Arbuthnot’s data on the sex ratio of births in London was examined in Chapter 3. Use a binomial logistic regression model to assess whether the proportion of male births varied with the variables Year, Plague, and Mortality in the Arbuthnot data set. Produce effect plots for the terms in this model. What do you conclude? Compare your results with a null model whereas you only have the intercept of the curve.
data(Arbuthnot, package="HistData")
View(Arbuthnot)
library(effects)
## Loading required package: carData
## lattice theme set by effectsTheme()
## See ?effectsTheme for details.
library(car)
##
## Attaching package: 'car'
## The following objects are masked from 'package:carData':
##
## Guyer, UN, Vocab
library(vcdExtra)
## Loading required package: vcd
## Loading required package: grid
## Loading required package: gnm
##
## Attaching package: 'vcdExtra'
## The following object is masked from 'package:car':
##
## Burt
## The following object is masked from 'package:carData':
##
## Burt
summary(Arbuthnot)
## Year Males Females Plague
## Min. :1629 Min. :2890 Min. :2722 Min. : 0.00
## 1st Qu.:1649 1st Qu.:4759 1st Qu.:4457 1st Qu.: 0.00
## Median :1670 Median :6073 Median :5718 Median : 3.00
## Mean :1670 Mean :5907 Mean :5535 Mean : 1240.70
## 3rd Qu.:1690 3rd Qu.:7576 3rd Qu.:7150 3rd Qu.: 22.25
## Max. :1710 Max. :8426 Max. :7779 Max. :68596.00
## Mortality Ratio Total
## Min. : 8393 Min. :1.011 Min. : 5.612
## 1st Qu.:12739 1st Qu.:1.048 1st Qu.: 9.199
## Median :17867 Median :1.065 Median :11.813
## Mean :17816 Mean :1.071 Mean :11.442
## 3rd Qu.:21030 3rd Qu.:1.088 3rd Qu.:14.723
## Max. :97306 Max. :1.156 Max. :16.145
Arbuthnot$Males<-as.numeric(Arbuthnot$Males>"None")
arbuthnot.logistic<-glm(formula = Males ~ Year + Plague + Mortality, family = binomial, data = Arbuthnot)
summary(arbuthnot.logistic)
##
## Call:
## glm(formula = Males ~ Year + Plague + Mortality, family = binomial,
## data = Arbuthnot)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -2.409e-06 -2.409e-06 -2.409e-06 -2.409e-06 -2.409e-06
##
## Coefficients:
## Estimate Std. Error z value Pr(>|z|)
## (Intercept) -2.657e+01 6.228e+06 0 1
## Year -6.945e-16 3.888e+03 0 1
## Plague -1.218e-18 2.196e+01 0 1
## Mortality 8.413e-19 1.794e+01 0 1
##
## (Dispersion parameter for binomial family taken to be 1)
##
## Null deviance: 0.0000e+00 on 81 degrees of freedom
## Residual deviance: 4.7573e-10 on 78 degrees of freedom
## AIC: 8
##
## Number of Fisher Scoring iterations: 25
plot(jitter(Year,2)~Males,data=Arbuthnot,ylab="Probability(year)")
coef(arbuthnot.logistic)
## (Intercept) Year Plague Mortality
## -2.656607e+01 -6.944593e-16 -1.217462e-18 8.412722e-19
exp(coef(arbuthnot.logistic))
## (Intercept) Year Plague Mortality
## 2.900701e-12 1.000000e+00 1.000000e+00 1.000000e+00
exp(10*coef(arbuthnot.logistic)[2])
## Year
## 1