dataReduction: Factor analysis or principal component analysis

In rosetta: Parallel Use of Statistical Packages in Teaching

Factor analysis or principal component analysis

Description

This is a wrapper for the psych functions psych::pca() and psych::fa() to produce output that it similar to the output produced by jamovi.

Usage

factorAnalysis(
  data,
  nfactors,
  items = names(data),
  rotate = "oblimin",
  covar = FALSE,
  na.rm = TRUE,
  kaiser = 1,
  loadings = TRUE,
  summary = FALSE,
  correlations = FALSE,
  modelFit = FALSE,
  eigenValues = FALSE,
  screePlot = FALSE,
  residuals = FALSE,
  itemLabels = items,
  colorLoadings = FALSE,
  fm = "minres",
  digits = 2,
  headingLevel = 3,
  ...
)

principalComponentAnalysis(
  data,
  items,
  nfactors,
  rotate = "oblimin",
  covar = FALSE,
  na.rm = TRUE,
  kaiser = 1,
  loadings = TRUE,
  summary = FALSE,
  correlations = FALSE,
  eigenValues = FALSE,
  screePlot = FALSE,
  residuals = FALSE,
  itemLabels = items,
  colorLoadings = FALSE,
  digits = 2,
  headingLevel = 3,
  ...
)

rosettaDataReduction_partial(
  x,
  digits = x$input$digits,
  headingLevel = x$input$headingLevel,
  echoPartial = FALSE,
  partialFile = NULL,
  quiet = TRUE,
  ...
)

## S3 method for class 'rosettaDataReduction'
knit_print(
  x,
  digits = x$input$digits,
  headingLevel = x$input$headingLevel,
  echoPartial = FALSE,
  partialFile = NULL,
  quiet = TRUE,
  ...
)

## S3 method for class 'rosettaDataReduction'
print(
  x,
  digits = x$input$digits,
  headingLevel = x$input$headingLevel,
  forceKnitrOutput = FALSE,
  ...
)

Arguments

data	The data frame that contains the items.
nfactors	The number of factors to extract, or 'eigen' to extract all factors with an eigen value higher than the number specified in kaiser. In the future, parallel can be specified here to extract the number of factors suggested by parallel analysis.
items	The items to analyse; if not specified, all variables in data will be used.
rotate	Which rotation to use; see psych::fa() for all options. The most common options are 'none' to not rotate at all; 'varimax' for an orthogonal rotation (assuming/imposing that the components or factors are not correlated); or 'oblimin' for an oblique rotation (allowing the components/factors to correlate).
covar	Whether to analyse the correlation matrix (FALSE) or the covariance matrix (TRUE).
na.rm	Whether to first remove all cases with missing values.
kaiser	The minimum eigenvalue when applying the Kaiser criterion (see nfactors).
loadings	Whether to display the component or factor loadings.
summary	Whether to display the factor or component summary.
correlations	Whether to display the correlations between factors of components.
modelFit	Whether to display the model fit Only for EFA).
eigenValues	Whether to display the eigen values.
screePlot	Whether to display the scree plot.
residuals	Whether to display the matrix with residuals.
itemLabels	Optionally, labels to use for the items (optionally, named, with the names corresponding to the items; otherwise, the order of the labels has to match the order of the items)
colorLoadings	Whether, when producing an Rmd partial (i.e. when calling the command while knitting) to colour the cells using kableExtra::kable_styling().
fm	The method to use for the factor analysis: 'fm' for Minimum Residuals; 'ml' for Maximum Likelihood; and 'pa' for Principal Factor.
digits	The number of digits to round to.
headingLevel	The number of hashes to print in front of the headings when printing while knitting
...	Any additional arguments are passed to psych::fa(), psych::pca(), to the default print method by the print method, and to rmdpartials::partial() when knitting an RMarkdown partial.
x	The object to print.
echoPartial	Whether to show the executed code in the R Markdown partial (TRUE) or not (FALSE).
partialFile	This can be used to specify a custom partial file. The file will have object x available.
quiet	Passed on to knitr::knit() whether it should b chatty (FALSE) or quiet (TRUE).
forceKnitrOutput	Force knitr output.

Details

The code in these functions uses parts of the code in jamovi, written by Jonathon Love and Ravi Selker.

Value

An object with the object resulting from the call to the psych functions and some extracted information that will be printed.

Examples

### Load example dataset
data("pp15", package="rosetta");

### Get variable names with expected
### effects of a high dose of MDMA
items <-
  grep(
    "highDose_AttBeliefs_",
    names(pp15),
    value=TRUE
  );

### Do a factor analysis
rosetta::factorAnalysis(
  data = pp15,
  items = items,
  nfactors = "eigen",
  scree = TRUE
);

if (FALSE) {
  ### To get more output, show the
  ### output as Rmd Partial in the viewer,
  ### and color/size the factor loadings
  rosetta::rosettaDataReduction_partial(
    rosetta::factorAnalysis(
      data = pp15,
      items = items,
      nfactors = "eigen",
      summary = TRUE,
      correlations = TRUE,
      colorLoadings = TRUE
    )
  );
}

rosetta documentation built on March 7, 2023, 7:40 p.m.