Easy task high credit, tough task low credit, reasonable results.
Nov 18th, 2009 by Nick
If you are assigning scores to questions on a test, I used to make hard questions worth more points than easy questions, because I thought that they should carry a greater reward for correct answers. This caused problems because many students who had a decent level of understanding of material were not getting the questions worth the most points on their assessments.
My current approach is the exact opposite, an easy question should be worth lots of points, and a tough questions should be worth a few. The few points are enough to separate out kids who are good from those who are excellent, but students who do well on a test should not be receiving terrible scores. This tends to agree with our intuition about what a student’s total score should be.
I think I figured this out sometime last year, and I’ve been thankful for it ever since.
Huh, this has never occurred to me before, but I like it! I think I’ll adopt the philosophy.
I keep the point values relatively flat, sometimes the harder stuff is worth a bit more, but the amount of easy stuff is always greater. I often start a test with “gimme” level problems for the first 30 – 50 points, then moderately easy, then one or two moderately hard (to help separate the top from the almost top…)
Slightly different take, but based on a similar line of thought.
[...] Subscribe « Easy task high credit, tough task low credit, reasonable results. [...]
Great point. It is amazing how something can seem like common sense, but nobody does it. I will definitely apply this idea.
I can imagine this scoring scheme creating a problem for students who attempt hard problems and then run out of time, but whether that’s likely to happen depends on how long the test is and how it’s presented. I definitely like the idea of reserving most of the point spread for distinguishing the middle of the class.
Ben – student problems concerning timing for the test is a separate issue from how you actually tally up the results and report a score in my mind. There’s an easy solution to the problem: remind students that it’s important to answer questions they are confident on before those they are “reaching for” or simply give more time.
I am frustrated typing that because I’d much rather be using Standards Based Grading – shouldn’t the central issue be assessing at what level a student understands and can perform? Anyway, if you are still (stuck) applying old-school tests it’s reasonable to me that the penalty for missing an easy question should be greater than the penalty for missing a tough one.