Path canvas
Visual model building
R² .41
Trust on the node
Q² .31
Predictive relevance
Draw latents, indicators, and paths on one canvas. Mode, R², and Q² print on the node after Estimate—same diagram you export.
SEMGate
The next generation of SEM, is here.
Build, estimate, and publish PLS, CB, Meta-SEM, Multilevel, Exploratory SEM, and Monte Carlo — Hair-ready diagnostics in one studio.
Occasional SEMGate news - unsubscribe anytime.
Dual engines
PLS-SEM & CB-SEM
.44
PLS path β
.94
CB-SEM CFI
IndexEstimateCut-off
CFI0.942≥ .90
TLI0.931≥ .90
RMSEA0.048≤ .08
SRMR0.041≤ .08
One visual model, two engines: variance-based PLS composites and covariance-based CB-SEM with χ², CFI, TLI, RMSEA, and SRMR.
Inference
Bootstrap & Hair reporting
5,000
Bootstrap samples
8.23
t Trust → Advocacy
Pathβtp
Exp. → Trust0.4369.10<.001
Trust → Adv.0.4418.23<.001
Value → Adv.0.2874.51<.001
Exp. → Adv.0.0941.420.156
Smart bootstrap fills SE, t, p, and CIs on the paths you drew. Hair-style Stage 1–2 tables stay linked to the same model.
Interactions
Moderation & Johnson–Neyman
.12–.68
JN significance region
±1 SD
Simple slopes
Latent interactions with simple slopes and Johnson–Neyman bands—plus PROCESS-style conditional indirect effects and IMM when you need them.
Nonlinear
Quadratic effects
.18
Quadratic β
.09
f² (small–medium)
Pathf²Size
Exp. → Trust0.210Medium
Trust → Adv.0.198Medium
Value → Adv.0.094Small
Exp. → Adv.0.012-
Estimate squared terms alongside linear paths. Effect sizes and plots land in Results next to the structural table.
Prediction suite
IPMA · PLSpredict · NCA
.27
Q²predict Trust
IPMA
Importance–performance
ConstructQ²RMSEMAE
Trust0.3100.7240.581
Advocacy0.3800.6920.548
Q²predict Trust 0.274 · Advocacy 0.331
Out-of-sample PLSpredict, importance–performance maps, and necessary-condition analysis run from the same PLS model.
Specialized
RSA & Ordinal PLS
a1–a4
Response surface
Ordinal
Polychoric PLS path
HTMT
.41
AVE
.71
CFI
.94
SRMR
.041
Response surface analysis for congruence hypotheses and ordinal PLS for Likert-style indicators—without leaving the studio.
Advanced SEM
Multilevel · HOC · CMV · Panel
ICC .31
Cluster nesting
HOC
Higher-order constructs
HTMT
.41
AVE
.71
CFI
.94
SRMR
.041
MSEM and multilevel PLS, higher-order constructs, CMV/CMB checks, Gaussian copula endogeneity, and panel/DSEM beyond single-level SEM.
Hair Stage 1
Measurement quality
.71
AVE Trust (≥ .50)
.41
HTMT Trust–Advocacy
ConstructαCRAVEStatus
Trust.88.91.71Established
Advocacy.86.90.69Established
Experience.84.89.67Established
AVE ≥ .50 · CR ≥ .70 · λ ≥ .70
α, ρA, CR, AVE, HTMT, Fornell–Larcker, and cross-loadings—the Stage 1 checklist before you interpret structural paths.
Publication
Results · figures · export
ggplot
Publication figures
docx
Word + Excel export
HTMT
.41
AVE
.71
CFI
.94
SRMR
.041
Rich Results tree with ggplot-ready figures and one-click Word and Excel export for manuscripts, supplements, and reviewer packs.
Path canvas
Visual model building
R² .41
Trust on the node
Q² .31
Predictive relevance
Draw latents, indicators, and paths on one canvas. Mode, R², and Q² print on the node after Estimate—same diagram you export.
Dual engines
PLS-SEM & CB-SEM
.44
PLS path β
.94
CB-SEM CFI
IndexEstimateCut-off
CFI0.942≥ .90
TLI0.931≥ .90
RMSEA0.048≤ .08
SRMR0.041≤ .08
One visual model, two engines: variance-based PLS composites and covariance-based CB-SEM with χ², CFI, TLI, RMSEA, and SRMR.
Visual model building
Auto model
Measurement that cleans itself — scientifically
Tap the wand. Auto model fixes failing Stage 1, then frees CB residuals only if fit still fails.
Auto model
Results → canvas → Results
Before · Stage 1 failsEnhance · on canvasAfter · Stage 1 clean
1
Before
Stage 1 fails
2
Enhance
On canvas
3
After
Stage 1 clean
Results
Stage 1 measurement
Weak loading
TR3 · λ .38
Trust α
.61
AVE
.42
HTMT
.94
CFI
.89
Same Results chrome as Studio — Hair Stage 1 flags fail in red before Auto model runs.
- 1BlinkWeak / reverse λ
- 2DropSafe remove
- 3Re-estimateRound complete
- 4CB residualIf fit still fails
Results
Stage 1 measurement
Clean loadings
TR3 dropped · Trust re-estimated
Trust α
.86
AVE
.68
HTMT
.71
CFI
.97
After tables pass in green — publication-ready Stage 1, Decline restores the pre-enhance model.
Analyses
Every major SEM workflow in one studio
From PLS and CB-SEM to Meta-SEM, multilevel, Monte Carlo, PROCESS, IPMA, NCA, MGA, method bias, and segmentation — run and report without switching tools.
PLS-SEM
- Estimate path coefficients on composites
- Outer loadings and measurement quality
- R², f², and predictive metrics
- Hair-style results tables in one canvas
CB-SEM
- Covariance-based factor models (lavaan)
- Global fit: χ², CFI, RMSEA, SRMR
- Modification indices and residuals
- Publication-ready measurement tables
Meta-SEM
- Stage 1: pool study-level correlations
- Stage 2: fit structural meta-models
- Keep path structure across samples
- Study library and meta variables
Multilevel SEM
- Within- and between-level CB-SEM
- ICCs and cross-level paths
- Cluster-aware estimation
- THREELEVEL / Mplus export options
Multilevel PLS
- Mode-A within/between composites
- ICC(1) for nested indicators
- Optional random slopes
- L2 / L3 nested structural paths
Exploratory SEM
- ESEM-within-CFA on CB-SEM (lavaan)
- Target-anchor identification
- Python target rotation on PLS
- Explore structure with paths retained
Ordinal PLS
- Treat Likert items as ordinal
- ALS category quantification
- Lohmoller optimal scaling
- Then run metric PLS on scaled data
Lohmoller PLS
- Factorial or centroid inner scheme
- Constrained exogenous Mode A
- Optional covariance-matrix-only input
- Classic extended PLS options
Monte Carlo
- Simulate from your model IR
- Power, bias, and RMSE curves
- Suggested N before data collection
- PLS, OLS, or CB-proxy recovery
Bootstrap & significance
- Resample for SEs and t-values
- p-values on structural paths
- Percentile confidence intervals
- SmartPLS-style inference workflow
PLSpredict
- Out-of-sample predictive assessment
- Q²predict for endogenous constructs
- Holdout / k-fold prediction metrics
- Compare predictive power of paths
IPMA
- Importance–performance maps
- Construct and indicator views
- Total effects vs performance (0–100)
- Prioritize high-importance gaps
NCA
- Necessary condition analysis
- Bottlenecks for an outcome
- Ceiling lines beside PLS paths
- Complement sufficiency with necessity
MGA
- Compare paths across groups
- Bootstrap or permutation (PLS)
- Wald contrasts with Holm (CB-SEM)
- Test measurement differences by segment
Measurement invariance
- Configural → metric → scalar → strict
- Group equivalence for CB-SEM
- Partial invariance options
- Gate fair multigroup comparisons
Path equality (Wald)
- Test whether two paths are equal
- lavaan lavTestWald / nested Δχ²
- Auto pairwise constraints by outcome
- CB-SEM structural equality checks
Gaussian copula
- Park & Gupta endogeneity check
- Control-function QC terms on paths
- Flag significant copula regressors
- Automatic subset screening option
Nonlinear / quadratic
- Quadratic (X²) and cubic (X³) terms
- Two-stage PLS (Hair / Basco)
- QE nodes on the canvas
- Self-moderation and curve shapes
CMV / CMB
- Harman single-factor screen
- Full-collinearity VIFs
- ULMC, MLMV, Lindell–Whitney
- CFA marker / method-factor checks
Segmentation
- FIMIX-PLS latent classes
- POS, GAS, and REBUS options
- Recover unobserved heterogeneity
- Segment when one model is too coarse
fsQCA
- Fuzzy-set QCA on PLS scores
- Calibration and necessity tests
- Truth table and Quine–McCluskey
- Complex / parsimonious / intermediate
RSA
- Edwards polynomial response surface
- a1–a4 surface parameters
- Shanock-style congruence cells
- Principal axes and optional bootstrap
Diagnostics & export
- Sample-size rules of thumb
- Hair checklist and model-fit flags
- Export Word and Excel tables
- Publication-ready diagrams
Connected intelligence
The power of dual SEM engines
Same visual model. SEMGate engines for variance-based (PLS-SEM) and covariance-based (CB-SEM) estimation.
PLS-SEM / PLSc
Prediction-oriented models with formative and reflective constructs.
- ✓Reflective & formative constructs
- ✓Bootstrapping & confidence intervals
- ✓PLSpredict & Q²predict
- ✓IPMA & higher-order constructs
- ✓MGA & measurement invariance
- ✓Total, direct & indirect effects
Anywhere
Use SEMGate from any device
Build and review models on desktop, tablet, or phone — same studio, adapted to your screen.
Invitation to verify
Help verify SEMGate against PLS-SEM and AMOS
We invite experienced SEM researchers to check how closely SEMGate matches popular packages—run the same model on your data in a PLS-SEM package or AMOS and in SEMGate, then rate accuracy on a clear Measurement and Structural checklist.
Who
Researchers published with PLS-SEM or AMOS
What
Side-by-side accuracy check on your own data
Thank you
1-year SEMGate activation if accepted
Pricing
Plans for researchers and institutions
The most affordable and user-friendly SEM platform — start with a free trial, or seat your lab on an institutional plan.
Most popular30 days free
Individual
Full research workspace
$70USD / 6 months
$11.67 a month. Save $20.
Start free trialNo charge for 30 days. No card required. Cancel anytime.
Powered by Stripe
Card details are entered on Stripe and never stored by SEMGate.
Enterprise
Institutions & labs
Shared seats, verified domains, and admin controls for departments, faculties, and research consortia.
$7,200/ year
Up to 100 seats · $6.00/user/mo · USD annual billing
Unlimited concurrent sessions
Departments & mid-size faculties
Opens institution signup — then choose your plan tier and continue.
Security
Trust is not a feature. It is our track record.
AES-256-GCM with a unique key per dataset. Decrypt briefly in memory during estimation, then discard.
Learn more →Testimonials
What researchers say
Approved notes from SEM researchers and instructors using SEMGate in teaching and research.
Be the first to share how SEMGate fits your SEM workflow — submit on the right.
Share a testimonial
Submissions are reviewed before they appear on this page.