show.correlations {affinitymatrix}

Print correlations of matching factors with matching and outcome variables

Description

This function returns a list of plots, one for each of the first ndims orthogonal sorting dimension. In the k-th plot, the correlation between a man's observed matching variable and the man's k-th matching factor is plotted on the x-axis; the correlation between a woman's observed matching variable and the woman's k-th matching factor is plotted on the y-axis. In addition, the user can supply additional variables stored in the matrix Z that were not previously used in the estimation ("outcome variables"). The function prints the correlation between the outcome variable and the man's k-th matching factor on the x-axis, while the correlation between the outcome variable and the woman's k-th matching factor is on the y-axis.

Usage

show.correlations(
  res,
  Z = matrix(0, nrow = N, ncol = 0),
  labels_x = paste0("Trait ", 1:Kx),
  labels_y = paste0("Trait ", 1:Ky),
  labels_z = if (Kz > 0) paste0("Outcome ", 1:Kz) else c(),
  ndims = min(Kx, Ky, 10),
  pr = 0.02,
  color_arrows = c("black", "red"),
  size_arrows = 0.5,
  font_labels = c("bold", "italic"),
  label_x_axis = "First partner",
  label_y_axis = "Second partner"
)

Arguments

res

A list corresponding to the output of estimate.affinity.matrix, estimate.affinity.matrix.lowrank or estimate.affinity.matrix.unipartite.

Z

A matrix Z with additional variables that were not previously used in the estimation. The i-th row of Z must contain information on the couple formed by the i-th row of X and the i-th row of Y, so that nrow(Z)=nrow(X). Defaults to an empty matrix: Z is optional.

labels_x

A vector of strings indicating the names of men's matching variables. Defaults to "Trait k" for every k matching variable.

labels_y

A vector of strings indicating the names of women's matching variables. Defaults to "Trait k" for every k matching variable.

labels_z

A vector of strings indicating the names of the outcome variables. Defaults to "Outcome k" for every k outcome variable.

ndims

An integer indicating the number of orthogonal matching dimensions that will be plotted. The function plots the first \ codendims dimensions. Defaults to all dimensions unless the latter are more than 10, in which case only the first 10 are plotted.

pr

A probability indicating the two-tailed significance level required for a matching or outcome variable to be displayed in a plot. In order to avoid having too many variables plotted at the same time, the function only selects those whose correlation with the matching factor is significantly different from zero (in a two-tailed test) at the pr level. Defaults to 0.02 and can be set to 1 to print all variables.

color_arrows

A string or a vector of strings containing color names for the arrows. All matching variables are assigned the first color given in the vector, while all outcome variables are assigned the second color. See ggplot. Defaults to "black" and "red" respectively.

size_arrows

A positive real number or a vector containing the size of the arrows. All matching variables are assigned the first size given in the vector, while all outcome variables are assigned the second size. See ggplot. Defaults to 0.5 for both.

font_labels

A string or a vector of strings containing font types for the labels. All matching variables are assigned the first font type given in the vector, while all outcome variables are assigned the second font type. See ggplot. Defaults to "bold" and "italic" respectively.

label_x_axis

A string containing a root for all x-axis names in different plots. Defaults to "First partner".

label_y_axis

A string containing a root for all y-axis names in different plots. Defaults to "Second partner".

Value

The function returns a list of ndims plots created with ggplot.

See Also

Chiappori, Pierre-Andre, Edoardo Ciscato, and Carla Guerriero. "Analyzing matching patterns in marriage: theory and application to Italian data." HCEO Working Paper no. 2020-080 (2020).