dat)
summary(mod4.2)##
## Call:
## glm(formula = abor ~ factor(gender) + tv * news + factor(affil) +
## ideol * relig + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.79503 -0.12748 0.01785 0.17160 0.54090
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.4530245 0.2143689 2.113 0.0398 *
## factor(gender)1 0.0753681 0.0864879 0.871 0.3879
## tv 0.0131638 0.0101251 1.300 0.1998
## news 0.0370344 0.0207994 1.781 0.0813 .
## factor(affil)2 -0.2552826 0.1774852 -1.438 0.1568
## factor(affil)3 0.0624513 0.1030913 0.606 0.5475
## ideol 0.0310701 0.0559502 0.555 0.5813
## relig 0.0817095 0.1135462 0.720 0.4753
## factor(life)2 0.1427751 0.1237359 1.154 0.2543
## factor(life)3 0.2075239 0.1090787 1.903 0.0631 .
## tv:news -0.0003639 0.0015408 -0.236 0.8143
## ideol:relig -0.0405705 0.0285681 -1.420 0.1620
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09523045)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.5711 on 48 degrees of freedom
## AIC: 41.797
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod4.2, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.8979427Anova(mod4.2)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 0.7594 1 0.38352
## tv 2.7951 1 0.09455 .
## news 5.7146 1 0.01682 *
## factor(affil) 4.1159 2 0.12771
## ideol 0.2718 1 0.60214
## relig 0.9859 1 0.32075
## factor(life) 3.7490 2 0.15343
## tv:news 0.0558 1 0.81330
## ideol:relig 2.0168 1 0.15557
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1mod4.3<-glm(abor~factor(gender)+ tv + news + factor(affil) + ideol+ hsgpa+relig + factor(life), data=dat)
summary(mod4.3)##
## Call:
## glm(formula = abor ~ factor(gender) + tv + news + factor(affil) +
## ideol + hsgpa + relig + factor(life), data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.78254 -0.15670 0.02471 0.17001 0.61509
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 0.922570 0.349856 2.637 0.0112 *
## factor(gender)1 0.108508 0.086199 1.259 0.2141
## tv 0.011479 0.006901 1.663 0.1027
## news 0.035818 0.014339 2.498 0.0159 *
## factor(affil)2 -0.297773 0.167562 -1.777 0.0818 .
## factor(affil)3 0.100567 0.102666 0.980 0.3321
## ideol -0.023921 0.041482 -0.577 0.5668
## hsgpa -0.104562 0.096287 -1.086 0.2828
## relig -0.055152 0.057571 -0.958 0.3428
## factor(life)2 0.146273 0.123552 1.184 0.2422
## factor(life)3 0.226448 0.108344 2.090 0.0418 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.09492278)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 4.6512 on 49 degrees of freedom
## AIC: 40.84
##
## Number of Fisher Scoring iterations: 2#P-value
with(mod4.3, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.8529233Anova(mod4.3)## Analysis of Deviance Table (Type II tests)
##
## Response: abor
## LR Chisq Df Pr(>Chisq)
## factor(gender) 1.5846 1 0.20810
## tv 2.7663 1 0.09627 .
## news 6.2401 1 0.01249 *
## factor(affil) 7.5496 2 0.02294 *
## ideol 0.3325 1 0.56417
## hsgpa 1.1793 1 0.27751
## relig 0.9177 1 0.33807
## factor(life) 4.4571 2 0.10768
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1AIC(mod4);AIC(mod4.1);AIC(mod4.2);AIC(mod4.3)## [1] 44.45505## [1] 39.86636## [1] 41.79668## [1] 40.83968#Based on AIC, model with interaction between ideology and religion performs better than with individual, however last modelalso improves the original mod and does not include the complexity #colnames(dat)
fit1<-glm(abor~factor(gender) + age + hsgpa + cogpa + dhome + dres + tv + sport + news+ aids + factor(veg) + factor(affil) + factor(ideol) + relig + abor + factor(affirm) + factor(life), data=dat)## Warning in model.matrix.default(mt, mf, contrasts): the response appeared on the
## right-hand side and was dropped## Warning in model.matrix.default(mt, mf, contrasts): problem with term 15 in
## model.matrix: no columns are assignedsummary(fit1)##
## Call:
## glm(formula = abor ~ factor(gender) + age + hsgpa + cogpa + dhome +
## dres + tv + sport + news + aids + factor(veg) + factor(affil) +
## factor(ideol) + relig + abor + factor(affirm) + factor(life),
## data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.65734 -0.13528 0.01858 0.15510 0.52924
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 7.348e-01 7.298e-01 1.007 0.32074
## factor(gender)1 1.391e-01 1.178e-01 1.181 0.24552
## age 1.128e-03 6.346e-03 0.178 0.85992
## hsgpa -1.988e-01 1.217e-01 -1.634 0.11103
## cogpa 1.811e-01 1.767e-01 1.025 0.31208
## dhome 1.527e-05 3.008e-05 0.508 0.61486
## dres -2.249e-02 1.414e-02 -1.590 0.12060
## tv 6.202e-03 8.357e-03 0.742 0.46280
## sport -9.483e-03 1.397e-02 -0.679 0.50158
## news 5.434e-02 1.687e-02 3.221 0.00271 **
## aids 2.752e-03 2.035e-02 0.135 0.89317
## factor(veg)1 1.571e-01 1.442e-01 1.089 0.28326
## factor(affil)2 -2.908e-01 2.314e-01 -1.257 0.21698
## factor(affil)3 1.627e-01 1.285e-01 1.267 0.21343
## factor(ideol)2 -1.483e-01 1.469e-01 -1.010 0.31934
## factor(ideol)3 -2.655e-01 2.196e-01 -1.209 0.23455
## factor(ideol)4 -2.401e-02 1.659e-01 -0.145 0.88574
## factor(ideol)5 1.306e-02 3.122e-01 0.042 0.96685
## factor(ideol)6 -3.680e-01 3.443e-01 -1.069 0.29224
## factor(ideol)7 -6.227e-01 4.144e-01 -1.502 0.14171
## relig -1.159e-02 6.961e-02 -0.167 0.86867
## factor(affirm)1 -1.911e-01 1.461e-01 -1.308 0.19911
## factor(life)2 1.158e-01 1.399e-01 0.828 0.41325
## factor(life)3 2.526e-01 1.346e-01 1.876 0.06872 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1037835)
##
## Null deviance: 10.1833 on 59 degrees of freedom
## Residual deviance: 3.7362 on 36 degrees of freedom
## AIC: 53.696
##
## Number of Fisher Scoring iterations: 2library(MASS)
library(bestglm)## Warning: package 'bestglm' was built under R version 4.1.3## Loading required package: leaps## Warning: package 'leaps' was built under R version 4.1.3stepAIC(fit1)## Start: AIC=53.7
## abor ~ factor(gender) + age + hsgpa + cogpa + dhome + dres +
## tv + sport + news + aids + factor(veg) + factor(affil) +
## factor(ideol) + relig + abor + factor(affirm) + factor(life)## Warning in model.matrix.default(object, data = structure(list(abor = c(0L, : the
## response appeared on the right-hand side and was dropped## Warning in model.matrix.default(object, data = structure(list(abor = c(0L, :
## problem with term 15 in model.matrix: no columns are assigned##
## Step: AIC=53.7
## abor ~ factor(gender) + age + hsgpa + cogpa + dhome + dres +
## tv + sport + news + aids + factor(veg) + factor(affil) +
## factor(ideol) + relig + factor(affirm) + factor(life)
##
## Df Deviance AIC
## - aids 1 3.7381 51.727
## - relig 1 3.7391 51.742
## - age 1 3.7395 51.749
## - factor(ideol) 6 4.4254 51.853
## - dhome 1 3.7629 52.124
## - sport 1 3.7840 52.459
## - tv 1 3.7934 52.607
## - cogpa 1 3.8453 53.423
## - factor(veg) 1 3.8593 53.642
## <none> 3.7362 53.696
## - factor(gender) 1 3.8808 53.975
## - factor(affirm) 1 3.9138 54.483
## - factor(life) 2 4.1023 55.304
## - dres 1 3.9986 55.768
## - hsgpa 1 4.0132 55.988
## - factor(affil) 2 4.1998 56.715
## - news 1 4.8130 66.891
##
## Step: AIC=51.73
## abor ~ factor(gender) + age + hsgpa + cogpa + dhome + dres +
## tv + sport + news + factor(veg) + factor(affil) + factor(ideol) +
## relig + factor(affirm) + factor(life)
##
## Df Deviance AIC
## - relig 1 3.7407 49.769
## - age 1 3.7427 49.800
## - factor(ideol) 6 4.4258 49.859
## - dhome 1 3.7639 50.139
## - sport 1 3.7840 50.459
## - tv 1 3.7938 50.614
## - cogpa 1 3.8519 51.526
## - factor(veg) 1 3.8609 51.666
## <none> 3.7381 51.727
## - factor(gender) 1 3.8859 52.053
## - factor(affirm) 1 3.9174 52.537
## - factor(life) 2 4.1027 53.311
## - dres 1 3.9986 53.769
## - hsgpa 1 4.0200 54.089
## - factor(affil) 2 4.2029 54.758
## - news 1 4.8145 64.910
##
## Step: AIC=49.77
## abor ~ factor(gender) + age + hsgpa + cogpa + dhome + dres +
## tv + sport + news + factor(veg) + factor(affil) + factor(ideol) +
## factor(affirm) + factor(life)
##
## Df Deviance AIC
## - age 1 3.7451 47.839
## - dhome 1 3.7670 48.188
## - sport 1 3.7892 48.541
## - tv 1 3.7993 48.701
## - factor(ideol) 6 4.5110 49.004
## - factor(veg) 1 3.8611 49.668
## <none> 3.7407 49.769
## - cogpa 1 3.8769 49.914
## - factor(gender) 1 3.8888 50.099
## - factor(affirm) 1 3.9195 50.570
## - factor(life) 2 4.1187 51.544
## - dres 1 4.0053 51.869
## - hsgpa 1 4.0459 52.474
## - factor(affil) 2 4.2063 52.807
## - news 1 4.8394 63.219
##
## Step: AIC=47.84
## abor ~ factor(gender) + hsgpa + cogpa + dhome + dres + tv + sport +
## news + factor(veg) + factor(affil) + factor(ideol) + factor(affirm) +
## factor(life)
##
## Df Deviance AIC
## - dhome 1 3.7828 46.439
## - sport 1 3.7894 46.544
## - tv 1 3.8005 46.720
## - factor(ideol) 6 4.5115 47.009
## - factor(veg) 1 3.8698 47.804
## <none> 3.7451 47.839
## - cogpa 1 3.8806 47.972
## - factor(gender) 1 3.8945 48.185
## - factor(affirm) 1 3.9680 49.308
## - factor(life) 2 4.1267 49.661
## - dres 1 4.0099 49.939
## - hsgpa 1 4.0500 50.535
## - factor(affil) 2 4.2095 50.853
## - news 1 4.8422 61.254
##
## Step: AIC=46.44
## abor ~ factor(gender) + hsgpa + cogpa + dres + tv + sport + news +
## factor(veg) + factor(affil) + factor(ideol) + factor(affirm) +
## factor(life)
##
## Df Deviance AIC
## - sport 1 3.8389 45.324
## - tv 1 3.8422 45.374
## - factor(ideol) 6 4.5465 45.473
## - factor(veg) 1 3.8965 46.217
## - cogpa 1 3.9103 46.429
## <none> 3.7828 46.439
## - factor(gender) 1 3.9223 46.612
## - factor(affirm) 1 3.9780 47.459
## - dres 1 4.0294 48.229
## - factor(life) 2 4.1757 48.369
## - hsgpa 1 4.0919 49.152
## - factor(affil) 2 4.2942 50.048
## - news 1 4.8834 59.762
##
## Step: AIC=45.32
## abor ~ factor(gender) + hsgpa + cogpa + dres + tv + news + factor(veg) +
## factor(affil) + factor(ideol) + factor(affirm) + factor(life)
##
## Df Deviance AIC
## - factor(ideol) 6 4.5727 43.818
## - tv 1 3.8878 44.083
## <none> 3.8389 45.324
## - factor(veg) 1 3.9776 45.453
## - cogpa 1 4.0160 46.030
## - factor(affirm) 1 4.0247 46.158
## - factor(gender) 1 4.0311 46.254
## - dres 1 4.0717 46.856
## - hsgpa 1 4.1278 47.677
## - factor(affil) 2 4.3780 49.207
## - factor(life) 2 4.4269 49.873
## - news 1 4.8836 57.766
##
## Step: AIC=43.82
## abor ~ factor(gender) + hsgpa + cogpa + dres + tv + news + factor(veg) +
## factor(affil) + factor(affirm) + factor(life)
##
## Df Deviance AIC
## - factor(affirm) 1 4.5811 41.928
## - factor(veg) 1 4.5884 42.024
## - dres 1 4.6792 43.199
## - factor(gender) 1 4.6915 43.357
## <none> 4.5727 43.818
## - cogpa 1 4.7343 43.902
## - hsgpa 1 4.7492 44.090
## - tv 1 4.9844 46.991
## - factor(life) 2 5.1856 47.365
## - news 1 5.3037 50.716
## - factor(affil) 2 6.5882 61.729
##
## Step: AIC=41.93
## abor ~ factor(gender) + hsgpa + cogpa + dres + tv + news + factor(veg) +
## factor(affil) + factor(life)
##
## Df Deviance AIC
## - factor(veg) 1 4.5925 40.078
## - dres 1 4.6795 41.203
## - factor(gender) 1 4.6936 41.383
## <none> 4.5811 41.928
## - hsgpa 1 4.7496 42.095
## - cogpa 1 4.7507 42.109
## - tv 1 5.0067 45.258
## - factor(life) 2 5.1885 45.399
## - news 1 5.3238 48.944
## - factor(affil) 2 6.6746 60.511
##
## Step: AIC=40.08
## abor ~ factor(gender) + hsgpa + cogpa + dres + tv + news + factor(affil) +
## factor(life)
##
## Df Deviance AIC
## - dres 1 4.6835 39.255
## - factor(gender) 1 4.7103 39.597
## <none> 4.5925 40.078
## - cogpa 1 4.7538 40.149
## - hsgpa 1 4.7638 40.275
## - factor(life) 2 5.1899 43.414
## - tv 1 5.0271 43.503
## - news 1 5.3339 47.057
## - factor(affil) 2 6.7481 59.167
##
## Step: AIC=39.25
## abor ~ factor(gender) + hsgpa + cogpa + tv + news + factor(affil) +
## factor(life)
##
## Df Deviance AIC
## - factor(gender) 1 4.7660 38.302
## - cogpa 1 4.8186 38.961
## <none> 4.6835 39.255
## - hsgpa 1 4.8735 39.641
## - tv 1 5.0551 41.836
## - factor(life) 2 5.3322 43.038
## - news 1 5.3642 45.397
## - factor(affil) 2 6.9786 59.183
##
## Step: AIC=38.3
## abor ~ hsgpa + cogpa + tv + news + factor(affil) + factor(life)
##
## Df Deviance AIC
## <none> 4.7660 38.302
## - hsgpa 1 4.9337 38.378
## - cogpa 1 4.9664 38.774
## - factor(life) 2 5.3612 41.363
## - tv 1 5.2130 41.681
## - news 1 5.3959 43.750
## - factor(affil) 2 7.3733 60.484##
## Call: glm(formula = abor ~ hsgpa + cogpa + tv + news + factor(affil) +
## factor(life), data = dat)
##
## Coefficients:
## (Intercept) hsgpa cogpa tv news
## 0.26347 -0.13477 0.19231 0.01436 0.03641
## factor(affil)2 factor(affil)3 factor(life)2 factor(life)3
## -0.43348 0.13360 0.15716 0.26063
##
## Degrees of Freedom: 59 Total (i.e. Null); 51 Residual
## Null Deviance: 10.18
## Residual Deviance: 4.766 AIC: 38.3#colnames(dat)
fit.v<-glm(veg~factor(gender) + age + hsgpa + cogpa + dhome + dres + tv + sport + news+ aids + factor(abor) + factor(affil) + factor(ideol) + relig + abor + factor(affirm) + factor(life), data=dat)
summary(fit.v)##
## Call:
## glm(formula = veg ~ factor(gender) + age + hsgpa + cogpa + dhome +
## dres + tv + sport + news + aids + factor(abor) + factor(affil) +
## factor(ideol) + relig + abor + factor(affirm) + factor(life),
## data = dat)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -0.44584 -0.18744 -0.06622 0.10833 0.85964
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -4.945e-02 8.416e-01 -0.059 0.9535
## factor(gender)1 -1.523e-01 1.342e-01 -1.135 0.2637
## age 2.812e-03 7.205e-03 0.390 0.6986
## hsgpa 2.088e-02 1.434e-01 0.146 0.8850
## cogpa -7.586e-02 2.035e-01 -0.373 0.7114
## dhome -2.376e-05 3.410e-05 -0.697 0.4905
## dres 1.978e-02 1.631e-02 1.213 0.2332
## tv 9.874e-03 9.434e-03 1.047 0.3022
## sport -1.357e-02 1.583e-02 -0.858 0.3968
## news -2.175e-02 2.147e-02 -1.013 0.3179
## aids -2.105e-03 2.315e-02 -0.091 0.9280
## factor(abor)1 2.032e-01 1.865e-01 1.089 0.2833
## factor(affil)2 2.232e-02 2.689e-01 0.083 0.9343
## factor(affil)3 -1.413e-01 1.474e-01 -0.958 0.3442
## factor(ideol)2 1.337e-01 1.679e-01 0.796 0.4312
## factor(ideol)3 3.111e-01 2.494e-01 1.247 0.2204
## factor(ideol)4 -8.891e-02 1.882e-01 -0.473 0.6394
## factor(ideol)5 -3.136e-01 3.512e-01 -0.893 0.3778
## factor(ideol)6 1.871e-02 3.977e-01 0.047 0.9627
## factor(ideol)7 -2.990e-02 4.859e-01 -0.062 0.9513
## relig 8.629e-02 7.788e-02 1.108 0.2752
## abor NA NA NA NA
## factor(affirm)1 3.372e-01 1.605e-01 2.101 0.0427 *
## factor(life)2 -2.987e-02 1.606e-01 -0.186 0.8535
## factor(life)3 -1.703e-01 1.579e-01 -1.079 0.2879
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 0.1342379)
##
## Null deviance: 7.6500 on 59 degrees of freedom
## Residual deviance: 4.8326 on 36 degrees of freedom
## AIC: 69.135
##
## Number of Fisher Scoring iterations: 2#stepAIC(fit.v)
library(leaps)
with(fit.v, 1-pchisq(null.deviance-deviance , df.null -df.residual))## [1] 0.9999999#stepAIC(fit.v, direction = "forward")
stepAIC(fit.v)## Start: AIC=69.13
## veg ~ factor(gender) + age + hsgpa + cogpa + dhome + dres + tv +
## sport + news + aids + factor(abor) + factor(affil) + factor(ideol) +
## relig + abor + factor(affirm) + factor(life)
##
##
## Step: AIC=69.13
## veg ~ factor(gender) + age + hsgpa + cogpa + dhome + dres + tv +
## sport + news + aids + factor(abor) + factor(affil) + factor(ideol) +
## relig + factor(affirm) + factor(life)
##
## Df Deviance AIC
## - factor(affil) 2 4.9697 66.813
## - factor(ideol) 6 5.7022 67.063
## - aids 1 4.8337 67.148
## - hsgpa 1 4.8354 67.170
## - factor(life) 2 5.0008 67.187
## - cogpa 1 4.8512 67.366
## - age 1 4.8530 67.388
## - dhome 1 4.8977 67.938
## - sport 1 4.9313 68.348
## - news 1 4.9703 68.820
## - tv 1 4.9796 68.933
## - factor(abor) 1 4.9918 69.080
## <none> 4.8326 69.135
## - relig 1 4.9974 69.146
## - factor(gender) 1 5.0056 69.245
## - dres 1 5.0300 69.537
## - factor(affirm) 1 5.4251 74.074
##
## Step: AIC=66.81
## veg ~ factor(gender) + age + hsgpa + cogpa + dhome + dres + tv +
## sport + news + aids + factor(abor) + factor(ideol) + relig +
## factor(affirm) + factor(life)
##
## Df Deviance AIC
## - factor(life) 2 5.0626 63.925
## - factor(ideol) 6 5.8411 64.507
## - cogpa 1 4.9697 64.813
## - aids 1 4.9698 64.815
## - hsgpa 1 4.9729 64.853
## - age 1 5.0207 65.426
## - news 1 5.0569 65.857
## - factor(abor) 1 5.0726 66.044
## - sport 1 5.0782 66.109
## - dhome 1 5.1014 66.382
## - factor(gender) 1 5.1185 66.584
## - dres 1 5.1351 66.778
## <none> 4.9697 66.813
## - relig 1 5.1578 67.043
## - tv 1 5.1825 67.329
## - factor(affirm) 1 5.6383 72.387
##
## Step: AIC=63.93
## veg ~ factor(gender) + age + hsgpa + cogpa + dhome + dres + tv +
## sport + news + aids + factor(abor) + factor(ideol) + relig +
## factor(affirm)
##
## Df Deviance AIC
## - factor(ideol) 6 5.8770 60.875
## - aids 1 5.0628 61.927
## - hsgpa 1 5.0659 61.964
## - cogpa 1 5.0676 61.984
## - age 1 5.1014 62.383
## - sport 1 5.1232 62.638
## - factor(abor) 1 5.1351 62.778
## - news 1 5.1515 62.970
## - dhome 1 5.1605 63.074
## - factor(gender) 1 5.1763 63.257
## - tv 1 5.2316 63.895
## <none> 5.0626 63.925
## - dres 1 5.2412 64.005
## - relig 1 5.2664 64.293
## - factor(affirm) 1 5.6757 68.783
##
## Step: AIC=60.88
## veg ~ factor(gender) + age + hsgpa + cogpa + dhome + dres + tv +
## sport + news + aids + factor(abor) + relig + factor(affirm)
##
## Df Deviance AIC
## - cogpa 1 5.8842 58.948
## - news 1 5.8941 59.049
## - hsgpa 1 5.9009 59.118
## - aids 1 5.9104 59.215
## - factor(gender) 1 5.9469 59.584
## - factor(abor) 1 5.9493 59.608
## - tv 1 5.9526 59.641
## - dres 1 6.0719 60.832
## - age 1 6.0750 60.863
## <none> 5.8770 60.875
## - relig 1 6.0850 60.962
## - dhome 1 6.1300 61.403
## - sport 1 6.1365 61.467
## - factor(affirm) 1 6.7371 67.069
##
## Step: AIC=58.95
## veg ~ factor(gender) + age + hsgpa + dhome + dres + tv + sport +
## news + aids + factor(abor) + relig + factor(affirm)
##
## Df Deviance AIC
## - news 1 5.9028 57.138
## - hsgpa 1 5.9144 57.255
## - aids 1 5.9151 57.263
## - factor(abor) 1 5.9593 57.709
## - tv 1 5.9608 57.724
## - factor(gender) 1 5.9699 57.816
## - dres 1 6.0776 58.888
## <none> 5.8842 58.948
## - age 1 6.0886 58.997
## - dhome 1 6.1365 59.467
## - sport 1 6.1371 59.473
## - relig 1 6.1393 59.495
## - factor(affirm) 1 6.8197 65.801
##
## Step: AIC=57.14
## veg ~ factor(gender) + age + hsgpa + dhome + dres + tv + sport +
## aids + factor(abor) + relig + factor(affirm)
##
## Df Deviance AIC
## - aids 1 5.9348 55.462
## - hsgpa 1 5.9508 55.623
## - factor(abor) 1 5.9624 55.740
## - tv 1 5.9709 55.825
## - factor(gender) 1 5.9739 55.856
## - dres 1 6.0779 56.892
## <none> 5.9028 57.138
## - age 1 6.1158 57.265
## - dhome 1 6.1500 57.599
## - relig 1 6.1522 57.620
## - sport 1 6.1970 58.056
## - factor(affirm) 1 6.8224 63.825
##
## Step: AIC=55.46
## veg ~ factor(gender) + age + hsgpa + dhome + dres + tv + sport +
## factor(abor) + relig + factor(affirm)
##
## Df Deviance AIC
## - hsgpa 1 5.9877 53.995
## - tv 1 5.9910 54.028
## - factor(gender) 1 5.9938 54.055
## - factor(abor) 1 6.0000 54.118
## <none> 5.9348 55.462
## - dres 1 6.1375 55.477
## - relig 1 6.1786 55.877
## - sport 1 6.1999 56.084
## - dhome 1 6.2240 56.317
## - age 1 6.2279 56.354
## - factor(affirm) 1 6.8647 62.195
##
## Step: AIC=53.99
## veg ~ factor(gender) + age + dhome + dres + tv + sport + factor(abor) +
## relig + factor(affirm)
##
## Df Deviance AIC
## - factor(abor) 1 6.0519 52.634
## - factor(gender) 1 6.0652 52.766
## - tv 1 6.0850 52.961
## - dres 1 6.1743 53.835
## <none> 5.9877 53.995
## - sport 1 6.2300 54.374
## - relig 1 6.2426 54.496
## - age 1 6.2427 54.497
## - dhome 1 6.2661 54.721
## - factor(affirm) 1 6.9587 61.011
##
## Step: AIC=52.63
## veg ~ factor(gender) + age + dhome + dres + tv + sport + relig +
## factor(affirm)
##
## Df Deviance AIC
## - factor(gender) 1 6.1127 51.234
## - tv 1 6.1971 52.056
## - dres 1 6.2125 52.206
## - relig 1 6.2433 52.503
## <none> 6.0519 52.634
## - sport 1 6.3069 53.110
## - dhome 1 6.3115 53.154
## - age 1 6.3238 53.271
## - factor(affirm) 1 7.1059 60.267
##
## Step: AIC=51.23
## veg ~ age + dhome + dres + tv + sport + relig + factor(affirm)
##
## Df Deviance AIC
## - dres 1 6.2318 50.392
## - tv 1 6.2405 50.475
## - relig 1 6.2758 50.814
## - sport 1 6.3197 51.232
## <none> 6.1127 51.234
## - dhome 1 6.3538 51.555
## - age 1 6.3699 51.707
## - factor(affirm) 1 7.1332 58.497
##
## Step: AIC=50.39
## veg ~ age + dhome + tv + sport + relig + factor(affirm)
##
## Df Deviance AIC
## - relig 1 6.4044 50.031
## - tv 1 6.4337 50.305
## <none> 6.2318 50.392
## - dhome 1 6.4447 50.407
## - sport 1 6.4948 50.872
## - age 1 6.5177 51.083
## - factor(affirm) 1 7.1371 56.530
##
## Step: AIC=50.03
## veg ~ age + dhome + tv + sport + factor(affirm)
##
## Df Deviance AIC
## - tv 1 6.5765 49.622
## <none> 6.4044 50.031
## - age 1 6.6467 50.259
## - sport 1 6.6599 50.378
## - dhome 1 6.6611 50.389
## - factor(affirm) 1 7.1728 54.830
##
## Step: AIC=49.62
## veg ~ age + dhome + sport + factor(affirm)
##
## Df Deviance AIC
## - age 1 6.7631 49.301
## <none> 6.5765 49.622
## - sport 1 6.8068 49.687
## - dhome 1 6.8512 50.078
## - factor(affirm) 1 7.2944 53.838
##
## Step: AIC=49.3
## veg ~ dhome + sport + factor(affirm)
##
## Df Deviance AIC
## - sport 1 6.9470 48.910
## - dhome 1 6.9548 48.978
## <none> 6.7631 49.301
## - factor(affirm) 1 7.3308 52.137
##
## Step: AIC=48.91
## veg ~ dhome + factor(affirm)
##
## Df Deviance AIC
## - dhome 1 7.1163 48.355
## <none> 6.9470 48.910
## - factor(affirm) 1 7.5536 51.934
##
## Step: AIC=48.36
## veg ~ factor(affirm)
##
## Df Deviance AIC
## <none> 7.1163 48.355
## - factor(affirm) 1 7.6500 50.694##
## Call: glm(formula = veg ~ factor(affirm), data = dat)
##
## Coefficients:
## (Intercept) factor(affirm)1
## -3.727e-16 2.093e-01
##
## Degrees of Freedom: 59 Total (i.e. Null); 58 Residual
## Null Deviance: 7.65
## Residual Deviance: 7.116 AIC: 48.36#affirm = support
#affirmative action (1 = yes, 0 = no)
#veg ~ factor(affirm)5.6 Refer to the previous exercise. The data file also shows responses on whether a person smokes frequently. Software reports model −2 log-likelihood values of 1130.23 with only an intercept term, 1124.86 with also the main effect predictors, and 1119.87with also all the two-factor interactions. a. Write the model for each case and show that the numbers of parameters are 1, 5,and 11.
Refer to Table 2.9 on death penalty decisions. Fit a logistic model with the two race predictors. Conduct a residual analysis and interpret.
dp<-read.table("http://www.stat.ufl.edu/~aa/cat/data/DeathPenalty.dat", header=TRUE)
dp<-lapply(dp, as.factor)
dp<-as.data.frame(dp)
dp2<-lapply(dp, as.numeric)
dp2<-as.data.frame(dp2)
#fit<-glm(P~factor(D)+factor(V), data=dp)