Psychology (PSYC) 757 is a course devoted to introducing graduate psychology students to Bayesian statistics. The course was redesigned from previous incarnations to increase student involvement and reduce the workload to a manageable level. Also, in previous versions, I attempted to implement a standard lecture (2010), problem-focused (2013), and flipped-learning (2016) formats - each offered some advantages to students but none satisfied my goal (see below). Last year (2018), I implemented a group, project-based approach with a blend of flipped-learning, lecture, and problem-focused pieces; this year, I plan to do the same given the enthusiastic feedback from students. My goal with this class is to provide every student the opportunity to implement basic Bayesian methods to standard social/behavioral science data analysis problems.

The purpose of the class remains focused on Bayesian statistics, however, unlike many Bayesian devotees, I consider Bayesian statistics a tool. Just like other tools, Bayesian methods have suitable and unsuitable applications. Students enrolled in my course may find my approach more challenging than standard graduate lectures because I emphasize mastery (i.e., learning) over achievement (i.e., grades). I demand more from you (and me) so that we all learn together. If this format is more difficult, why would I implement it? Simple. I want students to learn, not just get grades for taking a class. The more my students learn, the more future professional opportunities await them.

I require no textbooks for this course for two reasons. Reason 1: the material becomes outdated quickly given the quick pace of Bayesian tool development. Reason 2: an abundance of online and free resources exist so that you can continue to learn with the most updated materials. I link below and on the course website all the relevant material you ought to read. Feel free to send me any other resources you find useful.

I intend to assign one of three grades to you based upon your group’s performance over 10 projects. All students will participate in a single group for the duration of the semester. I assign students to the groups. Research indicates that group work in graduate classes translates well to professional activities. Moreover, your group will be graded as a whole - just like your professional work will be viewed as an end product by a team. Do your best on each assignment and grades will not be an issue. I have 10 projects assigned throughout the semester (roughly 1 assignment per week). Your group’s performance will be judged by both me and the rest of the class. I reserve the right to overrule the class if they collectively judge groups too stringently. The grades for all 10 projects will be communicated to students by demand. Your individual grade will be based upon the average of those projects throwing out the two worst grades. Thus, you will be graded on 80% (8 out of 10 projects) of your work. Each project is worth 10 points for a total of 80 points for the course.

Every Tuesday, teams present their project results to the class in 3 minute sessions. Each group will present their results using the following format (numbers correspond to the slide):

1. The Problem

2. The Solution

3. The Results

4. Lessons Learned

Please do not deviate from the format provided; no more and no fewer slides permitted. Think like scientists when presenting the group project. You may select one presenter or divide the slide presentation among your group members but the transitions must be swift because you only have 3 minutes to present your project.

Although a B- is a satisfactory grade for a course, students must maintain a 3.00 average in their degree program and present a 3.00 GPA on the courses listed on the graduation application.

The following projects and the descriptions for each will be found on the linked documents. Check back throughout the semester for the most up-to-date project details - they may change as I gather more information on student progress.

1. Week 2: Start using R to solve basic probability problems

2. Week 3: Compute conditional probabilities in R

3. Week 4: Use bayess package to solve one normal distribution problem

4. Week 5: Apply Bayes Rule

5. Week 6: Show me the money: Plotting your results

6. Week 7: Compare frequentist and subjectivists results from a single source

7. Week 8: Use rstan to solve a simple Bayesian problem

8. Week 9: Calculate and interpret correlations using Bayesian methods

9. Week 10: Stan and R for general linear model estimation and inference

10. Week 11: R, stan, python applications for Bayesian inference

I use Google Sites, Groups, Forms, Docs, and Sheets for the course - as I do for all my other courses. You will not find any of the material on Blackboard because I have found that platform too unreliable over the years. Google is a toaster - it simply works. I highly recommend you use Google Meet or Zoom for your team work outside regular class hours but use whatever tool works best for you and the rest of your team.

Disability Services at George Mason University is committed to upholding the letter and spirit of the laws that ensure equal treatment of people with disabilities. Under the administration of University Life, Disability Services implements and coordinates reasonable accommodations and disability-related services that afford equal access to university programs and activities. Students can begin the registration process with Disability Services at any time during their enrollment at George Mason University. If you are seeking accommodations, please visit http://ds.gmu.edu/ for detailed information about the Disability Services registration process. Disability Services is located in Student Union Building I (SUB I), Suite 2500. Email:ods@gmu.edu | Phone: (703) 993-2474