Questions tagged [structural-equation-modeling]
Structural Equation Modeling is a multivariate technique. It is based on formulating a set of linear relations between variables, some of which may be latent, and estimating the whole system, typically by analyzing the covariance matrix of the observed variables.
1,334 questions
2
votes
0
answers
19
views
Differences between equivalent second-order LCA models
I have run a second-order LCA model in Mplus.
In the %OVERALLA% , I merely changed "c1 c2 on c" to "c2 c1 on c", which caused changes in the "C Classes (Rows) by C1 Classes (...
- 21
2
votes
2
answers
55
views
Dependent and independent variables in logistic regression
Sometimes in observational studies in the health sciences we have a lot of variables, all of them considered as random variables. Is it possible that variables that are considered as "independent&...
- 31
3
votes
1
answer
46
views
Calculating effect size with effects coding
I am using a scalar CFA to compare latent variable means for different groups to the overall latent mean in a kind of effects coding approach (I constrain the overall latent mean of the whole sample ...
- 43
0
votes
0
answers
22
views
Proportion of Maximum Scaling (POMS) for a reverse coded item
I'm working with an SEM model with 7 indicators, 5 have a range from [0, X], where X is a better score. The other 2 are time based, so they range from [0, Y] where Y is a worse score. I'm planning to ...
- 1
1
vote
0
answers
27
views
How to plot path diagram for SEM with generalized additive model
I am working with piecewiseSEM package in R. I have included gam models in my psem. My question is how we can represent the relationships from psem with gam models in a path diagram. Since we are not ...
- 11
6
votes
1
answer
114
views
What is the meaning of $(I - A^{\intercal})^{-1}$ in the linear Structural Equation Model (SEM) formulation of the DAG-GNN paper?
In the DAG-GNN paper (Yu et al., NeurIPS 2019, paper link), the authors describe the linear Structural Equation Model (SEM) as follows:
$$X = A^{\intercal}X + Z,$$
where $A$ (an $m \times m$ matrix) ...
- 61
2
votes
1
answer
66
views
About measuring actual behaviour in PLS-SEM with single item
I am trying to find items for measuring actual behaviour (latent variable) in PLS-SEM. However, when I go to items for measuring actual behaviour, there would be a conflict between the number of items ...
- 137
3
votes
1
answer
89
views
measurement invariance with theta and lisrel parameterization
I’m trying to test a model of threshold invariance for ordinal indicators across 6 time points using semTools::measEq.syntax() with
parameterization = "theta",
ID.cat = "...
- 369
2
votes
2
answers
104
views
Can I use unstandardised factor loadings to compute communality?
For a larger problem I am trying to solve, I have been reading about communality/item-level reliability for a single-factor model.
From what I have read, I see that the formula for communality ($c$) ...
- 294
2
votes
1
answer
67
views
standardized indirect effects on SEM [closed]
I have 4 exogenous variables, 3 mediators, and 3 endogenous variables. AMOS has calculated direct and indirect effects, and I have relied on standardized indirect effects to report in my dissertation. ...
- 21
6
votes
1
answer
205
views
What are the pros and cons of a latent multivariate model vs a non-latent multivariate model
From what I understand, the advantage of a multivariate regression model - by which I mean two or more outcomes variables regressed on the same predictor - over several simple regressions is that the ...
- 2,345
2
votes
1
answer
79
views
Piecewise SEM with mixed resolutions: year-level environment, plot-level biota — how to avoid pseudo-replication and handle autocorrelation?
I’m fitting a piecewise SEM to understand drivers of a grass species' productivity. My data span ~20 years at a single location. Two variables are plot-level (3 plots) that were repeatedly sampled:
• ...
- 209
1
vote
0
answers
32
views
"Robustness test for results" in a multi-group SEM analysis
I have a study that used multi-group SEM analysis. The multi-groups intended to compare across different countries. Each group has a sample of about 500. There's 6 latent variables and 1 observable ...
- 71
3
votes
1
answer
153
views
Same-sign multiplication in model equation
I want to set up a regression with this model equation:
$y = \beta_0 + \beta_1.x_1 + \beta_2.x_2 + \beta_3.f(x_1,x_2)$
where $f$ is a "same-sign multiplication" operation
$f(x_1,x_2) = \...
- 238
1
vote
2
answers
148
views
Latent change score model with partial scalar invariance
I am investigating change in a latent factor measured by 3 ordinal items (each having three response categories) at three successive timepoints. I have determined that I have partial scalar invariance ...
- 7,022