pca.flex: Perform PCA

In mrajter/rajter.flex: A Flextable Based Statistics Package

Perform PCA

Description

Function is based on psych::principal function and in line with Field's Discovering statistics using R. The PCA is performed with listwise deletion and is based on correlation matrix. Due to mystic differences in algorthms, loadings obtained by oblimin rotation are marginally different from those obtained in SPSS...I still don't know why

Usage

pca.flex(
  data,
  form,
  n_factors = "KG",
  rotation = "oblimin",
  title = "",
  sing.solution = T,
  sorted = T,
  supress = 0.3,
  option = r.flex.opts
)

Arguments

data	data frame
form	formula with variables on rhs. E.g. 1 ~ var1 + var2 + var3
n_factors	number of components to extract. There are several methods exact number. Useful for forced single component and experienting "KG" - Kaiser-Guttman criteria (lambda>1) "MAP" - Velicer's Map test "VSS1" - very simple structure complexity 1 "VSS2" - very simple structure complexity 2 "Parallel" - parallel tests "ScreeAC" - Acceleration factor Scree test "ScreeOC" - Optimal coordinates Scree test
rotation	Rotation to use. Options are: "none", "varimax", "oblimin". Default is "oblimin".
title	Title of analysis. Will be determined if omitted
sing.solution	Should forced single component solution also be presented (default=T)
sorted	should the loadings be sorted in table (default=T)
supress	suppress coefficients below threshold (default=0.3)
option	inherited from r.flex.opts

Value

list with elements

• type - table type - used for inserting in word document

• title - used for table title. Can be set manually or automatically

• preqs - prerequisites - Bartlett and KMO

• df - results as data.frame

• table - flextable with results

• rotation - used rotation

• sing.solution - is there a forced single component solution

• extraction - how are components extracted (exact number or criteria) - see nfac.flex function for details

mrajter/rajter.flex documentation built on July 30, 2024, 1:31 a.m.