Skip to contents

Computes the between- and within-community strength of each variable for each community

Usage

net.loads(
  A,
  wc,
  loading.method = c("original", "revised"),
  scaling = 2,
  rotation = NULL,
  ...
)

Arguments

A

Network matrix, data frame, or EGA object

wc

Numeric or character vector (length = ncol(A)). A vector of community assignments. If input into A is an EGA object, then wc is automatically detected

loading.method

Character (length = 1). Sets network loading calculation based on implementation described in "original" (Christensen & Golino, 2021) or the "revised" (Christensen et al., 2024) implementation. Defaults to "revised"

scaling

Numeric (length = 1). Scaling factor for the magnitude of the "experimental" network loadings. Defaults to 2. 10 makes loadings roughly the size of factor loadings when correlations between factors are orthogonal

rotation

Character. A rotation to use to obtain a simpler structure. For a list of rotations, see rotations for options. Defaults to NULL or no rotation. By setting a rotation, scores estimation will be based on the rotated loadings rather than unrotated loadings

...

Additional arguments to pass on to rotations

Value

Returns a list containing:

unstd

A matrix of the unstandardized within- and between-community strength values for each node

std

A matrix of the standardized within- and between-community strength values for each node

rotated

NULL if rotation = NULL; otherwise, a list containing the rotated standardized network loadings (loadings) and correlations between dimensions (Phi) from the rotation

Details

Simulation studies have demonstrated that a node's strength centrality is roughly equivalent to factor loadings (Christensen & Golino, 2021; Hallquist, Wright, & Molenaar, 2019). Hallquist and colleagues (2019) found that node strength represented a combination of dominant and cross-factor loadings. This function computes each node's strength within each specified dimension, providing a rough equivalent to factor loadings (including cross-loadings; Christensen & Golino, 2021).

References

Original implementation and simulation
Christensen, A. P., & Golino, H. (2021). On the equivalency of factor and network loadings. Behavior Research Methods, 53, 1563-1580.

Demonstration of node strength similarity to CFA loadings
Hallquist, M., Wright, A. C. G., & Molenaar, P. C. M. (2019). Problems with centrality measures in psychopathology symptom networks: Why network psychometrics cannot escape psychometric theory. Multivariate Behavioral Research, 1-25.

Revised network loadings
Christensen, A. P., Golino, H., Abad, F. J., & Garrido, L. E. (2024). Revised network loadings. PsyArXiv.

Author

Alexander P. Christensen <alexpaulchristensen@gmail.com> and Hudson Golino <hfg9s at virginia.edu>

Examples

# Load data
wmt <- wmt2[,7:24]

# Estimate EGA
ega.wmt <- EGA(
  data = wmt,
  plot.EGA = FALSE # No plot for CRAN checks
)

# Network loadings
net.loads(ega.wmt)
#> The default 'loading.method' has changed to "revised" in {EGAnet} version >= 2.0.7.
#> 
#>  For the previous default (version <= 2.0.6), use `loading.method = "original"`
#> Loading Method: Revised
#> 
#>       1     2    
#> wmt2  0.608 0.145
#> wmt1  0.402 0.106
#> wmt3  0.344 0.204
#> wmt5  0.318 0.234
#> wmt4  0.298 0.231
#> wmt9        0.523
#> wmt7        0.461
#> wmt15       0.441
#> wmt14       0.433
#> wmt6  0.177 0.429
#> wmt16       0.392
#> wmt8        0.391
#> wmt10  0.21 0.368
#> wmt12       0.334
#> wmt18       0.326
#> wmt13       0.304
#> wmt17 0.122 0.299
#> wmt11       0.277
#> Standardized loadings >= |0.10| are displayed. To change this 'minimum', use `print(net.loads_object, minimum = 0.10)`