Causal Modelling

US 603
03.03.2023
Arnab Jana
 Causal models are mathematical models
representing causal relationships within an
individual system or population.

They facilitate inferences about causal

Causal relationships from statistical data.
Models
They can teach us a good deal about the
epistemology of causation, and about the
relationship between causation and
probability.
 SEM: very general, very powerful
multivariate technique.
What Is Specialized versions of other analysis
Structural methods.
Equation
Modeling?
Major applications of SEM:

Causal modeling or Confirmatory factor

path analysis. analysis (CFA).
What is path analysis?
• Path analysis, a precursor to and subset of structural equation modeling, is
a method to discern and assess the effects of a set of variables acting on a
specified outcome via multiple causal pathways.
• Developed nearly a century ago by Sewall Wright, a geneticist working at
the US Department of Agriculture,
• its early applications involved quantifying the contribution of genes vs.
environment on traits such as guinea pig coloration and assessing whether
temperature, humidity, radiation, or wind velocity had the greatest effect
on transpiration in plants.
• Social and life course epidemiologists subsequently adopted the method as
an effective way to distinguish direct from indirect effects and to test the
strength of hypothesized patterns of causal relationships.
What does it mean?
• Path analysis is based on a closed system of nested relationships
among variables that are represented statistically by a series of
structured linear regression equations.
• As such, path analysis is bound by the same set of assumptions as
linear regression, as well as some additional restrictions that describe
the allowable pattern of relations among variables.
• Variables are either exogenous, meaning their variance is not
dependent on any other variable in the model, or endogenous,
meaning their variance is determined by other variables in the model.
Exogenous variables may or may not be correlated with other
exogenous variables.
SEM

In figure, taken from

Pedhazur’s Multiple Regression in
Behavioral Research, variables 1
and 2 are exogenous and
correlated, while variables 3, and 4
are endogenous.
Interpretation
The structural equation that would
describe the relationship between
variables 1 and 3 is:
r13 = p31 + p32*r12

where r is the correlation coefficient

from a standard correlation matrix
containing all of the variables in the
model and the path coefficient p is the
standardized beta coefficient from the
linear regression model in which 1 and 2
are the independent variables and 3 is
the dependent variable.
 Test coefficients across multiple between-subjects groups

Ability to handle difficult data

Usefulness • Longitudinal with auto-correlated error

• Multi-level data
• Non-normal data
• Incomplete data
Terms, Nomenclature, Symbols,
and Vocabulary
• Experimental research
• independent and dependent variables.
• Non-experimental research
• predictor and criterion variables
• Observed (or manifest)

• Latent (or factors)

 • Rules of Thumb
Ratio of Sample Size to the Number of Free Parameters
Tanaka (1987): 20 to 1 (Most analysts now think
that is unrealistically high.)
Goal: Bentler & Chou (1987): 5 to 1
Several published studies do not meet this goal.
Sample Size
200 is seen as a goal for SEM research
Lower sample sizes can be used for
Sample Size Models with no latent variables
Models where all loadings are fixed (usually
to one)
Models with strong correlations
Simpler models

• Models for which there is an upper limit on N (e.g., countries or

years as the unit), 200 might be an unrealistic standard.
 Measurement model

• That part of a SEM model dealing

with latent variables and indicators.

Structural model
SEM details
• Contrasted with the above
• Set of exogenous and endogenous
variables in the model with arrows
and disturbance terms
Measurement Model: Confirmatory Factor
Analysis Observed or manifest variables

Hostility e1
D1

Hopelessness e2

Psychosocial
health e3
GHQ

Self-rated health e4

Latent construct or factor

Singh-Manoux, Clark and Marmot. 2002. Multiple measures of socio-economic

position and psychosocial health: proximal and distal measures.
Structural Model with Additional Variables
Observed or manifest variables

Hostility e1
D1
Education

Hopelessness e2

Occupation Psychosocial
health e3
GHQ

Income Self-rated health e4

Latent construct or factor

Causal Modeling or Path Analysis and
Confirmatory Factor Analysis
Education
Hostility e1
a= direct effect
b+c=indirect
Hopelessness e2
c Psychosocial
Income
health
GHQ e3
D1
D3
Self-rated health e4
Occupation

D2
What’s a Good Model?

• Fit measures:
–Chi-square test
–CFI (Comparative Fit Index)
–RMSE (Root Mean Square Error)
–TLI (Tucker Lewis Index)
–GFI (Goodness of Fit Index)
–And many, many, many more
–IFI, NFI, AIC, CIAC, BIC, BCC
CFAs & EFAs
CFAs and EFAs
• CFAs include a certain level of systematic measurement error in
the form of cross-loadings. Given that items are rarely pure
indicators of their corresponding constructs
• at least some degree of construct-relevant association can be
expected between items and the non-target
• recent review of simulation studies (Asparouhov et al., 2015)
showed that even small cross-loadings (as small as 0.100)
should be explicitly taken into account
CFAs and EFAs
• the Exploratory Structural Equation Modeling (ESEM)
framework (Asparouhov and Muthén, 2009; Marsh et al., 2014)
has been developed which incorporates the advantages of the
less restrictive EFA (i.e., allowing cross-loadings) and the more
advanced CFA (i.e., goodness-of-fit or multigroup models) at
the same time, providing a synergy that is “the best of both
worlds”
Software
• LISREL 9.1 from SSI (Scientific Software International)
• IBM’s SPSS Amos
• EQS (Multivariate Software)
• Mplus (Linda and Bengt Muthen)
• CALIS (module from SAS)
• Stata’s new sem module
• R (lavaan and sem modules)
AMOS
STATA
Books….
• Barbara M. Byrne (2012): Structural Equation Modeling with Mplus, Routledge Press
• She also has an earlier work using Amos
• Rex Kline (2010): Principles and Practice of Structural Equation Modeling, Guilford Press
• Niels Blunch (2012): Introduction to Structural Equation Modeling Using IBM SPSS Statistics and
Amos, Sage Publications
• James L. Arbuckle (2012): IBM SPSS Amos 21 User’s Guide, IBM Corporation (free from the Web)
• Rick H. Hoyle (2012): Handbook of Structural Equation Modeling, Guilford Press
• Great fit index site:
• http://www.psych-it.com.au/Psychlopedia/article.asp?id=277
Demonstration in AMOS
• C:\WORKING PAPERS\ICAER2022