Last year, a teacher called to complain that her students' grades were too low.  She felt that the gradebook was at fault.  I looked at the gradebook and reported back that it seemed fine, but she insisted, so I looked deeper. 

The students grades were accurate but unintended.  The culprit: Weighting.  The teacher weighted two categories, one 40% and one 60%.  The 40% category contained one 10 point assignment, and most student earned 5 points.  The other category contained hundreds of points. Hmmmmm...

The purpose of using points is to motivate and provide feedback.  This is based on the assumption that people understand the meaning of a point.  It's easy to see if you imagine that a point equals a dollar. A student earns $10 for getting everything right on something worth ten points, and $100 for 100 points. It's obvious that something worth 100 points is more valuable and worth a greater time investment than something worth ten points.  Right...well, not if those points are in given in a system of weighting.

In a weighted system, something worth 100 points isn't worth $100. It might be worth $10. Or it might be worth $150. It's impossible to tell.  The weighted value of an assigment is a function of 1) the value of the category 2) the number of points available and 3) the number of assignments given in the category.

For example, pretend that all homework and tests are worth 100 points. Homework is weighted at 25% and tests 75%. If there is one test and one homework assignment, and the student scores 70 on the test, and 90 on the homework, the weighted score is 75%.

Now, let's say there there was still one test and twenty homework assignments. The student still scores 70 on the test, and 90 on all twenty homework assignments. Does the weighted score go up or down? Neither. It stays the same.  Weighting multipies the percentage earned by the weight regardless of home many points are involved.  With weighting, a student can get a score that is better than his current average, and his average can go down. Try explaining that one.

Why does this matter? Points are both feedback and stimulus: They suggest to students how much time to spend on something.  The suggest relative worth.  And not just in a given teacher's class because students are comparing pulls on their time from ALL their classes.

A student's workday is a collection of classes and assignments. Imagine a student trying to make time-investment choices across multiple classes, some of which use weighting, and each with 100 point assignment due the next day.  How should they make that choice?

What about our teacher with the low scoring students?  She had given most students a 5 of 10, 50%, on one assignment, and then counted that result as 40% of their grade.  She didn't intend to do that, but that's what happened. 

 Here's a simple test -- Ask students to guess the impact on their aggregate score of earning 100% on assignment in a given category.  If they can't predict the outcome, what does that suggest?  At best, it makes grading seem arbitrary, and at worst  we get an email from a perplexed teacher.

What can staffs do?  One idea is to collaborate on scoring systems so that, at a minimum, the math all works similarly.   This way, at least if people weight, they all weight and the math is relative  Better would be to have the same weighted categories, so that math might approach being comprehensible.  Another path would be to contrast the merits of weighting with the downsides, and make a choice based on criteria of student understanding and clarity.  

 At School Loop, we get so much email from teachers confused about the impact of weighting that we sometimes joke that teachers should need to pass a test before they can use the weighting tools.  Maybe it's not as funny as we think.