EFAtools (version 1.1.0)

print.efa_compare: Print and format an efa_compare object

Description

print() shows a summarised output of the efa_compare() function: the mean (with its range), median, and root mean squared distance (RMSE) of the differences, the number of decimals to which all numbers agree, the minimum number of decimals provided, and (for matrices) the number of differing indicator-to-factor correspondences, followed (optionally) by the table of elementwise differences. format() assembles the same report and returns it as a character vector; print() is cat(format(x), sep = "\n"). The lines follow the active console theme, so they are plain when colours are disabled (for example when captured into a file or stripped with cli::ansi_strip()).

Usage

# S3 method for efa_compare
print(x, ...)
# S3 method for efa_compare
format(
  x,
  digits = NULL,
  m_red = NULL,
  range_red = NULL,
  round_red = NULL,
  print_diff = NULL,
  ...
)

Value

print() returns its argument x invisibly. format() returns a character vector with the report lines.

Arguments

x

An object of class efa_compare (output from efa_compare()).

...

Passed from print() to format(); not otherwise used.

digits, m_red, range_red, round_red, print_diff

Display controls, documented in efa_compare(). Each defaults to NULL, meaning the value efa_compare() recorded in x$settings is used; supplying one overrides it for this call only, so the comparison need not be recomputed to change the printed report.

Details

The line reporting the minimum number of decimals provided is shown only when it carries information: two ordinary double matrices carry the full double precision, for which the count is uninformative and the line is omitted.

The summary statistics are absolute differences, so they carry no direction. The elementwise differences are signed, and the table is headed by the direction of the subtraction, named with the x_labels recorded by efa_compare() ("x" and "y" by default): a negative cell means the first solution is the lower of the two there.

Examples

# A type SPSS EFA to mimick the SPSS implementation
EFA_SPSS_5 <- efa_fit(IDS2_R, n_factors = 5,
                      estimate_control = estimate_control(type = "SPSS"),
                      rotate_control = rotate_control(type = "SPSS"))

# A type psych EFA to mimick the psych::fa() implementation
EFA_psych_5 <- efa_fit(IDS2_R, n_factors = 5,
                       estimate_control = estimate_control(type = "psych"),
                       rotate_control = rotate_control(type = "psych"))

# compare the two
comp <- efa_compare(EFA_SPSS_5$unrot_loadings, EFA_psych_5$unrot_loadings,
                    x_labels = c("SPSS", "psych"))
comp

# format() returns the same lines as a character vector:
writeLines(format(comp))

# the display settings can be changed without recomputing the comparison:
print(comp, digits = 2, print_diff = FALSE)