Overview

Description

CALENDAR DESCRIPTION:

Course can be repeated for credit. Students may not take this course for further credit if similar topics are covered. See Psychology department website for course description.

COURSE DETAILS:

Provisional Structuring of Substantive Areas

Psyc402/715 is a proseminar in measurement. Structural equation modelling (sem) will be the focus of the spring 2021 offering. Students will, of course, learn how to fit models using the Lavaan package of R and interpret the output. The theoretical underpinnings of sem- matrix and covariance algebra, relation between path diagram and model implied covariance matrix, identification, etc.-, knowledge of which is necessary in order to employ sem in the service of fruitful science, will, however, receive extensive coverage. Three of the distinct branches of sem will be addressed: observed variable path models, measurement models, and full models involving both measurement and structural, regression, components. Students will gain much experience fitting sem models to actual data, and a diversity of models will be encountered. Some special emphasis will be given to longitudinal models (e.g., latent growth model, cross-lag regression model) and test theory models (congeneric-, tau-equivalent-, and parallel- structures, and hierarchical factor structure).

Psyc402/715 will be comprised of two parts:

i. Fundamentals/background
• Introduction [3 classes of model; path diagram; mathematization leading to model implied covariance matrix; identification analysis; testing (estimation, model fit, etc.)]
• Matrix Algebra
• Random variables, random vectors, multivariate distributions
• Expectation algrebra, covariance algebra

ii. Three classes of model
Detailed consideration of each of:
• Observed variable path models (i.e., multivariate regression models)
• Measurement models (i.e., confirmatory factor structures)
• Full models involving both measurement model components, and multivariate regression structures bearing on latent variables

Lectures and lecture notes
Lectures, and the lecture notes on which they are based, are the primary means by which will be covered, the materials of which the course is comprised. The lectures are structured sequentially in accordance with the lecture notes, and proceed through the course materials in a linear fashion. This year, due to Covid, lectures will be filmed and made available to you, along with the accompanying pdf lecture notes- both, on Canvas-, at the beginning of each week. In each lecture, I will be referring directly to the accompanying lecture notes. However, the notes comprise but a sketch. Accordingly, elucidation, fleshing out, and expanding upon will be the standard order of business. Once a filmed lecture is “released”, you will be able to return to it, and view it, throughout the semester.

Assignments
Each of the five assignments will be constituted of some number of parts, the parts made available to you- on Canvas- as the semester unfolds. A given assignment- all of its parts- will be due on the date and time specified on the schedule, below. Structural equation models will be fit using the Lavaan package of R. You should install R, and the Lavaan package, on your computer as soon as possible.
Workouts
There will be a number of “workouts”, relatively brief computational/theoretical exercises, each of which you will undertake on your own and in accordance with your own schedule. Though it is highly recommended that you complete each and every workout (for they are designed to help you consolidate theoretical knowledge and master practical application), they are not for marks and will not be submitted. Rather, there will be made available to you, at the end of the week on which a workout is assigned, an answer key. You may compare your answers to the answer key at your leisure (and, of course, discuss during office hours).
Office hours
There will be weekly office hours, held synchronously (see schedule, below), using the Zoom web conferencing app. Though optional, regular participation in the office hours is highly recommended. It is expected that, among other things, students will use this time to discuss matters related to the filmed lectures, the readings, workouts, and assignments. That is to say, anything whatsoever related to structural equation modeling.
Canvas Discussions
The text-based discussion feature of SFU’s Canvas platform will offer students the opportunity to post questions and comments, and me, the opportunity to answer, clear up confusions, etc. I will do my best to address questions and comments in a timely fashion.
Recapitulation of implications of the university’s response to Covid: The course will be, in the main, offered in a virtual and asynchronous format. There is, of course, the necessity, in a statistics course, of discussing lecture and reading material, and, accordingly, of scheduling some synchronous class time. The synchronous class time that will take place this semester is a weekly, optional (but highly recommended), office hours (see schedule, below), to be held using the Zoom web conferencing app. The following is a compact recapitulation of the implications for each component of the course:

• lectures: Asynchronous. I will film lectures and make them available to you- on Canvas- at the beginning of each week (see schedule, below). Once a lecture is made available, you will be able to view it, and return to it, throughout the semester.
• readings: Asynchronous. When there are readings, unless otherwise indicated, they will be made available to you in pdf form- on Canvas- at the beginning of the week on which they bear (see schedule, below).
• assignments: Asynchronous. Each part of each assignment will be made available to you- on Canvas- at the beginning of the week indicated on the schedule, below. You will undertake each assignment on your own and in accordance with your own schedule, and submit to Canvas, on the due date.
• workouts: Asynchronous. On weeks on which there is a workout, it will be made available to you- on Canvas- at the beginning of the week (see schedule, below). Workouts constitute optional (but highly recommended) work undertaken on your own and in
accordance with your own schedule. They are not submitted. An answer key will be made available to you on Canvas.
• office hours: Synchronous. Synchronous office hours will be held in accordance with the schedule, below. You will receive a link in the first announcement posted to Canvas.
• Canvas discussions: Asynchronous. When the time comes, this feature of Canvas will be activated, and you will be notified.

Note: I have done my best to anticipate the implications of alterations to the course structure, not to mention the simple reality of university life, in response to Covid. However, I have never taught a virtual, largely asynchronous, course in structural equation modeling. I would ask for your patience, and also, that, in the event that you encounter shortcomings, you let me know, and work with me to overcome the problems. We all want for the necessary mode of teaching to “work.” We are, however, sailing into uncharted waters.

Note: All asynchronous components of the course will be contained within modules, organized by week, and hosted on the Canvas platform.

Grading

Materials

REQUIRED READING: