What Is SEM.0.1?

Video Transcript
Download Video
Right-click and save to download

Detailed Description

This is the first in a series of planned screen-capture videos on Structural Equation Modeling.

Details

Image Dimensions: 480 x 360

Date Taken:

Length: 00:04:11

Location Taken: Lafayette, LA, US

Transcript

Slide 1, Speaker:
This video addresses the question, “What

is Structural Equation Modeling?”

Slide 2, Speaker:
First, equations that are “structural”

are those we intend to represent cause-effect
connections in quantitative models.

Typically, we are working with network hypotheses
in structural equation modeling, such as the

one shown here.

Slide 3, Speaker:
Structural equation modeling, or “SEM”,

can be thought of in a number of ways.

The first I would like to mention is as a
framework for quantitative analysis.

The graphic here is meant to make the point
that we use statistical and mathematical tools

within an SEM framework to seek a causal understanding
about the multiple processes operating in

systems.

Slide 4, Speaker:
SEM also provides us with a means for translating

abstract ideas into testable expectations.

Here we show a theoretical proposition.

SEM methods allow us to develop a so-called
causal diagram that translates the general

ideas into a network of expectations.

Using the causal diagram and data, we can
test the hypothesized network, and therefore

the theoretical proposition, arrive at a new
and quantitative understanding.

Slide 5, Speaker:
SEM unfolds as a process, moving from theory

to model to results and eventually back to
theory.

This process supports generating a clarity
of meaning, repeatability, learning, and building

on prior knowledge.

Slide 6, Speaker:
SEM is also a method for learning.

Here, we start with an observation, that older
forest stands that burn show poorer recovery

after fire.

We can propose and test ideas about the mechanisms
behind this observation by using SEM.

In this case, we explicitly test the idea
that fire severity might explain all or part

of the relationship between forest age and
plant recovery.

Stated in a different way, that older stands
burn hotter and that the stands that burn

hotter have weaker recovery.

This is called the “test of mediation.”

Through the use of this kind of test, we can
dig deeper and deeper into understanding systems

by bringing in more variables.

Slide 7, Speaker:
Structural equation models also have a particular

form.

In these models, responses (y-variables) can
depend on other responses (other y-variables)

as well as external predictors (x-variables).

This equational form allows networks of relationships
to be represented.

Slide 8, Speaker:
SEM is also a body of knowledge.

It is important to realize that there is a
great deal of information about SEM, but it

is derived from many different scientific
disciplines.

Slide 9, Speaker:
History has played an important role in shaping

the literature on SEM.

This diagram represents a citation map showing
the historical flow of knowledge among disciplines.

From this perspective, we can realize how
the flow of information, and especially the

lack of flow of information , shapes peoples’
perceptions of quantitative methods, as well

as the language people are familiar with.

Slide 10, Speaker:
SEM is also a community of practice.

There are active online discussion groups,
both general and also ones specifically related

to software packages.

There also exist a number of working groups
focused on particular scientific problems

that use SEM for the solution.

Slide 11, Speaker:
I hope this overview has been useful.

More detail can be found in the written document
summarized by this video presentation.

Also, you can search for examples that relate
to your subject of interest.