STRUCTURAL EQUATION

MODELING (SEM)
Structural Equation Modelling
(sumber:Niina Kotamäki)
 Covariance structure analysis
 Causal modeling
 Simultaneous equations modeling
 Path analysis
 Confirmatory factor analysis
 Latent variable modeling
 LISREL-modeling

 Highly flexible “modeling toolbox”

 Extension of the general linear model (GLM)

 SEM
 Quite recent innovation (late 1960s early 1970 )
 Extensively applied in social sciences, psychology, economy, chemistry
and biology
◦ Applications in ecology and environmental sciences are limited
◦ Even less common in aquatic ecosystems
 tests theoretical hypothesis about causal relationships
 tests relationships between observed and unobserved variables
 combines regression analysis (path analysis) and factor analysis
 researchers use SEM to determine whether a certain model is valid
 X1 a
Regression model:
Y=aX1+bX2+ε corr Y ε
b
X2
DEPENDENT INDEPENDEN
T M

LIMITATIONS
Multiple dependent (Y) variables are not permitted
Each independent variable (X) is assumed to be measured without error
 controlled experiments  measurement errors are negligible and
uncontrolled variation is at minimum
 observational studies  all variables are subject to measurement error and
uncontrolled variation

Strong correlation (multicollinearity) may cause biased parameter estimates

and inflated standard errors
Indirect effects (mediating variables) cannot be included
The error or residual variable is the only unobserved variable
 SEM deals with these limitations

 Works with multiple, related equations simultaneously

 Allows reciprocal relationships
 Ability to model constructs as latent variables
 Allows the modeller to explicitly capture unreliability of
measurement in the model
 Indirect effects / mediating variables
 Compares the performance of a model across multiple
populations
 Steps of SEM analysis
1. Development of hypothesis / theory
2. Construction of path diagram
3. Model specification
4. Model identification
5. Parameter estimation
6. Model evaluation
7. Model modification
1. Development of hypothesis
SEM is a confirmatory technique:
 researcher needs to have established theory about

the relationships
 suited for theory testing, rather than theory

development
 2. Construction of path diagram
error

coefficients
path η
error

ξ
correlati
path on
η

Endogenous latent error

variable
Exogenous latent
variable
 3. Model Specification
 Creating a hypothesized model that you think
explains the relationships among multiple
variables

 Converting the model to multiple equations

 4. Model Identification
 (Just) identified
◦ a unique estimate for each parameter
◦ number of equations = number of parameters to be estimated
◦ a+b=5, a-b=2

 Under-identified (not identified)

◦ number of equations < number of parameters
◦ infinite number of solutions
◦ a+b=7
◦ model can not be estimated

 Over-identified
◦ number of equations > number of parameters
◦ the model can be wrong
Just identified model

ξ1 η2

ξ2

η1
ξ3
Over-identified model (SEM usually)

ξ1 η1

ξ2

η2
ξ3
5. Parameter estimation
 technique used to calculate parameters

 testing how well a model fits the data

 expected covariance structure is tested against the

covariance matrix of oberved data H0: Σ=Σ(θ)

 estimating methods: e.g. maximum likelihood (ML),

ordinary least Squares (OLS), etc.
 Measurement Model
◦ The part of the model that relates indicators to latent factors
◦ The measurement model is the factor analytic part of SEM
◦ The respective regression coefficient is called lambda () /
loading

 Structural model
◦ This is the part of the model that includes the relationships
between the latent variables
◦ relation between endogenous and exogenous construct is
called gamma (γ) and relation between two endogenous
constructs is called beta (β)
 Measurement
δ1 X1 λx11
model
Structural model
λx21 ξ1
δ2 X2 γ11
λy11
ϕ21 y1 ε1

δ1 X3 λx32 γ12
η1 λy21
y2 ε2
ϕ31
λx42 ξ2 β21
δ2 X4 γ22
λy32
y3 ε3
ϕ32
η 2 λy42
δ1 X5 λx53 γ23 y4 ε4

λx63 ξ3
δ2 X6
Endogenous latent
variables
Exogenous latent
variables
 6. Model evaluation
 Total model
• Chi Square (2) test
• the theoretically expected values vs. the empirical
data

• Because we are dealing with a measure of misfit, the

p-value for 2 should be larger than .05 to decide that
the theoretical model fits the data

• fit indices e.g. RMSEA, CFI, NNFI etc.

 Model parts
• t-value for the estimated parameters showing whether
they are different from 0 (or any other value that we
want to fix!);
t > 1.96, p < .05
 7. Model modification
 Simplify
the model (i.e., delete non-significant
parameters or parameters with large standard
error)

 Expand the model (i.e., include new paths)

 Confirmatory vs. explanatory

◦ Don’t go too far with model modification!
 Advantages of SEM
 use of confirmatory factor analysis to reduce measurement
error by having multiple indicators per latent variable
 graphical modeling interface
 testing models overall rather than coefficients individually
 testing models with multiple dependents
 modeling indirect variables
 testing coefficients across multiple between-subjects

groups
 handling difficult data (time series with autocorrelated error,

non-normal data, incomplete data).

SEM in ecology, example
Structural model

Physical environment Water clarity

Phytoplankton
dynamics

Nutrients Herbivore

Example from: G.B. Arhonditsis, C.A. Atow, L.J.

Steinberg, M.A. Kenney, R.C. Lathrop, S.j. McBride, K.H.
Reckhow. Exploring ecological patterns with structural
equation modeling and Bayesian analysis.
Ecological Modeling 192 (2006) 385-409
 Biovolume Chlorophyll a

water
Epilimnion depth clarity

Phytoplankton
dynamics

Nutrients Herbivore

Phosphorus (SRP) Nitrogen (DIN) Daphnia Zooplankton

 ε1 ε2

Biovolume Chlorophyll a
ψ22
λ4 λ5
Epilimnion depth water
(physical environment)
γ1 clarity
β1

φ12 Phytoplankton
γ2 dynamics
β2
ψ33
Nutrients Herbivore
ψ11

λ2 λ6 λ7
λ3

Phosphorus (SRP) Nitrogen (DIN) Daphnia Zooplankton

δ2 δ3 ε5 ε4
 2 =22.473; df=19 0.67 0.79

p=0.261 >0.05 OK!

Biovolume Chlorophyll a 0.84

0.82 0.89
Epilimnion depth water
(physical environment) -0.07 -0.92 clarity

0.42 Phytoplankton
dynamics
-0.84 -0.66

0.43
Nutrients Herbivore
0.76

0.96 0.91
0.84 0.99

Phosphorus (SRP) Nitrogen (DIN) Daphnia Zooplankton

0.71 0.98 0.93 0.83

SEM Software packages
 LISREL
 AMOS
 Function sem in R
 MPlus
 EQS
 Mx
 SEPATH
References:
http://www.upa.pdx.edu/IOA/newsom/semrefs.htm