DriverDeaths: Single Vehicle Nighttime Driver Deaths in Utah

In glarma: Generalized Linear Autoregressive Moving Average Models

Description

This data set gives the number of single vehicle nighttime driver deaths in the state of Utah by month over the period August 1980 to July 1986, along with observations on a number of possible predictors. The aim of the study from which it was taken was to investigate the effect of the lowering of the legal blood alcohol concentration (BAC) while driving, from 0.1 to 0.08 units, and the simultaneous introduction of administrative license revocation. The time period for the observations is centred on the month of the intervention, August 1983.

Usage

data(DriverDeaths)

Format

A data frame containing the following columns:

[, 1]	Deaths	Number of single vehicle nighttime driver deaths monthly.
[, 2]	Intercept	A vector of ones, providing the intercept in the model.
[, 3]	ReducedBAC	Indicator of before or after lowering of legal blood alcohol level.0 for months prior to August 1983, 1 for months on or after August 1983.
[, 4]	FriSat	Number of Friday and Saturday nights in the month.
[, 5]	lnOMVDRate	Log of the number of other motor vehicle deaths per 100,000 of population.
[, 6]	Population	Adult population of the State of Utah.

Source

Debra H. Bernat, William T.M. Dunsmuir, and Alexander C. Wagenaar (2004) Effects of lowering the legal BAC to 0.08 on single-vehicle-nighttime fatal traffic crashes in 19 jurisdictions. Accident Analysis & Prevention, 36, 1089–1097.

Examples

### Model number of deaths
data(DriverDeaths)
y <- DriverDeaths[, "Deaths"]
X <- as.matrix(DriverDeaths[, 2:5])
Population <- DriverDeaths[, "Population"]

### Offset included
glarmamodOffset <- glarma(y, X, offset = log(Population/100000),
                          phiLags = c(12),
                          type = "Poi", method = "FS",
                          residuals = "Pearson", maxit = 100, grad = 1e-6)
print(summary(glarmamodOffset))
par(mfrow =c(3,2))
plot(glarmamodOffset)


### No offset included
glarmamodNoOffset <- glarma(y, X, phiLags = c(12),
                            type = "Poi", method = "FS",
                            residuals = "Pearson", maxit = 100, grad = 1e-6)
print(summary(glarmamodNoOffset))
par(mfrow=c(3,2))
plot(glarmamodNoOffset)

Example output

Call: glarma(y = y, X = X, offset = log(Population/1e+05), type = "Poi", 
    method = "FS", residuals = "Pearson", phiLags = c(12), maxit = 100, 
    grad = 1e-06)

Pearson Residuals:
    Min       1Q   Median       3Q      Max  
-1.7016  -0.7951   0.1192   0.5898   3.7190  

GLARMA Coefficients:
       Estimate Std.Error z-ratio Pr(>|z|)
phi_12 -0.03379   0.08994  -0.376    0.707

Linear Model Coefficients:
           Estimate Std.Error z-ratio Pr(>|z|)   
Intercept  -2.13517   0.80601  -2.649  0.00807 **
ReducedBAC -0.38594   0.15610  -2.472  0.01342 * 
FriSat      0.08662   0.09086   0.953  0.34043   
lnOMVDRate  0.55724   0.22975   2.425  0.01529 * 

    Null deviance: 95.499  on 71  degrees of freedom
Residual deviance: 67.556  on 67  degrees of freedom
AIC: 262.1928 

Number of Fisher Scoring iterations: 12

LRT and Wald Test:
Alternative hypothesis: model is a GLARMA process
Null hypothesis: model is a GLM with the same regression structure
          Statistic p-value
LR Test       0.160   0.689
Wald Test     0.141   0.707


Call: glarma(y = y, X = X, type = "Poi", method = "FS", residuals = "Pearson", 
    phiLags = c(12), maxit = 100, grad = 1e-06)

Pearson Residuals:
    Min       1Q   Median       3Q      Max  
-1.6890  -0.7972   0.1085   0.6110   3.6885  

GLARMA Coefficients:
       Estimate Std.Error z-ratio Pr(>|z|)
phi_12 -0.03347   0.09039   -0.37    0.711

Linear Model Coefficients:
           Estimate Std.Error z-ratio Pr(>|z|)  
Intercept   0.02327   0.80685   0.029   0.9770  
ReducedBAC -0.33442   0.15625  -2.140   0.0323 *
FriSat      0.08667   0.09093   0.953   0.3405  
lnOMVDRate  0.55127   0.22998   2.397   0.0165 *

    Null deviance: 93.273  on 71  degrees of freedom
Residual deviance: 67.359  on 67  degrees of freedom
AIC: 262.0302 

Number of Fisher Scoring iterations: 12

LRT and Wald Test:
Alternative hypothesis: model is a GLARMA process
Null hypothesis: model is a GLM with the same regression structure
          Statistic p-value
LR Test       0.157   0.692
Wald Test     0.137   0.711

glarma documentation built on May 2, 2019, 6:33 a.m.