CS 1538: Participation Grade Calculation

There are many ways to earn participation points, including:
- Ask on-topic questions during lecture
- Answer questions during lecture
- Visit office hours regularly
- Complete ungraded surveys on CourseWeb
There will be "quick quizzes" at the end of many lectures. These quick quizzes will ask basic questions from material covered during lecture that day and previous lectures. Because these quizzes focus on material from the lecture, it is not possible to make up missed quick quizzes.

The formula used for calculating participation is:

Grade = 100*(0.5*ParticipationScore + 0.5*QuickQuizAverage)

The QuickQuizAverage value is an average of the percentages of all of the quizzes for the month.

The ParticipationScore value is calculated below:

ParticipationScore = MIN(1, 0.25*ClassAttendance + 0.5*(Emails + OfficeHours + Surveys)/(0.75 * [3 + SurveysGiven]) + 0.7*RankedClassParticipation)

The ClassAttendance value is (# quizzes turned in that month) / (# quizzes given in that month). Excused absenses are taken into account here by excluding that class from the denominator.

Emails are the number of email conversations in the month regarding homeworks, class concepts, or grades (emails about missing class or non-course emails aren't included).

OfficeHours are the number of times in the month the student visited office hours and talked about something related to the course.

Surveys are the number of surveys in the month the student completed. This does not include surveys started but not completed. Also, for the project team formation survey, a student received credit for completing it if they completed it or if they formed a group on their own.

SurveysGiven are the number of surveys in the month that were due.

To explain this part of the formula:

(Emails + OfficeHours + Surveys)/(0.75 * [3 + SurveysGiven])

This calculates the ratio of communication/feedback to me over an estimate of what a good amount of communication/feedback should be. The 3 in the denominator was determined empirically as a good minimum number of times a student struggling in the course should talk with me each month to improve their grade. The denominator is multiplied by 75% to discount the influence of the estimate on the overall participation grade. Note that the ratio can be greater than 1, but is assumed to be on a 0 - 1 scale.

RankedClassParticipation is based off of in-class participation, but curved according to the following:

If ClassParticipation >= 90% (i.e. almost every class), then RankedClassParticipation = 1
Else If ClassParticipation >= 67% (i.e. 2/3 of all classes), then RankedClassParticipation = 0.9
Else If ClassParticipation >= 50% (i.e. half of all classes), then RankedClassParticipation = 0.75
Else If ClassParticipation >= 25% (i.e. a quarter of all classes), then RankedClassParticipation = 0.5
Else If ClassParticipation > 0% (i.e. at least one class), then RankedClassParticipation = 0.25
Else RankedClassParticipation = 0

ClassParticipation is (number of classes during the month in which the student asked at least one relevant question or answered at least one question) / (number of classes attended during the month).

Note that it is possible for students to score greater than 1 in the calculation copied below since the weights of the three parts add up 1.45 and each part is on a 0 - 1 scale. This is by design: I tried more complicated formulas that were always between 0 and 1, but they were messy (especially when converted into Excel) and ended up undervaluing certain forms of participation.

0.7*RankedClassParticipation + 0.25*ClassAttendance + 0.5*(Emails + OfficeHours + Surveys)/(0.75 * [3 + SurveysGiven])

This formula is much cleaner and offers multiple ways to get a high grade. One simple way is to attend all classes and ask/answer questions each class. Assuming you don't send emails, go to office hours, or complete any surveys, then the ParticipationScore would be 95%. Another way would be:

participate in 25% of classes (about 1/month)
attend all classes (or have excused absenses)
send one email
attend no office hours
complete the only survey offered that month

This would give a participation grade of 93%.

Since I want 50% of the participation grade to come from the quizzes and 50% to come from everything else, if a student receives over 100% for ParticipationScore, the score is adjusted to be 100%. This is why the MIN function is applied to ParticipationScore.