his class is designed to provide participants a working knowledge of structural equation
modeling and the LISREL program
Students in Soc402 can most easily satisfy these requirements by selecting a published
article that estimated a structural equation model, IF that article contains a covariance matrix (or
correlation matrix and standard deviations) for the modelled variables. The student’s model can be
a direct replication of the published model, and the student’s project can report about the things
they learned by doing the replication. The student’s selected article must be OK’d by me in the
first two weeks of class to ensure the article and model do not contain features that are too difficult
for a student project. If an article is used, a link to, or copy of, that article must be provided in the
final report. If you wish to use some other data set please talk to me after our first meeting.
The write-up of your project should be submitted electronically via canvas, and include comments
regarding:
1) the data set employed, the variables, and their coding — plus the article or a link
2) the academic context of your model (e.g. was it: used by others?, a modification of a model
that had been used by others?, your own theory?). Only a very-brief context statement is
required and at least data-documenting references will be required. If the model corresponds
to “your own theory” this should be reported and no theory-reference would be required, but
a data-reference should be provided.
3) difficulties in model specification (if relevant)
4) a model diagram (a clear hand-drawn diagram of the final model is acceptable; un-
polished/rough diagram drafts may be included to assist documenting/clarifying
modeling difficulties or model changes). Remind me to speak about drawing diagrams
using Word.
5) the most challenging and/or focal segments/features of the model
6) your initial/early modeling attempts, and their consequences and diagnostics
7) your final model (estimation adequacy, fit, estimates, diagnostics, measurement, selected
interpretation of key features, latent effects and explained variance, controlling, overall theory
consequences of the current estimates/model, remaining difficulties/challenges, and future
possibilities).
Models that fit, or that fail-to-fit, can both receive top marks if the modeling is
conducted competently, assessed honestly, and reported clearly/appropriately; so beware making
unreasonable/untenable/indefensible model modification to obtain a fitting model.
In the past, students who rigidly followed the above steps did OK in this class but did
slightly less-well than students who treated the above as guidelines covering the basics while they
attempted to focus on the overall impact or importance of what was learned via their modeling
attempt(s). It is possible to consider the above points as you tell “the story” of whatever you are led
to say about how future research could be improved, or in what direction(s) future research is
prodded, by your modeling attempt. For example, as a minor part of your write-up, or possibly as a
portion of your thinking about your model, you might consider what you would claim if another
researcher had a model/theory that contains the same indicators as your model, and that fits about as
