Grader guide

While the grader is the intended audience here, others are welcome to read it.

General info

• Responsibilities section of the job post

• Schedule

• In grading/Discussions:

  • We want to try giving students just enough hints to figure it out without giving them the answer.

    • If they seem totally lost, direct them to office hours.

  • Don’t spend a ton of time trying to figure out something that you don’t understand; feel free to escalate to the instructor.

• For students seeking one-on-one help, direct them to office hours.

  • If someone asks to meet with you specifically, you’re welcome to do so, but not required.

• Notifications:

  • You may want to tweak your Ed Discussion notification settings.

• Log hours for any time you put in related to this class, including any learning you’re needing to do yourself to answer questions.

• Subscribe to all the Discussion Topics so you get notified.

• Keep an eye out for students who I should encourage to apply as a grader next term. Things to look for:

  • Being consistently helpful in the Discussions

  • Clean, well-documented solutions for the homeworks

  • Asking good questions

Weekly cadence

Weeks start/end at the beginning of each class.

• Grading assignment submissions and resubmissions and releasing grades for your section

  • Feel free to grade things as they come in, in the order received, to give those students more time for resubmission

  • Please try and be done with grading of an assignment within four days after it’s due (so they have time for resubmission)

Between-class participation tracking

Helper notebook.

• We can be fairly forgiving/generous with what counts as completion

• Every student should have each week marked one way or the other

• The instructor will mark participation for students that came to office hours

• You can use the Analytics from Ed to help

Discussions

• Help page

• We are trying to strike a balance between students getting accurate answers quickly and encouraging students to help one another to cement their learning

• Ensure Discussion questions have answers within the specified timeline.

  • Wait 24 hours to respond to questions that could be answered by another student, giving them a chance to do so.

    • Make sure homework questions have an answer within 48 hours, since they are time-sensitive.

    • Within 24 hours of when homework is due, answer questions as soon as possible to get students unstuck.

• Please give corrections/clarifications on student answers where necessary.

• If posts have the wrong Category, are a Question when they should be a Post or vice versa, please fix.

• Mark correct answers as Accepted, if they aren’t already

Check-in meeting

• How’s the workload?

• Anything you need clarification on?

• Any Discussions the instructor should jump in on?

• What came up in Discussions/assignments (common problems, etc.) that might be useful to cover in class?

• Are all cells in the gradebook filled in (through last week)?

Assignments

• Use annotations to leave comments within the PDFs.

• Scoring and regrade rules

• How to give extensions — see “Add special access to an assignment”

  • Grant any request for 1-2 days made before the deadline; escalate others to the instructor

  • Set the End Date to the original late submission deadline or the new due date, whichever is later

  • Don’t give extensions on the resubmission deadline unless authorized by the instructor

• Solutions folder will be shared with you from Google Drive

  • Instructor will share them with students via JupyterHub

  • The students don’t need to match the provided solution exactly, as long as they do what the question is asking

• Grader will manually apply late penalty

Grading

How to grade in Brightspace

You are checking student submissions against the solutions. That said, student code/output doesn’t need to look exactly like what’s in the solution, as long as they’re doing what’s asked for in each Step. When grading, points should only be deducted based on these criteria. Please leave comments for:

• Point deductions, explaining what it’s being deducted for

• Feedback like “this could be done better/differently,” even if there isn’t a corresponding point deduction

Checks

The following should be true for each Assignment:

• The description is a link to the assignment page on this site

• Points

  • 100 points per Assignment, except for Homework 3 and the Final Project Proposal which are 50 each

  • Percentage of the overall grade matches the breakdown in the syllabus

• Grouped and ordered in a logical way

• Display Grade as: Percentage

• Submission Type: Online, Website URL

• Final Project Proposal is a Discussion

• Dates match the schedule:

  • Due date

  • End Date

• Associated with the Homework gradebook category

• Visible

Plagiarism

See the class policies for more details for what constitues plagiarsm vs. fair reuse.

It isn’t your responsibility to look for potential instances of cheating/plagiarism. That said, if you have suspicions of those occuring, you must report them to the instructor. Things you might notice:

• Use of a package we haven’t covered in class

• Using a technique we didn’t cover in class

• Multiple student submisions:

  • Being identical

  • Solving a problem in the same unusual way

Final Project

Proposals

• Students are encouraged to submit before the deadline to get feedback sooner.

  • Please provide feedback on the proposals within four days of submission so that students can get started.

• If the proposal shows effort and follows the format, full credit should be given

• Things to look for (don’t spend too long on these):

  • Will their dataset answer their question?

  • Do they have a question that is objectively answerable?

  • Will it be the right level of challenge for the duration of the project and their skills, not too much, not too little?

• The feedback you will likely give the most often will be something like:

  Your question is good, but you’ll probably be able to answer it in relatively few lines of code. Think about what your follow-up question(s) will be.

Grading

The Final Projects themselves are peer graded. We’re using PeerMark to facilitate the peer grading. Once the peer review deadline passes:

1. Open each submission.

2. For similarity scores over 20%, look at the report.

   • Matches in data output can be ignored.

   • If it seems like there may be instances of plagiarism or you’re not sure, let the instructor know.

3. Calculate the median of the scores from the peers, using that as the final grade.

4. In the Gradebook, give points to the reviewer under the Final Project Peer Review.

Scoring details.