EFA: Exploratory factor analysis (EFA)
In EFAtools: Fast and Flexible Implementations of Exploratory Factor Analysis Tools
EFA R Documentation
Exploratory factor analysis (EFA)
Description
This function does an EFA with either PAF, ML,
or ULS with or without subsequent rotation.
All arguments with default value NA can be left to default if type
is set to one of "EFAtools", "SPSS", or "psych". The respective specifications are
then handled according to the specified type (see details). For all rotations
except varimax and promax, the GPArotation package is needed.
Usage
EFA(
x,
n_factors,
N = NA,
method = c("PAF", "ML", "ULS"),
rotation = c("none", "varimax", "equamax", "quartimax", "geominT", "bentlerT",
"bifactorT", "promax", "oblimin", "quartimin", "simplimax", "bentlerQ", "geominQ",
"bifactorQ"),
type = c("EFAtools", "psych", "SPSS", "none"),
max_iter = NA,
init_comm = NA,
criterion = NA,
criterion_type = NA,
abs_eigen = NA,
use = c("pairwise.complete.obs", "all.obs", "complete.obs", "everything",
"na.or.complete"),
varimax_type = NA,
k = NA,
normalize = TRUE,
P_type = NA,
precision = 1e-05,
order_type = NA,
start_method = "psych",
cor_method = c("pearson", "spearman", "kendall"),
...
)
Arguments
x
data.frame or matrix. Dataframe or matrix of raw data or matrix with
correlations. If raw data is entered, the correlation matrix is found from the
data.
n_factors
numeric. Number of factors to extract.
N
numeric. The number of observations. Needs only be specified if a
correlation matrix is used. If input is a correlation matrix and N = NA
(default), not all fit indices can be computed.
method
character. One of "PAF", "ML", or "ULS" to use principal axis
factoring, maximum likelihood, or unweighted least squares (also called minres),
respectively, to fit the EFA.
rotation
character. Either perform no rotation ("none"; default),
an orthogonal rotation ("varimax", "equamax", "quartimax", "geominT",
"bentlerT", or "bifactorT"), or an oblique rotation ("promax", "oblimin",
"quartimin", "simplimax", "bentlerQ", "geominQ", or "bifactorQ").
type
character. If one of "EFAtools" (default), "psych", or "SPSS" is
used, and the following arguments with default NA are left with
NA, these implementations are executed according to the respective program
("psych" and "SPSS") or according to the best solution found in Grieder &
Steiner (2020; "EFAtools"). Individual properties can be adapted using one of
the three types and specifying some of the following arguments. If set to
"none" additional arguments must be specified depending on the method
and rotation used (see details).
max_iter
numeric. The maximum number of iterations to perform after which
the iterative PAF procedure is halted with a warning. If type is one of
"EFAtools", "SPSS", or "psych", this is automatically specified if max_iter is
left to be NA, but can be overridden by entering a number. Default is
NA.
init_comm
character. The method to estimate the initial communalities
in PAF. "smc" will use squared multiple correlations, "mac" will use
maximum absolute correlations, "unity" will use 1s (see details).
Default is NA.
criterion
numeric. The convergence criterion used for PAF.
If the change in communalities from one iteration to the next is smaller than
this criterion the solution is accepted and the procedure ends.
Default is NA.
criterion_type
character. Type of convergence criterion used for
PAF. "max_individual" selects the maximum change in any of the
communalities from one iteration to the next and tests it against the
specified criterion. This is also used by SPSS. "sum" takes the difference of
the sum of all communalities in one iteration and the sum of all communalities
in the next iteration and tests this against the criterion. This procedure is
used by the psych::fa function. Default is NA.
abs_eigen
logical. Which algorithm to use in the PAF
iterations. If FALSE, the loadings are computed from the eigenvalues. This is
also used by the psych::fa function. If TRUE the
loadings are computed with the absolute eigenvalues as done by SPSS.
Default is NA.
use
character. Passed to stats::cor if raw data
is given as input. Default is "pairwise.complete.obs".
varimax_type
character. The type of the varimax rotation performed.
If "svd", singular value decomposition is used, as stats::varimax does. If "kaiser", the varimax procedure performed in SPSS is used.
This is the original procedure from Kaiser (1958), but with slight alterations
in the varimax criterion (see details, and Grieder & Steiner, 2020). Default is NA.
k
numeric. Either the power used for computing the target matrix P in
the promax rotation or the number of 'close to zero loadings' for the simplimax
rotation (see GPArotation::GPFoblq). If left to
NA (default), the value for promax depends on the specified type.
For simplimax, nrow(L), where L is the matrix of unrotated loadings,
is used by default.
normalize
logical. If TRUE, a kaiser normalization is
performed before the specified rotation. Default is TRUE.
P_type
character. This specifies how the target
matrix P is computed in promax rotation. If "unnorm" it will use the
unnormalized target matrix as originally done in Hendrickson and White (1964).
This is also used in the psych and stats packages. If "norm" it will use the
normalized target matrix as used in SPSS. Default is NA.
precision
numeric. The tolerance for stopping in the rotation
procedure. Default is 10^-5 for all rotation methods.
order_type
character. How to order the factors. "eigen" will reorder
the factors according to the largest to lowest eigenvalues of the matrix of
rotated loadings. "ss_factors" will reorder the factors according to descending
sum of squared factor loadings per factor. Default is NA.
start_method
character. How to specify the starting values for the
optimization procedure for ML. Default is "psych" which takes the
starting values specified in psych::fa. "factanal" takes the
starting values specified in the stats::factanal function.
Solutions are very similar.
cor_method
character. Passed to stats::cor.
Default is "pearson".
...
Additional arguments passed to rotation functions from the GPArotation package (e.g., maxit for maximum number of iterations).
Details
There are two main ways to use this function. The easiest way is to
use it with a specified type (see above), which sets most of the other
arguments accordingly. Another way is to use it more flexibly by explicitly
specifying all arguments used and set type to "none" (see examples).
A mix of the two can also be done by specifying a type as well as
additional arguments. However, this will throw warnings to avoid unintentional
deviations from the implementations according to the specified type.
The type argument is evaluated for PAF and for all rotations (mainly
important for the varimax and promax rotations). The type-specific settings
for these functions are detailed below.
For PAF, the values of init_comm, criterion, criterion_type,
and abs_eigen depend on the type argument.
type = "EFAtools" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "sum",
abs_eigen = TRUE.
type = "psych" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "sum",
abs_eigen = FALSE.
type = "SPSS" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "max_individual",
abs_eigen = TRUE.
If SMCs fail, SPSS takes "mac". However, as SPSS takes absolute eigenvalues,
this is hardly ever the case. Psych, on the other hand, takes "unity" if SMCs
fail, but uses the Moore-Penrose Psudo Inverse of a matrix, thus, taking "unity"
is only necessary if negative eigenvalues occur afterwards in the iterative
PAF procedure. The EFAtools type setting combination was the best in terms of accuracy
and number of Heywood cases compared to all the
other setting combinations tested in simulation studies in Grieder & Steiner
(2020), which is why this type is used as a default here.
For varimax, the values of varimax_type and order_type depend on
the type argument.
type = "EFAtools" will use the following argument specification:
varimax_type = "kaiser", order_type = "eigen".
type = "psych" will use the following argument specification:
varimax_type = "svd", order_type = "eigen".
type = "SPSS" will use the following argument specification:
varimax_type = "kaiser", order_type = "ss_factors".
For promax, the values of P_type,
order_type, and k depend on the type argument.
type = "EFAtools" will use the following argument specification:
P_type = "norm", order_type = "eigen", k = 4.
type = "psych" will use the following argument specification:
P_type = "unnorm", order_type = "eigen", k = 4.
type = "SPSS" will use the following argument specification:
P_type = "norm", order_type = "ss_factors", k = 4.
The P_type argument can take two values, "unnorm" and "norm". It controls
which formula is used to compute the target matrix P in the promax rotation.
"unnorm" uses the formula from Hendrickson and White (1964), specifically:
P = abs(A^(k + 1)) / A,
where A is the unnormalized matrix containing varimax rotated loadings.
"SPSS" uses the normalized varimax rotated loadings. Specifically it used the
following formula, which can be found in the SPSS 23 and SPSS 27 Algorithms manuals:
P = abs(A / sqrt(rowSums(A^2))) ^(k + 1) * (sqrt(rowSums(A^2)) / A).
As for PAF, the EFAtools type setting combination for promax was the best
compared to the other setting combinations tested in simulation studies in
Grieder & Steiner (2020).
The varimax_type argument can take two values, "svd", and "kaiser". "svd" uses
singular value decomposition, by calling stats::varimax. "kaiser"
performs the varimax procedure as described in the SPSS 23 Algorithms manual and as described
by Kaiser (1958). However, there is a slight alteration in computing the varimax criterion, which
we found to better align with the results obtain from SPSS. Specifically, the original varimax
criterion as described in the SPSS 23 Algorithms manual is
sum(n*colSums(lambda ^ 4) - colSums(lambda ^ 2) ^ 2) / n ^ 2, where n is the
number of indicators, and lambda is the rotated loadings matrix. However, we found the following
to produce results more similar to those of SPSS:
sum(n*colSums(abs(lambda)) - colSums(lambda ^ 4) ^ 2) / n^2.
For all other rotations except varimax and promax, the type argument
only controls the order_type argument with the same values as stated
above for the varimax and promax rotations. For these other rotations, the
GPArotation package is needed. Additional arguments can also be
specified and will be passed to the respective GPArotation function
(e.g., maxit to change the maximum number of iterations for the rotation procedure).
The type argument has no effect on ULS and ML. For ULS, no additional
arguments are needed. For ML, an additional argument
start_method is needed to determine the starting values for the
optimization procedure. Default for this argument is "factanal" which takes
the starting values specified in the stats::factanal function.
Value
A list of class EFA containing (a subset of) the following:
orig_R
Original correlation matrix.
h2_init
Initial communality estimates from PAF.
h2
Final communality estimates from the unrotated solution.
orig_eigen
Eigen values of the original correlation matrix.
init_eigen
Initial eigenvalues, obtained from the correlation matrix
with the initial communality estimates as diagonal in PAF.
final_eigen
Eigenvalues obtained from the correlation matrix
with the final communality estimates as diagonal.
iter
The number of iterations needed for convergence.
convergence
Integer code for convergence as returned by
stats:optim (only for ML and ULS).
0 indicates successful completion.
unrot_loadings
Loading matrix containing the final unrotated loadings.
vars_accounted
Matrix of explained variances and sums of squared loadings. Based on the unrotated loadings.
fit_indices
For ML and ULS: Fit indices derived from the unrotated
factor loadings: Chi Square, including significance level, degrees of freedom
(df), Comparative Fit Index (CFI), Root Mean Square Error of Approximation
(RMSEA), including its 90% confidence interval, Akaike Information Criterion
(AIC), Bayesian Information Criterion (BIC), and the common part accounted
for (CAF) index as proposed by Lorenzo-Seva, Timmerman, & Kiers (2011).
For PAF, only the CAF and dfs are returned.
rot_loadings
Loading matrix containing the final rotated loadings
(pattern matrix).
Phi
The factor intercorrelations (only for oblique rotations).
Structure
The structure matrix (only for oblique rotations).
rotmat
The rotation matrix.
vars_accounted_rot
Matrix of explained variances and sums of squared
loadings. Based on rotated loadings and, for oblique rotations, the factor
intercorrelations.
settings
A list of the settings used.
Source
Grieder, S., & Steiner, M.D. (2020). Algorithmic Jingle Jungle:
A Comparison of Implementations of Principal Axis Factoring and Promax Rotation
in R and SPSS. Manuscript in Preparation.
Hendrickson, A. E., & White, P. O. (1964). Promax: A quick method for
rotation to oblique simple structure. British Journal of Statistical Psychology,
17 , 65–70. doi: 10.1111/j.2044-8317.1964.tb00244.x
Lorenzo-Seva, U., Timmerman, M. E., & Kiers, H. A. L. (2011). The
Hull Method for Selecting the Number of Common Factors, Multivariate Behavioral
Research, 46, 340-364, doi: 10.1080/00273171.2011.564527
Kaiser, H. F. (1958). The varimax criterion for analytic rotation in
factor analysis. Psychometrika, 23, 187–200. doi: 10.1007/BF02289233
Examples
# A type EFAtools (as presented in Steiner and Grieder, 2020) EFA
EFAtools_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "PAF", rotation = "none")
# A type SPSS EFA to mimick the SPSS implementation (this will throw a warning,
# see below)
SPSS_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "SPSS", method = "PAF", rotation = "none")
# A type psych EFA to mimick the psych::fa() implementation
psych_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "psych", method = "PAF", rotation = "none")
# Use ML instead of PAF with type EFAtools
EFAtools_ML <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "ML", rotation = "none")
# Use oblimin rotation instead of no rotation with type EFAtools
EFAtools_oblim <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "PAF", rotation = "oblimin")
# Do a PAF without rotation without specifying a type, so the arguments
# can be flexibly specified (this is only recommended if you know what your
# doing)
PAF_none <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "none", method = "PAF", rotation = "none",
max_iter = 500, init_comm = "mac", criterion = 1e-4,
criterion_type = "sum", abs_eigen = FALSE)
# Add a promax rotation
PAF_pro <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "none", method = "PAF", rotation = "promax",
max_iter = 500, init_comm = "mac", criterion = 1e-4,
criterion_type = "sum", abs_eigen = FALSE, k = 3,
P_type = "unnorm", precision= 1e-5, order_type = "eigen",
varimax_type = "svd")
EFAtools documentation built on Jan. 6, 2023, 5:16 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| EFA | R Documentation |
Exploratory factor analysis (EFA)
Description
This function does an EFA with either PAF, ML,
or ULS with or without subsequent rotation.
All arguments with default value NA can be left to default if type
is set to one of "EFAtools", "SPSS", or "psych". The respective specifications are
then handled according to the specified type (see details). For all rotations
except varimax and promax, the GPArotation package is needed.
Usage
EFA(
x,
n_factors,
N = NA,
method = c("PAF", "ML", "ULS"),
rotation = c("none", "varimax", "equamax", "quartimax", "geominT", "bentlerT",
"bifactorT", "promax", "oblimin", "quartimin", "simplimax", "bentlerQ", "geominQ",
"bifactorQ"),
type = c("EFAtools", "psych", "SPSS", "none"),
max_iter = NA,
init_comm = NA,
criterion = NA,
criterion_type = NA,
abs_eigen = NA,
use = c("pairwise.complete.obs", "all.obs", "complete.obs", "everything",
"na.or.complete"),
varimax_type = NA,
k = NA,
normalize = TRUE,
P_type = NA,
precision = 1e-05,
order_type = NA,
start_method = "psych",
cor_method = c("pearson", "spearman", "kendall"),
...
)
Arguments
x |
data.frame or matrix. Dataframe or matrix of raw data or matrix with correlations. If raw data is entered, the correlation matrix is found from the data. |
n_factors |
numeric. Number of factors to extract. |
N |
numeric. The number of observations. Needs only be specified if a
correlation matrix is used. If input is a correlation matrix and |
method |
character. One of "PAF", "ML", or "ULS" to use principal axis factoring, maximum likelihood, or unweighted least squares (also called minres), respectively, to fit the EFA. |
rotation |
character. Either perform no rotation ("none"; default), an orthogonal rotation ("varimax", "equamax", "quartimax", "geominT", "bentlerT", or "bifactorT"), or an oblique rotation ("promax", "oblimin", "quartimin", "simplimax", "bentlerQ", "geominQ", or "bifactorQ"). |
type |
character. If one of "EFAtools" (default), "psych", or "SPSS" is
used, and the following arguments with default NA are left with
NA, these implementations are executed according to the respective program
("psych" and "SPSS") or according to the best solution found in Grieder &
Steiner (2020; "EFAtools"). Individual properties can be adapted using one of
the three types and specifying some of the following arguments. If set to
"none" additional arguments must be specified depending on the |
max_iter |
numeric. The maximum number of iterations to perform after which
the iterative PAF procedure is halted with a warning. If |
init_comm |
character. The method to estimate the initial communalities
in |
criterion |
numeric. The convergence criterion used for PAF.
If the change in communalities from one iteration to the next is smaller than
this criterion the solution is accepted and the procedure ends.
Default is |
criterion_type |
character. Type of convergence criterion used for
PAF. "max_individual" selects the maximum change in any of the
communalities from one iteration to the next and tests it against the
specified criterion. This is also used by SPSS. "sum" takes the difference of
the sum of all communalities in one iteration and the sum of all communalities
in the next iteration and tests this against the criterion. This procedure is
used by the |
abs_eigen |
logical. Which algorithm to use in the PAF
iterations. If FALSE, the loadings are computed from the eigenvalues. This is
also used by the |
use |
character. Passed to |
varimax_type |
character. The type of the varimax rotation performed.
If "svd", singular value decomposition is used, as stats::varimax does. If "kaiser", the varimax procedure performed in SPSS is used.
This is the original procedure from Kaiser (1958), but with slight alterations
in the varimax criterion (see details, and Grieder & Steiner, 2020). Default is |
k |
numeric. Either the power used for computing the target matrix P in
the promax rotation or the number of 'close to zero loadings' for the simplimax
rotation (see |
normalize |
logical. If |
P_type |
character. This specifies how the target
matrix P is computed in promax rotation. If "unnorm" it will use the
unnormalized target matrix as originally done in Hendrickson and White (1964).
This is also used in the psych and stats packages. If "norm" it will use the
normalized target matrix as used in SPSS. Default is |
precision |
numeric. The tolerance for stopping in the rotation procedure. Default is 10^-5 for all rotation methods. |
order_type |
character. How to order the factors. "eigen" will reorder
the factors according to the largest to lowest eigenvalues of the matrix of
rotated loadings. "ss_factors" will reorder the factors according to descending
sum of squared factor loadings per factor. Default is |
start_method |
character. How to specify the starting values for the optimization procedure for ML. Default is "psych" which takes the starting values specified in psych::fa. "factanal" takes the starting values specified in the stats::factanal function. Solutions are very similar. |
cor_method |
character. Passed to |
... |
Additional arguments passed to rotation functions from the |
Details
There are two main ways to use this function. The easiest way is to
use it with a specified type (see above), which sets most of the other
arguments accordingly. Another way is to use it more flexibly by explicitly
specifying all arguments used and set type to "none" (see examples).
A mix of the two can also be done by specifying a type as well as
additional arguments. However, this will throw warnings to avoid unintentional
deviations from the implementations according to the specified type.
The type argument is evaluated for PAF and for all rotations (mainly
important for the varimax and promax rotations). The type-specific settings
for these functions are detailed below.
For PAF, the values of init_comm, criterion, criterion_type,
and abs_eigen depend on the type argument.
type = "EFAtools" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "sum",
abs_eigen = TRUE.
type = "psych" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "sum",
abs_eigen = FALSE.
type = "SPSS" will use the following argument specification:
init_comm = "smc", criterion = .001, criterion_type = "max_individual",
abs_eigen = TRUE.
If SMCs fail, SPSS takes "mac". However, as SPSS takes absolute eigenvalues, this is hardly ever the case. Psych, on the other hand, takes "unity" if SMCs fail, but uses the Moore-Penrose Psudo Inverse of a matrix, thus, taking "unity" is only necessary if negative eigenvalues occur afterwards in the iterative PAF procedure. The EFAtools type setting combination was the best in terms of accuracy and number of Heywood cases compared to all the other setting combinations tested in simulation studies in Grieder & Steiner (2020), which is why this type is used as a default here.
For varimax, the values of varimax_type and order_type depend on
the type argument.
type = "EFAtools" will use the following argument specification:
varimax_type = "kaiser", order_type = "eigen".
type = "psych" will use the following argument specification:
varimax_type = "svd", order_type = "eigen".
type = "SPSS" will use the following argument specification:
varimax_type = "kaiser", order_type = "ss_factors".
For promax, the values of P_type,
order_type, and k depend on the type argument.
type = "EFAtools" will use the following argument specification:
P_type = "norm", order_type = "eigen", k = 4.
type = "psych" will use the following argument specification:
P_type = "unnorm", order_type = "eigen", k = 4.
type = "SPSS" will use the following argument specification:
P_type = "norm", order_type = "ss_factors", k = 4.
The P_type argument can take two values, "unnorm" and "norm". It controls
which formula is used to compute the target matrix P in the promax rotation.
"unnorm" uses the formula from Hendrickson and White (1964), specifically:
P = abs(A^(k + 1)) / A,
where A is the unnormalized matrix containing varimax rotated loadings.
"SPSS" uses the normalized varimax rotated loadings. Specifically it used the
following formula, which can be found in the SPSS 23 and SPSS 27 Algorithms manuals:
P = abs(A / sqrt(rowSums(A^2))) ^(k + 1) * (sqrt(rowSums(A^2)) / A).
As for PAF, the EFAtools type setting combination for promax was the best
compared to the other setting combinations tested in simulation studies in
Grieder & Steiner (2020).
The varimax_type argument can take two values, "svd", and "kaiser". "svd" uses
singular value decomposition, by calling stats::varimax. "kaiser"
performs the varimax procedure as described in the SPSS 23 Algorithms manual and as described
by Kaiser (1958). However, there is a slight alteration in computing the varimax criterion, which
we found to better align with the results obtain from SPSS. Specifically, the original varimax
criterion as described in the SPSS 23 Algorithms manual is
sum(n*colSums(lambda ^ 4) - colSums(lambda ^ 2) ^ 2) / n ^ 2, where n is the
number of indicators, and lambda is the rotated loadings matrix. However, we found the following
to produce results more similar to those of SPSS:
sum(n*colSums(abs(lambda)) - colSums(lambda ^ 4) ^ 2) / n^2.
For all other rotations except varimax and promax, the type argument
only controls the order_type argument with the same values as stated
above for the varimax and promax rotations. For these other rotations, the
GPArotation package is needed. Additional arguments can also be
specified and will be passed to the respective GPArotation function
(e.g., maxit to change the maximum number of iterations for the rotation procedure).
The type argument has no effect on ULS and ML. For ULS, no additional
arguments are needed. For ML, an additional argument
start_method is needed to determine the starting values for the
optimization procedure. Default for this argument is "factanal" which takes
the starting values specified in the stats::factanal function.
Value
A list of class EFA containing (a subset of) the following:
orig_R |
Original correlation matrix. |
h2_init |
Initial communality estimates from PAF. |
h2 |
Final communality estimates from the unrotated solution. |
orig_eigen |
Eigen values of the original correlation matrix. |
init_eigen |
Initial eigenvalues, obtained from the correlation matrix with the initial communality estimates as diagonal in PAF. |
final_eigen |
Eigenvalues obtained from the correlation matrix with the final communality estimates as diagonal. |
iter |
The number of iterations needed for convergence. |
convergence |
Integer code for convergence as returned by
|
unrot_loadings |
Loading matrix containing the final unrotated loadings. |
vars_accounted |
Matrix of explained variances and sums of squared loadings. Based on the unrotated loadings. |
fit_indices |
For ML and ULS: Fit indices derived from the unrotated factor loadings: Chi Square, including significance level, degrees of freedom (df), Comparative Fit Index (CFI), Root Mean Square Error of Approximation (RMSEA), including its 90% confidence interval, Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), and the common part accounted for (CAF) index as proposed by Lorenzo-Seva, Timmerman, & Kiers (2011). For PAF, only the CAF and dfs are returned. |
rot_loadings |
Loading matrix containing the final rotated loadings (pattern matrix). |
Phi |
The factor intercorrelations (only for oblique rotations). |
Structure |
The structure matrix (only for oblique rotations). |
rotmat |
The rotation matrix. |
vars_accounted_rot |
Matrix of explained variances and sums of squared loadings. Based on rotated loadings and, for oblique rotations, the factor intercorrelations. |
settings |
A list of the settings used. |
Source
Grieder, S., & Steiner, M.D. (2020). Algorithmic Jingle Jungle: A Comparison of Implementations of Principal Axis Factoring and Promax Rotation in R and SPSS. Manuscript in Preparation.
Hendrickson, A. E., & White, P. O. (1964). Promax: A quick method for rotation to oblique simple structure. British Journal of Statistical Psychology, 17 , 65–70. doi: 10.1111/j.2044-8317.1964.tb00244.x
Lorenzo-Seva, U., Timmerman, M. E., & Kiers, H. A. L. (2011). The Hull Method for Selecting the Number of Common Factors, Multivariate Behavioral Research, 46, 340-364, doi: 10.1080/00273171.2011.564527
Kaiser, H. F. (1958). The varimax criterion for analytic rotation in factor analysis. Psychometrika, 23, 187–200. doi: 10.1007/BF02289233
Examples
# A type EFAtools (as presented in Steiner and Grieder, 2020) EFA
EFAtools_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "PAF", rotation = "none")
# A type SPSS EFA to mimick the SPSS implementation (this will throw a warning,
# see below)
SPSS_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "SPSS", method = "PAF", rotation = "none")
# A type psych EFA to mimick the psych::fa() implementation
psych_PAF <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "psych", method = "PAF", rotation = "none")
# Use ML instead of PAF with type EFAtools
EFAtools_ML <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "ML", rotation = "none")
# Use oblimin rotation instead of no rotation with type EFAtools
EFAtools_oblim <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "EFAtools", method = "PAF", rotation = "oblimin")
# Do a PAF without rotation without specifying a type, so the arguments
# can be flexibly specified (this is only recommended if you know what your
# doing)
PAF_none <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "none", method = "PAF", rotation = "none",
max_iter = 500, init_comm = "mac", criterion = 1e-4,
criterion_type = "sum", abs_eigen = FALSE)
# Add a promax rotation
PAF_pro <- EFA(test_models$baseline$cormat, n_factors = 3, N = 500,
type = "none", method = "PAF", rotation = "promax",
max_iter = 500, init_comm = "mac", criterion = 1e-4,
criterion_type = "sum", abs_eigen = FALSE, k = 3,
P_type = "unnorm", precision= 1e-5, order_type = "eigen",
varimax_type = "svd")
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.