Grading for 02805, 2015 edition

This year (2015), due to incorporating peer-grading, there was a new grading-process in course 02805.

I guess I thought the peer-feedback was enough for you class participants, but given the number of emails I have received from you all, it’s clear that I’ve miscalculated somewhat. In retrospect, I agree that I have done a poor job informing you about the process. Sorry about that!

I did, however, put a lot of thought and effort into the grading this year, so to make things more clear, I thought I’d briefly describe how I’ve done the grading, so we have more transparency in the process.

The final grade has 4 components.

  • Assignment 1
  • Assignment 2
  • Final Assignment
  • Peer feedback performance.

Assignment 1 and 2

For assignment 1 and 2, your evaluation is based on a mix between evaluations given by the TAs and your peer-feedback. The TAs carry more weight than the peer-feedback. The reason I chose to incorporate peer-feedback in the evaluation of Assignment 1 & 2 was a careful analysis of how the TA evaluations (and my own evaluations) correlate with the averaged peer-feedback. It turns out that you guys are really good at giving feedback on the assignments!

Final assignment

When I looked into your performance on peer-evaluating the final assignments, there is much less correspondence between the averaged peer-feedback and my evaluations. I think the reason that the averaged peer-feedback is less accurate in reproducing my evaluations is that I asked the wrong questions. The questions were too general and non-specific. (It takes a lot of experience to reliably answer high-level questions such as “Evaluate the quality of the theory used”, or similar.)

It is not just that I didn’t agree with your evaluations. There are other signs that the peer evaluation of the final assignments are less reliable For example, there was much more variance in your evaluation of individual assignments (a sign that you were agreeing less with each other on some of the questions).

As a consequence there is no peer-component of the evaluation of the final assignment – the evaluation is given by me in collaboration with the TAs.

Peer feedback performance

Your evaluation for peer feedback performance is a mix of (1) the fraction of the assignments that you provided peer-feedback on (not counting the final assignment, since lots of people were informed of that late) and (2) the rating others gave you on how helpful your feedback was.

Combining those evaluations

In combining those 4 elements, the final project counts for roughly 50% of the final grade, and the remaining 50% are from Assignment 1 & 2 (each roughly 20%) as well as peer feedback (approximately 10%).

More details everything

Below, I’m linking to two spreadsheets.

In spreadsheet1 (this is a .tsv file), you can check out your evaluation within each of the 4 categories discussed above.

And in spreadsheet2 (a pdf) you can check out your detailed evaluations for the various components of the final project.

Hope this helps, otherwise send me an email to let me know!