Family demographers have long been interested in the effects of socioeconomic variables on the likelihood of marrying and how they differ for men and women. You will run analyses that allow you to address this issue.
1) Produce histograms of income for men and women. Describe the shape of the distributions.
2) Now create a variable for logged income and produce histograms for this variable. Use code like this: fmarriagemen$logincome<-log(fmarriagemen$inctot). Describe the shape of the distributions.
3) Now create a variable for the square root of income and produce histograms for this variable. Use code like this: fmarriagemen$sqrtincome<-sqrt(fmarriagemen$inctot). Describe the shape of the distributions.
4) Examine the association between income and fmarry for men and women separately using confidence intervals of fmarry by income. See for example the confidence intervals for the age gaps between partners by age group in Part 1 for sample code. To do this, you will need to first put income into categories using code like this but with more breaks:
fmarriagemen$inccat<-cut(fmarriagemen$inctot, breaks = c(0, 50000, 100000, Inf), right = FALSE)
5) Describe how income and marriage are associated for men and women. Does gender appear to substantially moderate the association income and marriage? What is your evidence?
6) Run a crosstab of degree and fmarry for men and women separately. Also conduct chi-square tests in R. Describe the association between these two variables for men and women.
7) Run logistic regression models of marriage for men and women to examine the effects of educational attainment. Interpret any coefficients that are significant (with the exception of the intercept).
8) Run logistic regression models of marriage for men and women to examine the effects of income. Interpret the effects of income based on the models for men and women.
9) Run logistic regression models of marriage for men and women to examine the effects of logged income. Does the relative importance of income for men and women depend on how income is specified (i.e., logged versus not)? Do you prefer the results from models with logged income or those from the models with income not logged? Justify your choice.
10) What role does income play in terms of the association education and marriage (i.e., a confounding, mediating, or moderating role)? Explain your evidence for this.