---
tags: stat340, syllabus
---

# STAT 340 Syllabus, Fall 2022
**Instructor**: Adam Loy

## How to use this syllabus

This document contains all the information you need to navigate the course. **This syllabus is meant to be read once, then searched as needed.** If you need to find something, use the table of contents found in the left sidebar of the web version. Click "Expand All" to see the subsections in the table. You can also hit `Control-F` on windows or `Command-F` on a Mac to bring up a search field, and then type in the text you are looking for.

+ When you see blue- or purple-underlined text in the syllabus or any other document, it's a clickable link. For example, [click here for a cat video](https://www.youtube.com/watch?v=aFuUidBR1aQ).
+ Links to all important documents and information can be found on the course website. **Please check the course website daily.**
+ The syllabus is available in electronic form and as a PDF. The electronic version will be consistently updated as needed, with updates ==highlighted yellow like this==. *The PDF version will only be updated occasionally and you should use it only for archival purposes.*

## Key Information

**Meetings:** MWF 5a (i.e., MW 1:50-3pm, F 2:20-3:20pm) in CMC 319.

**Student drop-in hours:** No appointment is necessary for drop-in hours, just stop by **CMC 307**

- Monday 3:30-4:30 pm
- Tuesday 1:30-2:30 pm
- Wednesday 3:30-4:30 pm
- Friday 12:30-1:30 pm

**Schedule an appointment:** Use my [Calendly page](https://calendly.com/aloy-meetings) to schedule an individual meeting time.

**Contact me:** Slack is preferred for most question, but e-mail me ([e-mail]) with personal (e.g., illness- or grade-related) questions. You can also schedule an appointment on [my Calendly page](https://calendly.com/aloy-meetings). Be sure to read [my availability/response policy](https://hackmd.io/@aloy/HkZv3Q6d5#Instructor-availability-and-message-responses).

**Course calendar:** The official course calendar is in [Appendix B](https://hackmd.io/@aloy/HkZv3Q6d5#Appendix-B-Class-Calendar). *In case of a date conflict on assignments or course documents, the Class Calendar is assumed to be correct.*

**Textbook:** The textbook is [Probability and Bayesian Modeling](https://bayesball.github.io/BOOK/probability-a-measurement-of-uncertainty.html) by Albert and Hu. You can buy a physical copy of the book, but you can access a free e-version through the link.

**Software and technology:**
* We will use the R programming language via the RStudio IDE for computational work in the course. You can access Carleton’s RStudio Server at https://maize.mathcs.carleton.edu using your Carleton credentials. If you plan to be off campus, then you will need a VPN to access maize.
* The [Gradescope](https://www.gradescope.com/courses/435021) grading platform will be used to collect and return most graded assignments. Be sure to check Gradescope for feedback on your submissions.

**Course webpage:** Our course webpage is at [carleton-stat340.github.io/fall2022](https://carleton-stat340.github.io/fall2022/). Course materials, including notes, daily prep, study guides, and assigments, will be posted here. Check the course webpage daily.

## Course Overview

For decades the world of statistics was dominated by “frequentist” methods. Bayesian statistics is an alternative school of thought founded upon the idea that our beliefs about the world are constantly revised with the incorporation of new information. While this idea is intuitive, Bayesian statistics was held back by the mathematical intractability of common inferential tasks. Computers have changed that. Today, Markov chain Monte Carlo (MCMC) methods are used by Bayesians to conduct statistical inference. In this course, we will explore the Bayesian philosophy and approach to statistical inference. We will start with the basic building blocks of inference and then explore the Bayesian regression model. Along the way, you will learn how and why MCMC works, enabling you to simulate from posterior distributions so that you are not restricted to models that can be “fit” using pencil and paper.

## Course Goals

After completing this course you will be able to do the following:
1. Given a question of interest, develop an appropriate Bayesian model that incorporates your prior belief about the parameter(s) of interest.
2. Given a Bayesian model and question of interest, derive or approximate the appropriate distribution (posterior or predictive) needed to answer the question.
3. Given a posterior distribution and question of interest, construct and interpret appropriate summaries to answer the question.
4. Comment on the adequacy of a model and compare competing models.
5. Clearly communicate the results of an analysis and its implications.

## Assessments and Grades

You can only learn by doing things. The things you will do to learn include many things that are not assessed or graded, along with some items that are formally assessed and graded.

#### Formally assessed and graded work

* **Learning Target completions**: Learning Targets are the basic skills that you should learn in the course. You will be asked to "complete" as many Learning Targets as possible, done mainly by taking short quizzes on them at regular intervals. See "Learning Targets" below for more.
* **Case studies and a final project**: Case studies allow you to apply the skills found in the Learning Targets in a complete analysis cycle. These will not only require you to construct statistical models to solve a problem, but also to communicate your reasoning clearly to an appropriate audience. In a case study, I will provide you with the data, or instructions for data collection. The final project is similar to a case study, but requires that you find your own data set and devise your own research questions.

#### Ungraded work

* **Class attendance and engagement.** Regular class attendance is the easiest way to learn the course material, practice new skills, ask questions, and get immediate feedback; however, attendance in class is not mandatory. When you attend class, I expect you to be fully engaged.
* **Reading and review.** I recommend reading the assigned sections of the text before we discuss them, so that you are already working with the ideas in advance of hearing about them from me. In addition, review your class notes after each class, carefully reconstructing for yourself the ideas, arguments, and overall story that is developing. Coming to class three times a week without this extra work will not lead to deep learning.
* **Homework.** You will complete and submit weekly homework assignments, which will generally be due on Fridays. While homework will not receive a formal grade, it will receive useful feedback that will help you learn. It may seem odd that homework is ungraded, but this provides a **no-stakes environment to receive feedback** as you are initially grappling with challenging topics, and only after you have thought about them and received feedback will you be assessed.

:::info
### About the grading system in STAT 340

:exclamation: The way grading works in this course is a little different than you might be used to. Please read this part carefully and ask questions as needed.

**Learning takes place over time through feedback loops.** Consequently, I do not expect to fully understand a new concept after a single exposure. Instead, **I expect that you will need to practice, make mistakes, get constructive feedback, reflect on that feedback, and reattempt.** Allowing multiple attempts will allow you to learn from your past mistakes and demonstrate not only your understanding, but also your growth. In addition, you will not be penalized for having a bad day that results in a bad test score, so long as you can show evidence that you've *eventually* learned what you need to learn.

I also believe that **assigning points to your work is unhelpful** for all parties involved. Points give the appearance of a scientific measurement, but in reality *all grading involves a judgment call by the instructor* based on their professional expertise. I believe your work should be evaluated just like everyone else's work in the real world is evaluated: Have *clearly defined standards* for quality, then I *give detailed verbal feedback* on your work instead of points, then *give you the opportunity to try again* based on the feedback. This gets you into a **feedback loop**, a conversation between you and me about your work, that continues (within reason) until your work meets the standards.

So in STAT 340:

* **There are no points on your work**.
* What you get instead of points is a **simple mark** that summarizes your current progress on the work, and **feedback** on what went well and what needs attention.
* **There is also no partial credit** awarded on your work, because there are no points.
* What you get instead of partial credit is **the ability to revise and resubmit almost any piece of work**, using the feedback to improve and grow.

This is not as weird as it sounds. *It's actually the way all human learning works, and the way most professionals are evaluated in their work*. It only seems strange because it's not how school work is typically graded.

Keep reading for more details.
:::

### How each assignment is graded

Each kind of graded assignment is different, and here is how each is graded:

| Assignment | How it's graded | What's recorded |
| :---- | :------------ | :-------------- |
| Learning Targets | Completeness and correctness | *Successful* or *Retry* (see below) |
| Case Study/<br>Project Components | Completeness, effort, correctness, and communication quality | *Successful* or *Retry* |
| Homework | Feedback is given, marks are not | N/A |

[The "Standards for Assessments in STAT 340" document](https://hackmd.io/@aloy/B1MRYILlj) contains details on the quality standards for each kind of assessment in the course. Please read this carefully and review before each submission you make.

### Learning Targets

Throughout the course, you will demonstrate your learning on each [Learning Target](https://hackmd.io/@aloy/B1MRYILlj?view#Appendix-A-Learning-Targets). The primary way you'll demonstrate your understanding is through **Learning Target quizzes**, given roughly every other week. Each quiz will contain several problems, and each problem covers one of the Learning Targets. Work on a quiz problem for a Learning Target with only a limited number of errors will constitute a "successful demonstration" of skill and will be marked *Successful*. Work on a Learning Target quiz problem that doesn't meet the standards will be marked *Retry*, and you'll have a chance to try again later after further study and practice. See the [Standards for Assessments document](https://hackmd.io/@aloy/B1MRYILlj) for the details on what you'll be asked to do for each Learning Target and the criteria for success.

Each quiz will contain problems for each *new* Learning Target along with *new versions* of quiz problems for earlier Learning Targets. **You do not need to do all the problems on each quiz;** just attempt the problems for Learning Targets that you haven't completed yet and feel ready to try. **Questions for each learning target will appear on at most two written quizzes.**

Your work on Learning Target quizzes will be done by hand, then scanned and uploaded to Gradescope.

**Alternative way of demonstrating skill on Learning Targets:** You can also demonstrate your skill on a Learning Target in the following ways other than quizzes (but you will need to spend a token to do so):

- *An oral learning target quiz in appointment hours*. In this option, you schedule time with me via Calendly. During this meeting I will give you a new problem and you work it out "live". ==You can work on any learning target during an oral quiz, not just those from the most recent quiz.==

Details on using these alternative assessment methods are found in the [Instructions for alternative Learning Target assessments](https://hackmd.io/@aloy/S1GKxpPlj) document. Other ways of demonstrating skill may be introduced later, especially for Learning Targets appearing near the end of the course.

### Case studies and Final Project

There will be two case studies and a final project in this course. Each of these assignments will have multiple components on that will be individually assessed on a *Successful*/*Retry* basis. See the [Standards for Assessments document](https://hackmd.io/@aloy/B1MRYILlj) for more details. You may revise case studies, but revisions are not possible for the final project.

## Earning a Course Grade

Your grade in the course is determined by how many requirements for each kind of assessment you eventually meet. To earn a course grade, satisfy all of the requirements in the row for that grade in the table below:

| | Learning Targets completed (20) | Case study marks (12) | Final project marks (6) |
| :---: | :---: | :---: | :---: |
| A | 18 | 10 | 5 |
| B | 16 | 8 | 4 |
| C | 14 | 6 | 3 |
| D | 9 | 3 | none |

Numbers in parentheses indicate the total number of marks available. An "F" is assigned if not all of the requirements for a D are met.

:::success
**Example:** Alice finishes the semester with the following things accomplished:

* They completed (successfully demonstrated understanding of) 16 Learning Targets.
* They earned 9 marks on the Case Studies.
* They earned 4 marks on the final project.

Looking across the row for a grade of "B", Alice has satisfied all the requirements for that grade, so they earn a "B" in the course.
:::

**Plus/minus grades:** To earn a "plus" grade, complete all the requirements for a basic letter grade **and** the requirements for the next grade up in one of the other categories. A "minus" grade is assigned if you meet all of the requirements for a grade except for one category, and that category is no more than one level lower than the rest.

:::success
**Example:** Briar finishes the semester with the following things accomplished:

* They completed 17 Learning Targets.
* They earned 8 marks on the Case Studies.
* They earned 5 marks on the final project.

Looking across the row for a grade of "B", Briar has satisfied all the requirements for that grade and also satisfied the requirements for an "A" on the final project; thus they earn a "B+" in the course.
:::

:::warning
**Important facts about this grading system**

- **Different assignments do not "average together".** You can't make up for less-than-great work on Case Studies by doing more Learning Targets, for instance. Each course grade requires *consistent quality on all assignments* to earn the grade.
- **Once you have satisfied all the requirements for a grade, you've earned that grade, and it never goes down.** In fact grades only ever go *up* (or remain the same) in this system.
- **You do not have to do everything.** If you want an "A" in the class, for example, you do not have to complete *every* Learning Target, only 18 of them.
:::

## Revision and Resubmission

Except for the final project, **you can revise and resubmit any work you turn in, without penalty**. How this works depends on the assignment:

**Learning Targets:** As detailed above, work on a Learning Target (via a quiz or other means) that doesn't meet the standards for "success" can be reattempted, through work on a new version of the same problem---on a later quiz or an oral quiz in appointment hours.

**Case studies:** A case study can be revised and resubmitted by uploading a revised version of the work to the same Gradescope "assignment" where it was previously submitted. All versions of the case study will be archived for my reference.

### Limitations on revisions

Although there is no penalty for revising and resubmitting work, I need to place reasonable limitations on revisions to be able to provide feedback quickly enough:

+ **Only two case study revisions are allowed.** This can be two revisions of the same case study or one revision of each case study.
+ **A case study must be completed with a good-faith effort to be eligible for revision.** If I deem a submission to be "not assessable" due to lack of effort, then it cannot be revised. See the [Standards for Assessments document](https://hackmd.io/@aloy/B1MRYILlj) for details.
+ **Only one request per week for oral quizzes on Learning Targets may be made**, and only one Learning Target per quiz is allowed.
+ **No requests for oral Learning Target quizzes may be made after 11:59 pm on Sunday, November 13** so that the last week of class can be used to complete the requests that have been made.
+ **No revisions are allowed after 11:59 pm on Wednesday, November 16** in order to provide time and space to mark all pending revisions for final course grade determinations.

## Academic integrity

You are expected to follow the [policies regarding academic integrity](https://apps.carleton.edu/handbook/academics/?policy_id=21359) established by Carleton College. Any work you submit for a grade should be your own, not a facsimile of a classmate's work or an online solution, which would constitute academic dishonesty. To check if your homework meets this standard, imagine I asked you to explain your reasoning for each problem---you should be able to do so with ease using language similar to your submission. Finally, cell phones are prohibited during learning target quizzes.

Cases of academic dishonesty will be dealt with strictly. Each such case will be referred to the Academic Standing Committee via the Associate Dean of Students or the Associate Provost. A formal finding of responsibility can result in disciplinary sanctions ranging from a censure and a warning to permanent dismissal in the case of repeated and serious offenses.

Here are the specific parameters for how much and with whom you may collaborate on each piece of graded work in the course:

| Assessment | Collaboration allowed |
| :-------- | :--------------------- |
| Learning Targets | **No collaboration is allowed at all** with other people (including online forums such as Stack Exchange or Reddit). You are allowed to use resources from the class such as the textbook, your notes, and your past quiz work; but getting answers or significant parts of solutions from outside print or electronic resources is not allowed. |
| Case Studies | Each case study will be comprised of group and individual components. You must adhere to the guidelines given for each case study. |
| Homework | You are encouraged to collaborate on these, and you can use outside sources. But your work must be your own and reflect your own understanding (i.e., it can't be copied). |
| Final Project | You are expected to collaborate with your group, but cannot rely on external sources (people or online sources) to guide the *statistical content* of the project. You are allowed to ask for R help in the Stats Lab or to search R-related tutorials. |

**There's no need to be academically dishonest here, because you can revise and resubmit almost everything.** Rather than cheat, *ask questions and use the feedback loop* so that you really grow and learn in the course.

## How to Get Help

You will almost certainly find yourself lost, stuck, or confused on *something* in this course. This is not a defect in your character or intelligence; it's a sign you are being challenged, and you can turn that challenge into real growth by **seeking out help as soon as you need it**.

Make every effort to get yourself unstuck and resolve your questions on your own first. But then:

* **Post a question on Slack.**
* **Attend drop-in hours and ask questions there.**
* [Schedule an appointment through Calendly](https://calendly.com/aloy-meetings) if drop-in hours don't work for you.
* **Work with a classmate** especially on homework, as long as you're staying within bounds on academic honesty.

## Course Policies

### Classroom culture

All people in this class deserve to feel safe, respected, and valued. That means that all members of our class community are responsible to each other to make sure that all voices get heard, all comments are considered respectfully, and everyone has a chance at success. Determination, cooperation, and hard work are highly valued in this class; helping your neighbor understand the material is more important than trying to be the first to answer. We flourish as a community when every individual participates and learns.

### Attendance, absences, and participation

**Attendance:** Attenance is expected at all class meetings, as I believe it is critical to your learning, but will not be graded.

**Absences:** If you must be absent, you do not need prior permission or justification, but a heads-up is appreciated. However *please avoid any non-essential absences* such as skipping class, or scheduling a trip during a day where class is scheduled. If you miss, you are solely responsible for catching up.

**Missing Learning Target quizzes:** If you miss a class meeting during which there is a Learning Target quiz and don't make prior arrangements with me (these are clearly marked on the Class Calendar), then *there are no makeups offered*. Instead, just attempt the problems you intended to take on the next quiz, or through an oral quiz.

**Participation in class:** You are expected to participate actively during each class meeting, but this will not be graded.

### Deadlines and late work

Homework assignments, case studies, and the final project all have deadlines. Those are handled in different ways:

+ *Homework* assignments have a deadline posted on Gradescope. Since homework does not contribute to your grade, only assignments received by this deadline will receive feedback. This will also allow me to post solutions to the assignments quickly.
+ *Case studies* have "soft" deadlines --- suggested deadlines that give you something to put on the calendar and motivate you to complete it. If you intend to turn in a Case Study but don't think you can finish by the deadline, **you can ask for an extension** by emailing me. Tell me what the issue is, and tell me when you plan to turn in the work. Extension **requests must be received before the original deadline, and I have the right to refuse or give you a different extended deadline if the situation warrants**.
+ The *Final Project* has a "hard" deadline. I cannot accept projects that are submitted late unless a formal extension is granted by the Dean of Students.

### COVID-19 related policies

#### What are the health and safety protocols?

Carleton's culture of accountability and respect remains essential as we move into the 2022-23 academic year. We have an obligation to protect one another and the members of the Northfield community, and we all must continue to take that responsibility seriously. To ensure the well-being of each other and the broader community, we will

- follow the College policies on testing, quarantine, and isolation.
- [report positive COVID-19 test results](https://docs.google.com/forms/d/e/1FAIpQLSf-W0_2mei7QQ7pWK4f0kAYTIBB-fHbwnkvdyE54KeRcPWCgQ/viewform) to the College as soon as possible.
- stay home when sick. (Even if you don’t have COVID-19, you should stay home if you aren’t feeling well.)
- wear a mask if you were a close contact, regardless of whether you have symptoms.
- follow the College mask-wearing policy.
- not eat in class while masking is required.

#### What happens if you (the instructor) cannot be in class?

While I am vaccinated, breakthrough infections can still occur. There is a possibility that I will need to miss class this term either due to COVID symptoms or to care for a family member. If I cannot be physically present for a class meeting, then you will receive a notification through both Slack and email on or before that day. In most cases, class will still be held, but in a synchronous online format using Zoom. Please check Slack and email daily.

#### What happens if I (the student) cannot be in class?

Similarly, it is likely some folks will need to miss class for illness throughout the term. Please reach out to me as soon as possible if you need to miss class because of COVID-19 symptoms, the need to quarantine, or the need to isolate so that we can make arrangements for your continued engagement in the course. While each person's needs may be unique, there are a few guiding principles we'll use to help you remain engaged in the course.

- All materials (slide, handouts, etc.), assignments, and announcements will be posted on the course webpage.
- There are options for virtual student hours. In addition, you can set up an appointment via Calendly.
- Each class day, we will have two designated note takers that will add notes to a Google Doc. These notes will be shared with the entire class so that you can "get the notes from class" without any additional hassle. This has the added benefit of creating a community of learning and caring, and also allows me to check on class understanding.
- I will not record individual lectures, or replicate lectures during office hours. I understand that this might help you engage in the course; however, this would be a huge time investment that would detract from the energy I have to meet all of our course goals---I'd be teaching the course twice. The collaborative daily notes provides a rich alternative and allow me to provide the best course experience for the entire group.
- If, after working through the collaborative notes and readings, you feel that you need a "lecture" to truly learn the material, please talk to me. There are alternative materials that I can suggest on a case-by-case basis.

### Instructor availability and message responses

You can ask a question at any time. You can use Slack (preferred), email ([e-mail]), drop-in hours (either in person or virtual), talk to me after class, or make an appointment. Slack is usually the best venue for conceptual questions since someone likely has the same question.

Please note that I do not always respond immediately to messages. I do not typically check email or Slack messages between 9 pm and 7 am on weekdays, and I check them infrequently over the weekend in order to devote time to family, rest, and finding balance. Messages received during these times will receive attention once I am back online. Otherwise you can expect to receive a response to your message within a few hours, often much sooner. If you post questions to Slack, you are likely to receive responses faster.

### Accommodations and assistive technology

Carleton College is committed to providing equitable access to learning opportunities for all students. The Office of Accessibility Resources (Henry House, 107 Union Street) is the campus office that collaborates with students who have disabilities to provide and/or arrange reasonable accommodations. If you have, or think you may have, a disability (e.g., mental health, attentional, learning, autism spectrum disorders, chronic health, traumatic brain injury and concussions, vision, hearing, mobility, or speech impairments), please contact [e-mail] or call Sam Thayer (’10), Director of the Office of Accessibility Resources (x4464), to arrange a confidential discussion regarding equitable access and reasonable accommodations.

The Assistive Technologies program brings together academic and technological resources to complement student classroom and computing needs, particularly in support of students with physical or learning disabilities. Accessibility features include text-to-speech (Kurzweil), speech-to-text (Dragon) software, and audio recording Smartpens. If you would like to know more, contact [e-mail] or visit go.carleton.edu/aztech

### Course materials assistance

I recognize the potential financial burden of textbook costs. If you are in need of assistance to cover course expenses, please speak with me by the end of Week 1.

## Appendix A: Learning Targets

There are 20 Learning Targets in the course overall. These represent the basic skills that are available to learn in the course.

### Univariate Models
1. Given your prior belief, specify an appropriate prior distribution for a univariate model.
2. Given the prior distribution and data, derive the posterior distribution for a univariate model.
3. Given the posterior distribution and a research question, estimate the parameter (or function of parameters) of interest and interpret the results in context.
4. Given the posterior distribution and a research question, conduct a Bayesian hypothesis test and interpret the results in context.
5. Given a univariate Bayesian model, derive the posterior predictive distribution and use it to make predictions about future observations.
6. Assess the adequacy of a univariate Bayesian model.

### Multivariate models
7. Given your prior belief, specify an appropriate prior distribution for a multivariate model.
8. Given the prior distribution and data, derive the posterior distribution for a multivariate model.
9. Given a model specification, explain the steps of the Metropolis algorithm.
10. Given a model specification, explain the steps of the Gibbs sampler.
11. Given MCMC draws, check whether they have converged to the posterior.
12. Given draws from the (approximate) posterior distribution, draw inferences about the appropriate parameters in the context of the research question.

### Hierarchical models
13. Given your prior belief, specify an appropriate prior distribution for a hierarchical model.
14. Given a research question and your prior belief, write out a hierarchical model using two-stage priors in statistical notation.
15. Given draws from the (approximate) posterior distribution of your hierarchical model, draw inferences about the appropriate parameters in the context of the research question.
16. Given a fitted hiearchical model, check whether the fitted model is appropriate/adequate.

### Regression models
17. Given your prior belief, specify an appropriate prior distribution for a regression model.
18. Given a research question and your prior belief, write out a regression model in statistical notation.
19. Given draws from the (approximate) posterior distribution of your regression model, draw inferences about the appropriate parameters in the context of the research question.
20. Given a fitted regression model, check whether the fitted model is appropriate/adequate.

## Appendix B: Case study learning targets

### Multivariate model
1. Given a research question, construct a Bayesian model that captures the information on the parameter prior to conducting the study.
2. Clearly communicate the prior information you incorporate into the model.
3. Given your data and model, implement an appropriate procedure to sample from the posterior distribution in R.
4. Clearly summarize the relevant posterior information.
5. Clearly communicate an analysis and its implications in written format.
6. Adhere to standards for technical writing.

### Hierarchical model

1. Given a research question, construct a Bayesian hierarchical model that captures the information on the parameter prior to conducting the study.
2. Clearly communicate the prior information your are incoporating into the model.
3. Given your data and model, implement an appropriate MCMC algorithm to sample from the posterior distribution using JAGS via R.
4. Clearly summarize the relevant posterior information.
5. Clearly communicate an analysis and its implications in written format.
6. Adhere to standards for technical writing.

## Appendix B: Class Calendar

The calendar can be accessed directly [at this link](https://calendar.google.com/calendar/embed?src=c_nppromkgk7tm2iclrqdtn6idfo%40group.calendar.google.com); it is also available on the course webpage.

<iframe src="https://calendar.google.com/calendar/embed?src=c_nppromkgk7tm2iclrqdtn6idfo%40group.calendar.google.com&ctz=America%2FChicago" style="border: 0" width="800" height="600" frameborder="0" scrolling="no"></iframe>