CountEigen.PA: Number of observed eigenvalues that exceed a given set of...

In pcaPA: Parallel Analysis for Ordinal and Numeric Data using Polychoric and Pearson Correlations with S3 Classes

Description

Counts the number of observed eigenvalues that exceed the given percentiles.

Usage

CountEigen.PA(PA, percentiles = NULL)

Arguments

PA	an object of class "PA".
percentiles	the percentiles that ought to be plotted, defaults to those in the object.

Value

A named numeric vector indicating the number of eigenvalues that are greater than the eigenvalues distribution percentiles under independence.

Author(s)

Carlos A. Arias [e-mail] and Victor H. Cervantes [e-mail]

See Also

PA, print.PA, coef.PA, Check.PA, plot.PA, quantile.PA

Examples

# # Run Parallel Analysis for binary data conforming to the Rasch model
data(simRaschData)
binaryRaschPA <- PA(simRaschData, percentiles = c(0.95, 0.99), nReplicates = 200,
                    type = "binary", algorithm = "polychoric")
print(binaryRaschPA)

# # Number of retained factors
nComponents <- CountEigen.PA(binaryRaschPA, percentiles = .99)
nComponents["p99"]

# # Run Parallel Analysis for binary data conforming to the 2PL model
data(sim2plData)
binary2plPA <- PA(sim2plData, percentiles = c(0.95, 0.99), nReplicates = 200,
                  type = "binary", algorithm = "polychoric")
print(binary2plPA)

# # Number of retained factors
nComponents <- CountEigen.PA(binary2plPA, percentiles = .99)
nComponents["p99"]

pcaPA documentation built on May 29, 2017, 6:53 p.m.