| Type: | Package |
| Title: | Network Structural Equation Modeling |
| Version: | 0.4 |
| Maintainer: | Zhiyong Zhang <johnnyzhz@gmail.com> |
| Description: | Several methods have been developed to integrate structural equation modeling techniques with network data analysis to examine the relationship between network and non-network data. Both node-based and edge-based information can be extracted from the network data to be used as observed variables in structural equation modeling. To facilitate the application of these methods, model specification can be performed in the familiar syntax of the 'lavaan' package, ensuring ease of use for researchers. Technical details and examples can be found at https://bigsem.psychstat.org. |
| License: | GPL-2 | GPL-3 [expanded from: GPL] |
| Imports: | sna, influential, lavaan, network, igraph, methods |
| Depends: | latentnet, R (≥ 3.5) |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 7.2.3 |
| NeedsCompilation: | no |
| Packaged: | 2025-05-09 14:11:15 UTC; ziqianxu |
| Author: | Zhiyong Zhang 
[aut, cre],
Ziqian Xu [aut] |
| Repository: | CRAN |
| Date/Publication: | 2025-05-09 14:50:10 UTC |
Friendship network data
Description
A dataset with friendship network data of 165 undergraduate students in China, and also contains information extroversion and depression scores
Usage
data("friend_data")
Format
A list with two elements
- network
A list containing a 165 x 165 adjacency matrix indicating friendship ties.
- non_network
A dataframe with 165 rows and 11 columns containing variables on depression from the Patient Health Questionnaire and extroversion from the mini-IPIP scale.
Source
<https://bigsem.org/>
Examples
data("friend_data")
length(friend_data$network)
head(friend_data$nonnetwork)
Calculate a mediation effect from a networksem model
Description
Calculate a mediation effect from a networksem model
Usage
path.networksem(res, predictor, mediator, outcome)
Arguments
res |
a networksem output file |
predictor |
a character string of the predictor variable |
mediator |
a character string of the mediator variable |
outcome |
a character string of the outcome variable |
Value
a target path, associated estimates, and z-score
Fit a Structural Equation Model (SEM) with both network and non-network data by incorporating node-level network statistics as variables.
Description
Fit a Structural Equation Model (SEM) with both network and non-network data by incorporating node-level network statistics as variables.
Usage
sem.net(
model = NULL,
data = NULL,
netstats = NULL,
ordered = NULL,
sampling.weights = NULL,
data.rescale = FALSE,
netstats.rescale = FALSE,
group = NULL,
cluster = NULL,
constraints = "",
WLS.V = NULL,
NACOV = NULL,
netstats.options = NULL,
...
)
Arguments
model |
A model string specified in lavaan model syntax that includes relationships among the network and non-network variables. |
data |
A list containing the data. The list has two named components, "network" and "nonnetwork"; "network" is a list of named adjacency matrices for the network data, and "nonnetwork" is the dataframe of non-network covariates. |
netstats |
A user-specified list of network statistics to be calculated and used in the SEM. Available options include "degree", "betweenness", "closeness", "evcent", "stresscent", and "infocent" from the "sna" package and "ivi", "hubness.score", "spreading.score" and "clusterRank" from the "influential" package. |
ordered |
Parameter same as "ordered" in the lavaan sem() function; whether to treat data as ordinal. |
sampling.weights |
Parameter same as "sampling.weights" in the lavaan sem() function; whether to apply weights to data. |
data.rescale |
TRUE or FALSE, whether to rescale the whole dataset (with restructured network and nonnetwork data) to have mean 0 and standard deviation 1 when fitting it to SEM, default to FALSE. |
netstats.rescale |
TRUE or FALSE, whether to rescale the network statistics to have mean 0 and standard deviation 1, default to FALSE. |
group |
Parameter same as "group" in the lavaan sem() function; whether to fit a multigroup model. |
cluster |
Parameter same as "cluster" in the lavaan sem() function; whether to fit a cluster model. |
constraints |
Parameter same as "constraints" in the lavaan sem() function; whether to apply constraints to the model. |
WLS.V |
Parameter same as "WLS.V" in the lavaan sem() function; whether to use WLS.V estimator. |
NACOV |
Parameter same as "NACOV" in the lavaan sem() function; whether to use NACOV estimator. |
netstats.options |
A user-specified named list with element names corresponding to the network statistics names and element values corresponding to options for that network statistics used as optional arguments in the corresponding functions in the "sna" or "influential" packages. e.g., netstats.options=list("degree"=list("cmode"="freeman"), "closeness"=list("cmode"="undirected"), "clusterRank"=list("directed"=FALSE)). |
... |
Optional arguments for the sem() function. |
Value
A networksem object containing the updated model specification string with the reconstructed network statistics as variables and a lavaan SEM object.
Examples
set.seed(100)
nsamp = 20
net <- ifelse(matrix(rnorm(nsamp^2), nsamp, nsamp) > 1, 1, 0)
mean(net) # density of simulated network
lv1 <- rnorm(nsamp)
lv2 <- rnorm(nsamp)
nonnet <- data.frame(x1 = lv1*0.5 + rnorm(nsamp),
x2 = lv1*0.8 + rnorm(nsamp),
x3 = lv2*0.5 + rnorm(nsamp),
x4 = lv2*0.8 + rnorm(nsamp))
model <-'
lv1 =~ x1 + x2
lv2 =~ x3 + x4
net ~ lv2
lv1 ~ net + lv2
'
data = list(network = list(net = net), nonnetwork = nonnet)
set.seed(100)
res <- sem.net(model = model, data = data, netstats = c('degree'))
summary(res)
Compute user-specified network statistics for a specific network.
Description
Compute user-specified network statistics for a specific network.
Usage
sem.net.addvar(
model.network.stat.var.list = NULL,
data = NULL,
netstats = NULL,
model.network.var.i = NULL,
netstats.rescale = TRUE,
netstats.options = NULL
)
Arguments
model.network.stat.var.list |
a list of elements with names corresponding to the network names and values corresponding to lists of network statistics that will be calculated for the corresponding network |
data |
a list containing both the non-network and network data |
netstats |
a list of user-specified network statistics |
model.network.var.i |
the index of a network within all networks |
netstats.rescale |
a logical value indicating whether to rescale network statistics to have mean 0 and sd 1 |
netstats.options |
a list with element names corresponding to the network statistics and element values corresponding to another list. The list corresponding to each network statistics has element names being the argument names for calculating the network statistics, and values being the argument values |
Value
a list with the first value being the list of network statistics names and the second value being the data frame with added network statistics variables
Compute a list of user-specified network statistics using the "influential" package and add it to the existing data.
Description
Compute a list of user-specified network statistics using the "influential" package and add it to the existing data.
Usage
sem.net.addvar.influential(
model.network.stat.var.list,
data,
model.network.var.i,
stats,
statsname,
netstats.rescale,
netstats.options = NULL
)
Arguments
model.network.stat.var.list |
a list of elements with names corresponding to the network names and values corresponding to lists of network statistics that will be calculated for the corresponding network |
data |
a list containing both the non-network and network data |
model.network.var.i |
an index indicating a specific network within all networks |
stats |
a network statistics that can be calculated using package "influential" |
statsname |
name of the network statistics |
netstats.rescale |
a logical value indicating whether to rescale network statistics to have mean 0 and sd 1 |
netstats.options |
a list with names being the argument names for calculating the network statistics, and values being the argument values |
Value
a list with the first value being the list of network statistics names and the second value being the data frame with added network statistics
Compute a list of user-specified network statistics values using the "sna" package and add them to the non-network data.
Description
Compute a list of user-specified network statistics values using the "sna" package and add them to the non-network data.
Usage
sem.net.addvar.stat(
model.network.stat.var.list,
data,
model.network.var.i,
stats,
statsname,
netstats.rescale,
netstats.options = NULL
)
Arguments
model.network.stat.var.list |
a list of elements with names corresponding to the network names and values corresponding to lists of network statistics that will be calculated for the corresponding network |
data |
a list containing both the non-network and network data |
model.network.var.i |
an index indicating a specific network within all networks |
stats |
a network statistics that can be calculated using package "sna" |
statsname |
name of the network statistics |
netstats.rescale |
a logical value indicating whether to rescale network statistics to have mean 0 and sd 1 |
netstats.options |
a list with names being the argument names for calculating the network statistics, and values being the argument values |
Value
a list with the first value being the list of network statistics names and the second value being the data frame with added network statistics
Fit a Structural Equation Model (SEM) with both network and non-network data by transforming nonnetwork data into paired values corresponding to network edge values.
Description
Fit a Structural Equation Model (SEM) with both network and non-network data by transforming nonnetwork data into paired values corresponding to network edge values.
Usage
sem.net.edge(
model = NULL,
data = NULL,
type = "difference",
ordered = NULL,
sampling.weights = NULL,
data.rescale = FALSE,
group = NULL,
cluster = NULL,
netstats.rescale = FALSE,
constraints = "",
WLS.V = NULL,
NACOV = NULL,
...
)
Arguments
model |
A model string specified in lavaan model syntax that includes relationships among the network and non-network variables. |
data |
A list containing the data. The list has two named components, "network" and "nonnetwork"; "network" is a list of named adjacency matrices for the network data, and "nonnetwork" is the dataframe of non-network covariates. |
type |
Option for transforming nonnework data; "difference" for using the difference between two individuals as the edge covariate; "average" for using the average between two individuals as the edge covariate. |
ordered |
Parameter same as "ordered" in the lavaan sem() function; whether to treat data as ordinal. |
sampling.weights |
Parameter same as "sampling.weights" in the lavaan sem() function; whether to apply weights to data. |
data.rescale |
TRUE or FALSE, whether to rescale the whole dataset (with restructured network and nonnetwork data) to have mean 0 and standard deviation 1 when fitting it to SEM, default to FALSE. |
group |
Parameter same as "group" in the lavaan sem() function; whether to fit a multigroup model. |
cluster |
Parameter same as "cluster" in the lavaan sem() function; whether to fit a cluster model. |
netstats.rescale |
TRUE or FALSE, whether to rescale the network statistics to have mean 0 and standard deviation 1, default to FALSE. |
constraints |
Parameter same as "constraints" in the lavaan sem() function; whether to apply constraints to the model. |
WLS.V |
Parameter same as "WLS.V" in the lavaan sem() function; whether to use WLS.V estimator. |
NACOV |
Parameter same as "NACOV" in the lavaan sem() function; whether to use NACOV estimator. |
... |
Optional arguments for the sem() function. |
Value
A networksem object containing the updated model specification string with the reconstructed network statistics as variables and a lavaan SEM object.
Examples
set.seed(100)
nsamp = 20
net <- data.frame(ifelse(matrix(rnorm(nsamp^2), nsamp, nsamp) > 1, 1, 0))
mean(net) # density of simulated network
lv1 <- rnorm(nsamp)
lv2 <- rnorm(nsamp)
nonnet <- data.frame(x1 = lv1*0.5 + rnorm(nsamp),
x2 = lv1*0.8 + rnorm(nsamp),
x3 = lv2*0.5 + rnorm(nsamp),
x4 = lv2*0.8 + rnorm(nsamp))
model <-'
lv1 =~ x1 + x2
lv2 =~ x3 + x4
lv1 ~ net
lv2 ~ lv1
'
data = list(network = list(net = net), nonnetwork = nonnet)
set.seed(100)
res <- sem.net.edge(model = model, data = data, type = 'difference')
summary(res)
Fit a Structural Equation Model (SEM) with both network and non-network data by transforming nonnetwork data into paired values corresponding to network latent distance pairs.
Description
Fit a Structural Equation Model (SEM) with both network and non-network data by transforming nonnetwork data into paired values corresponding to network latent distance pairs.
Usage
sem.net.edge.lsm(
model = NULL,
data = NULL,
type = "difference",
latent.dim = 2,
data.rescale = FALSE,
ordered = NULL,
sampling.weights = NULL,
group = NULL,
cluster = NULL,
netstats.rescale = FALSE,
constraints = "",
WLS.V = NULL,
NACOV = NULL,
...
)
Arguments
model |
A model string specified in lavaan model syntax that includes relationships among the network and non-network variables. |
data |
A list containing the data. The list has two named components, "network" and "nonnetwork"; "network" is a list of named adjacency matrices for the network data, and "nonnetwork" is the dataframe of non-network covariates. |
type |
"difference" for using the difference between the network statistics of the two actors as the edge covariate; "average" for using the average of the network statistics of the two actors as the edge covariate. |
latent.dim |
The number of network latent dimensions to use in extracting latent positions of network nodes. |
data.rescale |
TRUE or FALSE, whether to rescale the whole dataset (with restructured network and nonnetwork data) to have mean 0 and standard deviation 1 when fitting it to SEM, default to FALSE. |
ordered |
Parameter same as "ordered" in the lavaan sem() function; whether to treat data as ordinal. |
sampling.weights |
Parameter same as "sampling.weights" in the lavaan sem() function; whether to apply weights to data. |
group |
Parameter same as "group" in the lavaan sem() function; whether to fit a multigroup model. |
cluster |
Parameter same as "cluster" in the lavaan sem() function; whether to fit a cluster model. |
netstats.rescale |
TRUE or FALSE, whether to rescale the network statistics to have mean 0 and standard deviation 1, default to FALSE. |
constraints |
Parameter same as "constraints" in the lavaan sem() function; whether to apply constraints to the model. |
WLS.V |
Parameter same as "WLS.V" in the lavaan sem() function; whether to use WLS.V estimator. |
NACOV |
Parameter same as "NACOV" in the lavaan sem() function; whether to use NACOV estimator. |
... |
Optional arguments for the sem() function. |
Value
A networksem object containing the updated model specification string with the reconstructed network statistics as variables, a lavaan SEM output object, and a latentnet ergm object.
Examples
set.seed(10)
nsamp = 20
lv1 <- rnorm(nsamp)
net <- ifelse(matrix(rnorm(nsamp^2) , nsamp, nsamp) > 1, 1, 0)
lv2 <- rnorm(nsamp)
nonnet <- data.frame(x1 = lv1*0.5 + rnorm(nsamp),
x2 = lv1*0.8 + rnorm(nsamp),
x3 = lv2*0.5 + rnorm(nsamp),
x4 = lv2*0.8 + rnorm(nsamp))
model <-'
lv1 =~ x1 + x2
lv2 =~ x3 + x4
net ~ lv1
lv2 ~ net
'
data = list(network = list(net = net), nonnetwork = nonnet)
set.seed(100)
res <- sem.net.edge.lsm(model = model, data = data, latent.dim = 1)
summary(res)
Fit a Structural Equation Model (SEM) with both network and non-network data by incorporating network latent positions as variables.
Description
Fit a Structural Equation Model (SEM) with both network and non-network data by incorporating network latent positions as variables.
Usage
sem.net.lsm(
model = NULL,
data = NULL,
latent.dim = 2,
ordered = NULL,
sampling.weights = NULL,
data.rescale = FALSE,
netstats.rescale = FALSE,
group = NULL,
cluster = NULL,
constraints = "",
WLS.V = NULL,
NACOV = NULL,
...
)
Arguments
model |
A model string specified in lavaan model syntax that includes relationships among the network and non-network variables. |
data |
A list containing the data. The list has two named components, "network" and "nonnetwork"; "network" is a list of named adjacency matrices for the network data, and "nonnetwork" is the dataframe of non-network covariates. |
latent.dim |
The number of network latent dimensions to use in extracting latent positions of network nodes. |
ordered |
Parameter same as "ordered" in the lavaan sem() function; whether to treat data as ordinal. |
sampling.weights |
Parameter same as "sampling.weights" in the lavaan sem() function; whether to apply weights to data. |
data.rescale |
TRUE or FALSE, whether to rescale the whole dataset (with restructured network and nonnetwork data) to have mean 0 and standard deviation 1 when fitting it to SEM, default to FALSE. |
netstats.rescale |
TRUE or FALSE, whether to rescale the network statistics to have mean 0 and standard deviation 1, default to FALSE. |
group |
Parameter same as "group" in the lavaan sem() function; whether to fit a multigroup model. |
cluster |
Parameter same as "cluster" in the lavaan sem() function; whether to fit a cluster model. |
constraints |
Parameter same as "constraints" in the lavaan sem() function; whether to apply constraints to the model. |
WLS.V |
Parameter same as "WLS.V" in the lavaan sem() function; whether to use WLS.V estimator. |
NACOV |
Parameter same as "NACOV" in the lavaan sem() function; whether to use NACOV estimator. |
... |
Optional arguments for the sem() function. |
Value
A networksem object containing the updated model specification string with the reconstructed network statistics as variables, a lavaan SEM output object, and a latentnet ergmm object.
Examples
set.seed(10)
nsamp = 20
net <- ifelse(matrix(rnorm(nsamp^2), nsamp, nsamp) > 1, 1, 0)
mean(net) # density of simulated network
lv1 <- rnorm(nsamp)
lv2 <- rnorm(nsamp)
nonnet <- data.frame(x1 = lv1*0.5 + rnorm(nsamp),
x2 = lv1*0.8 + rnorm(nsamp),
x3 = lv2*0.5 + rnorm(nsamp),
x4 = lv2*0.8 + rnorm(nsamp))
model <-'
lv1 =~ x1 + x2
lv2 =~ x3 + x4
net ~ lv2
lv1 ~ net + lv2
'
data = list(network = list(net = net), nonnetwork = nonnet)
set.seed(100)
res <- sem.net.lsm(model = model, data = data, latent.dim = 2)
summary(res)
Summarize output from networksem functions includeing sem.net, sem.net.lsm, sem.net.edge, sem.net.edge.lsm.
Description
Summarize output from networksem functions includeing sem.net, sem.net.lsm, sem.net.edge, sem.net.edge.lsm.
Usage
## S3 method for class 'networksem'
summary(object, ...)
Arguments
object |
A networksem output object. |
... |
Optional arguments. |
Value
a summary sheet of the networksem output.