The dySEM helps automate the process of scripting, fitting, and
reporting on latent models of dyadic data via
lavaan. The package was developed and used
in the course of the research described in Sakaluk, Fisher, &
Kilshaw (2021).
The dySEM logo was designed by Lowell Deranleau (for logo design
inquiries, email: [email protected]).
You can install the released version of dySEM from CRAN with:
install.packages("dySEM")You can install the development version from GitHub with:
# install.packages("devtools")
devtools::install_github("jsakaluk/dySEM")The package currently provides functionality regarding the following types of latent dyadic data models:
- Dyadic Confirmatory Factor Analysis
- Latent Actor-Partner Interdependence Models (APIM)
- Latent Common Fate Models (CFM)
- Latent Bifactor Dyadic (Bi-Dy) Models
- Observed Actor-Partner Interdependence (APIM)
- Dyadic Exploratory Factor Analysis (NEW)
Additional features currently include:
- Automated specification of invariance constraints for any model, including full indistinguishability
- Functions to assist with the specification of I-SAT Models and I-NULL Models for calibrated model fit indexes with indistinguishable dyad models
- Functions to assist with reproducible creation of path diagrams and tables of statistical output
- Functions to calculate supplemental statistical information (e.g., omega reliability, noninvariance effect sizes, corrected model fit indexes)
Functionality targeted for future development of dySEM is tracked
here. Current
high-priority items include:
- Longitudinal dyadic model scripting functions (e.g., curve of factors, common fate growth)
- Latent dyadic response surface analysis scripting and visualization functions
- Multi-group dyadic model scripting (e.g., comparing models from samples of heterosexual vs. LGBTQ+ dyads)
- Covariate scripting and optionality
- Improved ease of item selection in scraper functions
Please submit any feature requests via the dySEM
issues page, using the
“Wishlist for dySEM Package Development” tag.
If you are interested in collaborating on the development of dySEM,
please contact Dr. Sakaluk.
A dySEM workflow typically involves five steps, which are covered
in-depth in the Overview
vignette.
Briefly, these steps include:
- Import and wrangle data
- Scrape variables from your data frame
- Script your preferred model
- Fit and Inspect your model via
lavaan - Output statistical visualizations and/or tables
There are additional optional functions, as well, that help users to calculate certain additional quantitative values (e.g., reliability, corrected model fit indexes in models with indistinguishable dyad members).
Structural equation modeling (SEM) programs like lavaan require dyadic
data to be in dyad structure data set, whereby each row contains the
data for one dyad, with separate columns for each observation made for
each member of the dyad. For example:
DRES
#> # A tibble: 121 × 18
#> PRQC_1.1 PRQC_2.1 PRQC_3.1 PRQC_4.1 PRQC_5.1 PRQC_6.1 PRQC_7.1 PRQC_8.1
#> <int> <int> <int> <int> <int> <int> <int> <int>
#> 1 7 7 7 7 7 7 7 5
#> 2 6 6 6 7 7 6 5 5
#> 3 7 7 7 7 7 7 7 6
#> 4 6 6 6 7 7 6 5 6
#> 5 7 7 7 7 7 6 7 6
#> 6 6 6 6 6 6 3 6 5
#> 7 7 6 7 6 6 6 5 6
#> 8 6 7 7 7 7 6 5 6
#> 9 7 7 7 7 7 6 6 6
#> 10 6 6 6 7 7 7 4 4
#> # ℹ 111 more rows
#> # ℹ 10 more variables: PRQC_9.1 <int>, PRQC_1.2 <int>, PRQC_2.2 <int>,
#> # PRQC_3.2 <int>, PRQC_4.2 <int>, PRQC_5.2 <int>, PRQC_6.2 <int>,
#> # PRQC_7.2 <int>, PRQC_8.2 <int>, PRQC_9.2 <int>The dySEM scrapers consider appropriately repetitiously named
indicators as consisting of at least three distinct elements: stem,
item, and partner. Delimiter characters (e.g., “.”, “_“) are
commonly–but not always–used to separate some/all of these
elements.dySEM scrapers largely function by asking you to specify in
what order the elements of variable names are ordered.
dvn <- scrapeVarCross(DRES, x_order = "sip", x_stem = "PRQC", x_delim1="_",x_delim2=".", distinguish_1="1", distinguish_2="2")Scripter functions like
scriptCFA
typically require only three arguments to be specified:
- the
dvnobject (e.g., fromscrapeVarCross) to be used to script the model 1.arbitrary name(s) for the latent variable(s) you are modeling - the kind of parameter equality constraints that you wish to be imposed (if any)
qual.indist.script <- scriptCFA(dvn, lvname = "Quality")This function returns a character object with lavaan compliant syntax
for your chosen model, as well as exporting a reproducible .txt of the
scripted model to a /scripts folder in your working directory.
You can immediately pass any script(s) returned from a dySEM scripter
to your preferred lavaan wrapper, with your estimator and missing data
treatment of choice. For example:
qual.indist.fit <- lavaan::cfa(qual.indist.script, data = DRES, std.lv = FALSE, auto.fix.first= FALSE, meanstructure = TRUE)At this point, the full arsenal of lavaan model-inspecting tools are
at your disposal. For example:
summary(qual.indist.fit, fit.measures = TRUE, standardized = TRUE, rsquare = TRUE)dySEM also contains functionality to help you quickly, correctly, and
reproducibly generate output from your fitted model(s), in the forms of
path diagrams and/or tables of statistical values. By default these save
to a temporary directory, but you can specify a directory of your choice
by replacing tempdir() (e.g., with ".", which will place it in your
current working directory).
outputModel(dvn, model = "cfa", fit = qual.indist.fit,
table = TRUE, tabletype = "measurement",
figure = TRUE, figtype = "unstandardized",
writeTo = tempdir(),
fileName = "dCFA_indist")Please note that the dySEM project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.
The development of dySEM has been generously supported by Internal
Grants from Western University, including:
- a Research Mobilization, Creation & Innovation Grants for SSHRC-Related Research
- a Western Knowledge Mobilization Innovation Grants