Fits a Rasch PCM using eRm, and conducts a PCA of residuals to get eigenvalues using psych::pca() and reports the top 5 values.

Proportion of explained variance is calculated using stats::prcomp().

Usage

RIpcmPCA(
  dfin,
  output = "table",
  fontsize = 15,
  maxiter = 5000,
  rotation = "oblimin"
)

Arguments

dfin

Dataframe with item data only

output

Defaults to "table", optional "dataframe"

fontsize

Set font size for table

maxiter

Maximum number of iterations. Increase if convergence not obtained.

rotation

Defaults to "oblimin"

Details

Note from ?psych::pca: The eigenvectors are rescaled by the sqrt of the eigenvalues to produce the component loadings more typical in factor analysis.

Possible rotations are: "none", "varimax", "quartimax", "promax", "oblimin", "simplimax", and "cluster".

See Chou & Wang (2010, DOI: 10.1177/0013164410379322) for a simulation study testing PCA eigenvalues across multiple conditions.