Wk 11 - SEM

Question 1

What are latent variables? (x1)

Answer 1

Unobservable psych constructs

Question 2

How does SEM relate to CFA? (x3)

Answer 2

An extension of:

* Theory-driven approach to hypothesis testing
* Based on the same mathematical principles

Question 3

In characterising relationships between observed and latent varaibles, how is SEM like (x1), and different (x1) to CFA?

Answer 3

Can model correlational associations

Can also be used to model causal relationships (‘causal’ may be a little strong…)

Question 4

What assumptions are made by SEM (that it shares with CFA)? (x2)

Answer 4

Multivariate normality

Linear relationships among variables

Question 5

What is the typical focus of SEM? (x1)

Answer 5

Testing hypotheses about relationships between latent variables

Question 6

What methods are subsumed by SEM? (x3)

Which make it…(x1)
Despite… (x1)

Answer 6

CFA,
ANOVA,
Regression

Very powerful and flexible analytic technique
(but some restrictions on use)

Question 7

What are ‘manifest’ or ‘indicator’ variables in SEM? (x1)

Answer 7

Observed/measured variables

Question 8

How are manifest/indicator variables’ variance calculated in SEM? (x2)

Answer 8

Unique for each variable,

Estimated from data

Question 9

What constitutes the variance in latent variables in SEM? (x1)

Answer 9

Variance shared by subsets of manifestt/indicator variables

Question 10

In what ways might latent variables relate in SEM? (x2)

Answer 10

Correlations (bidirectinal arrows)
Causal paths (unidirectional)

Question 11

What are exogenous latent variables in SEM? (x2)

Answer 11

Those not caused by other FACTORS in the model

Summarise variance shared by subsets of OBSERVABLE variables

Question 12

What are endogenous latent variables in SEM? (x2)

Answer 12

Those predicted by other FACTORS (latent variables) in the model
Summarise hypothesised causal relations between constructs summarised by the exogenous VARIABLES

Question 13

What is ‘disturbance’ in SEM? (x1)

Answer 13

Error terms associated with endogenous latent variables

Question 14

What do direct causal paths describe in SEM? (x2)

Answer 14

Effects of a predictor variable,

Controlling for effects of all other variables

Question 15

What are the 2 stages of conducting an SEM?

Answer 15

Measurement model:
*CFA to specify factor structure of subsets of variables

Structural model:
*Examine relationships among latent variables ID’d by measurement model

Question 16

When do you specify order of variables’ association in SEM? (X2)

Answer 16

For masurement model,

And then for structural

Question 17

What model fit evaluations are required in SEM? (x3, x2)

Answer 17

Fit indices (same as CFA):

* Chi-square, plus
* Relative indices

Comparison with other plausible models
*Can matter less for measurement models

Question 18

Under what conditions might you refine the model in SEM? (x2)

Answer 18

Free up additional parameters if initial fit is poor

Fix parameters to zero if they don’t contribute to fit (i.e., are ≈ 0)

Question 19

Considering fit as well as parsimony, parameters should only be free/estimated if they… (x2)

Answer 19

Clearly improve model fit

*Contribute to more accurate depiction of data

Question 20

How are estimated parameters evaluated in SEM? (x2)

But they should also be… (x1)

Answer 20

Chi-square, or
Akaike Information Criteria

Theoretically justified, not just ad hoc data fitting

Question 21

Considering fit as well as parsimony in SEM, we should ask ourselves? (x1)
Because… (x1)

Answer 21

Is the model working due to complexity, or a good fit?

SEM has buckets of free parameters, so lots of additional flexibility…

Question 22

Why is the traditional CFA chi-square test not necessarily useful when comparing alternative SEM models? (x3)

Answer 22

Because models in CFA are nested
SEM might compare qualitatively different structures
*ie differences in order of variable associations

Question 23

What issue arises when comparing non-nested models in SEM? (x2)

Answer 23

A comparison model with (eg) more free-parameters
Will generate a higher chi-square
Meaning that more parsimonious models are disadvantaged in analysis

Question 24

What is the solution to issue of biased chi-squares generated by comparing non-nested models in SEM? (x1, plus explain x2)

Answer 24

AIC (Akaike Information Criterion)
= chi-square + 2k
*where k = number of free parameters

Question 25

What does the AIC do in practice? (x2)

Answer 25

Introduces a penalty for model complexity

So that simpler models can compete

Question 26

How should the AIC(Akaike Information Criteria) be interpreted? (x2)

Answer 26

It summarises the fit/flexibility trade-off

So choose model with lowest AIC

Question 27

How should you plan sample size for SEM? (x3)

Answer 27

Same as CFA:

* Aim for 10+/parameter
* 5+ may suffice (reduced power)

Question 28

What are the 3 steps for conducting SEM? (x2, x2, x3)

Answer 28

Specify measurement and structural models
*ID fixed and free parameters
Estimate parameters of measurement model
*Max observed/predicted fit for variance/covariance matrices
Examine model fit/parameter estimates
*Refine till well fitting/parsimonious
*Test alternatives and compare fits

Question 29

What issues might constrain the use of SEM?

Answer 29

Sample size too small - need big for reliable estimates
Insufficient number of indicator variables (measurement model)
Poor fit of measurement model to data

Question 30

How many indicator variables are required for measurement models in SEM? (x2)

Answer 30

Single-factor: 3

2+ factors: 2+ for each latent variable

Question 31

Why is a poor-fitting measurement model an issue for SEM? x1

Answer 31

Cannot successfully identify latent variables underlying the data set

Question 32

What method could you use instead of SEM if you don’t quite meet all the requirements? (x1, plus describe x4)

Answer 32

Path analysis:

* Creates composite variables by averaging observed measures
* These play role of latent variables 
* So determine order of association, and
* Evaluate fit/alternative models (as per CFA/SEM

Question 33

While asking similar questions to the structural models in SEM, the earlier technique of Path Analysis was limited by… (x1)
Because of… (x1)

Answer 33

Inability to explicitly model error/unique variance

Because of lack of latent variables to account for shared variance