randomLCA (version 1.1-4)

genderrole: Gender Role Opinion Items

Description

Opinions collected on gender roles in a study by Felling et al (1987). This was originally published in Heinen (1996) and subsequently in Galindo Garre and Vermunt (2006).

Usage

genderrole

Format

A data frame with 16 observations on the following 5 variables.

Q1

Women's liberation sets women against men.

Q2

It's better for a wife not to have a job because that always poses problems in the household, especially if there are children.

Q3

The most natural situation occurs when the man is the breadwinner and the woman runs the household and takes care of the children.

Q4

It isn't really as important for a girl to get a good education as it is for a boy.

Q5

A woman is better suited to raise small children than a man.

Freq

Number of subjects

References

Felling, A., Peters, J., and Schreuder, O. (1987) Religion in Dutch society 85: Documentation of a national survey on religious and secular attitudes in 1985. Amsterdam: Steinmetz Archive.

Galindo Garre, F. and Vermunt, J.K. (2006) Avoiding boundary estimates in latent class analysis by Bayesian posterior mode estimation. Behaviormetrika, 33, 43--59.

Heinen, T. (1996) Latent Class and Discrete Latent Trait Models: Similarities and Differences.

Examples

# \donttest{
# standard latent class
genderrole.lca1 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, nclass = 1, cores = 1)
genderrole.lca2 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, cores = 1)
genderrole.lca3 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, nclass = 3, cores = 1)
# repeat with random effect with constant loading
# increase quadrature points and/or use higher penalty to obtain
# convergence
genderrole.lca1random <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	nclass = 1, random = TRUE, cores = 1)
genderrole.lca2random <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	random = TRUE, penalty = 0.1, quadpoints = 61, cores = 1)
genderrole.lca3random <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	nclass = 3, random = TRUE, penalty = 0.1, quadpoints = 61, cores = 1)
# improved BIC for 1 class random
print(c(BIC(genderrole.lca1), BIC(genderrole.lca2), BIC(genderrole.lca3)))
print(c(BIC(genderrole.lca1random), BIC(genderrole.lca2random), 
	BIC(genderrole.lca3random)))
# can also repeat fits without constant loading to give mixture of IRT models
genderrole.lca1random2 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	nclass = 1, random = TRUE, constload = FALSE, cores = 1)
genderrole.lca2random2 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	random = TRUE, constload = FALSE, quadpoints = 61, penalty = 0.1, cores = 1)
genderrole.lca3random2 <- randomLCA(genderrole[, 1:5], freq = genderrole$Freq, 
	nclass = 3, random = TRUE, constload = FALSE, quadpoints = 61, penalty = 0.1, cores = 1)
# no improvement in fit
print(c(BIC(genderrole.lca1random2), BIC(genderrole.lca2random2), 
	BIC(genderrole.lca3random2)))
# }