NES {cmm}

Political Orientation in the US, three-wave panel study

Description

Data from the US National Election Studies. Three-wave panel study measuring political orientation on a seven-point scale. The data are tabulated in Bergsma, Croon, and Hagenaars (2009, 4.4).

Sections 4.2.1 and 4.3 in Bergsma, Croon and Hagenaars (2009).

Usage

data(NES)

Format

A data frame with 408 observations on the following variables.

T1

Political orientation at time 1 (ordered): 1 = Extremely liberal 2 = Liberal 3 = Slightly liberal" 4 = Moderate 5 = Slightly conservative 6 = Conservative 7 = Extremely conservative

T2

Political orientation at time 2 (ordered): see T1

T3

Political orientation at time 3 (ordered): see T1

Source

US National Election Studies.

References

Bergsma, W. P., Croon, M. A., & Hagenaars, J. A. P. (2009).Marginal models for dependent, clustered, and longitudinal categorical data.New York: Springer.

Examples in book: http://stats.lse.ac.uk/bergsma/cmm/R files/NES.R

Examples

data(NES)

####################################################################################
# Models for marginal homogeneity over time (Section 4.2.1)

# Marginal homogeneity : no change in political orientation over time

at <- MarginalMatrix(c("T1", "T2", "T3"), list(c("T1"), c("T2"), c("T3")), c(7,7,7));
bt1 <- ConstraintMatrix(c("T", "P"), list(c("T"), c("P")), c(3, 7));

model1 <- list(bt1, "identity", at);

start <- c(t(ftable(NES))) + .001;
pihat <- MarginalModelFit(NES, model1,
    MaxSteps = 1000, StartingPoint = start,
    ShowProgress = 100, MaxError = 1e-28,
    CoefficientDimensions = c(3,7), ShowCoefficients = TRUE,
    ShowParameters = FALSE, Labels = c("T", "P"));