Qualitative Participation Grading (and Quantifying it)

As discussed on the Grading page there is a qualitative aspect to discussion grading. While your initial posts are checked to make sure they meet the minimum (those grades you’ve been getting), the quality of your postings and participation counts as part of your discussion grade and your participation grade. We also said we won’t look at individual weeks: we’ll try to take a “holistic” view (since some weeks you may have more to say than others).

A downside of this qualitative holistic aspect is that it is hard to provide you with feedback: both in terms of “real” feedback (to help your learning), and “grading feedback” (so you can make sure you are on track to get the grade you want, if you care about such things). But, we will try to give you some feedback.

One thing we will do is provide quantitative measures of your online discussion performance. We emphasize that these are crude estimates and will really read things in determining grades.

I hesitate to even give you the quantitative metrics, because they don’t tell the whole story, and I do not want to imply that they are what we base grades on. However, I think they can be useful feedback.

So…

We will provide you with quantitative measures of your discussions for a range of weeks (we’ll probably do it for the first four weeks). They can be helpful to give you some idea of where you stand, but remember, they are no substitute for actual content reading (that we will do for real grading).

The metrics are computed over both the “online discussions” and the “seek and finds”.

Some metrics to consider:

  • Number of posts: Each week there are 3 required posts, and we’ve asked people to try to average at least 3 postings beyond that. But this is 6 posts per week (on average). The numbers here aren’t precise (some people put responses to both prompts in one post in early weeks), and quantity is no substitute for quality. But if you are way below expectations, you might consider being a more active participant. (if you’re curious, the median was 22 posts in the first 4 weeks, there were 4 people who posted more than 40).

  • Number of replies received: If your posts receive replies from others, that’s usually a good sign that your posts are interesting (or, you have a chatty group). These numbers are not corrected for group size. Not a strong signal, but something.

  • Number of likes received: If others like your posts, that’s a good sign that you wrote something that someone appreciated. Not a strong signal, but something.

  • Number of characters written: Quantity of characters is not quality of content - its a rough measure of real length, and length doesn’t necessarily mean that the post has good content. A post can be verbose and say nothing, or be short and make a good point. However, there is usually some correlation if you average over several weeks: there is a minimum that it takes to say something sufficient, and generally, people tend to write more when they have something useful to say. (if you’re curious, the median is around 18000 characters written in postings over the first 4 weeks)

  • Self Replies: This is a count of posts that are a reply in a chain that has an earlier post from the same author. It isn’t that useful (I thought it would help me answer a specific question, but it really doesn’t).

  • Initial Posts: This is a count of posts that do not have parent postings. Your “initial” postings each week (of which there should be generally be 3) are counted here. This is a metric of how well you have been following directions.

We will provide you with these metrics, for the first few weeks.