CB-SEM Studio: free online covariance-based structural equation modeling software
Data
One row per respondent, one column per survey item. CSV, TSV and Excel files are read on your device.
Need to recode, reverse-code or clean your data first? Use the free data cleaning and coding tool, then send the clean data straight back here.
No data yet
Try the app with a sample survey of 300 respondents measuring technology acceptance.
Model
Group your items into constructs, drag them where you want, then draw arrows or pick paths from the lists. Add moderators and covariances from the boxes below the canvas. With no arrows, the model is a confirmatory factor analysis.
Paths (hypotheses)
0Moderation
0Choose a moderator and the path whose strength it changes. A latent interaction is built from product indicators automatically.
Covariances
0Exogenous constructs are correlated automatically. Add correlated residuals between items, or between the disturbances of outcome constructs, only with a theoretical reason.
Mediation
0Indirect routes found in your paths. Each is tested automatically.
lavaan syntax
# measurement model: construct =~ items TRUST =~ TR1 + TR2 + TR3 SAT =~ SAT1 + SAT2 + SAT3 LOY =~ LOY1 + LOY2 + LOY3 # structural model: outcome ~ predictors SAT ~ TRUST LOY ~ SAT + TRUST # covariances (e.g. correlated residuals) SAT1 ~~ SAT2
Edits here update the visual builder, and the reverse. Lines starting with # are comments.
Run
Maximum likelihood estimation, as in AMOS and lavaan. The defaults suit most studies.
Starting
Advanced settings
Model fit
How closely the model reproduces the observed covariances, with common cut-offs (Hu & Bentler, 1999).
Measurement model
Loadings, reliability, convergent and discriminant validity.
Structural model
Path coefficients, hypothesis tests, explained variance and covariances.
Mediation
Every indirect route with delta-method and Monte Carlo confidence intervals, classified after Zhao, Lynch & Chen (2010).
Moderation
Latent interactions estimated with double-mean-centred product indicators, simple slopes at ±1 SD of the moderator and moderated mediation.
Modification indices
How much the chi-square would drop if a fixed parameter were freed. Free only what theory supports, one change at a time.
Diagram
The figure keeps the layout from the builder. Drag constructs or single items, click a construct to change its item side, and zoom with the buttons in the corner.
Solid arrows are significant at the chosen level; dashed arrows are not. Curved two-headed arrows are covariances (correlations when standardized). * p < .05 · ** p < .01 · *** p < .001.
Export & cite
All tables, the diagram and the model syntax, in the format your manuscript or next tool needs.
Please cite CB-SEM Studio
This tool is free. If you use it in a thesis, article, conference paper or report, please cite the software and the website. Created by Bibin Xavier.
Methods paragraph
A starting point for your write-up. Check it against your study and add the citations your field expects.