VoterTurnout: Voter Turnout in U.S. Presidential Elections 1984
In hetprobit: Heteroscedastic Probit Regression Models

Description Usage Format Details Source References See Also Examples

Census data from the U.S. Presidential Elections 1984

1	data("VoterTurnout")

A data frame with 98857 observations on the following 6 variables.

vote: factor. Did the respondent vote?
education: Years of education of the respondent
age: Age of the respondent
south: factor. Is the respondent from the South?
govelection: factor. Were gubernatorial elections held?
closing: How many days before the election has the registration been closed?

Nagler (1991) first analyzed the data using the standard probit model. In 1994 he introduced the skewed logit model (a.k.a scobit-model) and Altman and McDonald (2003) replicated this study with focus on numerical accuracy.

Supplementary material to Altman and McDonald (2003). https://doi.org/10.1093/pan/mpg016

Altman M., McDonald M.P. (2003) Replication with Attention to Numerical Accuracy. Political Analysis, 11, 302–307.

Nagler J. (1991) The Effects of Registration Laws and Education on U.S. Voter Turnout. The American Political Science Review, 85(4), 1393–1405.

Nagler J. (1994) Scobit: An Alternative Estimator to Logit and Probit. Political Science, 38(1), 230–55.

hetprobit

data("VoterTurnout")

## homoscedastic probit model
## regressors are the same as in Table 3 in Nagler (1991)
m0 <- glm(vote ~ age + I(age^2) + south + govelection + 
                (education + I(education^2)) * closing, data = VoterTurnout,
                family = binomial(link = "probit"))
summary(m0)

## Not run: 
## heteroscedastic probit model 
## main effects in the mean model and one regressor in the scale model 
m1 <- hetprobit(vote ~ education + age + south + govelection + closing | 
                education, data = VoterTurnout) 

summary(m1)

## End(Not run)