Convenience function for simultaneously plotting two networks containing the same nodes.
PROCRUSTESnet(
qgraph_net1,
qgraph_net2,
type1 = c("ordinal", "interval", "ratio", "mspline"),
type2 = type1,
MDSadj1 = NULL,
MDSadj2 = NULL,
stressTxt = F,
congCoef = F,
repulse = F,
repulsion = 1,
mdsArgs = list(),
...
)an object of type qgraph
an object of type qgraph. Contains the same nodes as qgraph_net2
transformation function for first MDS, defaults to "ordinal"
transformation function for second MDS, defaults to the same as type1
to use a proximities matrix other than the adjacency matrix
stored in qgraph_net1, provide it in this argument
to use a proximities matrix other than the adjacency matrix
stored in qgraph_net2, provide it in this argument
logical. Print the stress value in the lower left corner of the plots?
logical. Print the congruence coefficient for the two layouts?
logical. Add a small repulsion force with wordcloud package to avoid node overlap?
scalar for the repulsion force. Larger values add more repulsion
additional arguments in list format passed to smacof::mds
additional arguments passed to qgraph
Each network's layout is determined by multidimensional scaling, and then the layouts are brought into a similar space by using the Procrustes algorithm.
A network plotted with multidimensional scaling can be interpreted based on the distances between nodes. Nodes close together represent closely associated nodes, whereas nodes that are far apart represent unassociated or negatively associated nodes.
The Procrustes algorithm brings the two layouts into a similar space through rotations and dilations that do not impact the fit of the MDS solutions. In this implementation, the second network is rotated and dilated to fit the first.
Jones, P. J., Mair, P., & McNally, R. J. (2018). Visualizing psychological networks: A tutorial in R. Frontiers in Psychology, 9, 1742. https://doi.org/10.3389/fpsyg.2018.01742