So the last time I posted I discussed my testing and how I was trying to determine how long a test should take.
So here's the update: the elementary algebra test was fine. I was worried for nothing. The college algebra test was a little long, but that only affect 2 people and I let them come in and finish it later.
However, I did get a surprise. The third test was intermediate algebra, and I had not even thought it would be a problem. It took me 12 minutes. It ended up taking the students 2 hours. Yes, 2 hours. At the end of the first hour no one had finished, so I told them I'd let them finish on Monday. For Monday, they again took the whole hour. That was a disaster, and I was surprised. How did I go so wrong?
The intermediate algebra test was exclusively graphing. There were a couple interpretation problems, but I think the main thing was that they were just not as quick as I am on graphing. In fact, instead of the multiplier of 3, I should have had a multiplier of 5.
To remedy the situation I curved the test using a square-root curve. I fully admitted it was my fault in making it too long. The number of points on the test should have tipped me off, as previous tests were in the 60-90 point range, and this was 106. But I don't know if that is significant or not.
So for my next test that includes graphing, the one for elementary algebra, I have used a multiplier of 5. We'll see tomorrow if this is correct.
Musings on Teaching Mathematics
Thursday, October 17, 2013
Saturday, October 5, 2013
Assessing Test Effectiveness
When I write tests, I generally go through the review at the end of the chapter, and choose questions that embody what I think my students need to know. I then take the test myself to see how long it takes, and to see how many points to give each problem. In general this has worked well: the tests are timed appropriately, and no one has complained that they are unfair.
I have 3 tests scheduled for tomorrow. Let's consider two of them:
Test 1 is elementary algebra. I had a student come in to take it today because she will be gone for sports tomorrow when it will be given. She worked on it for an hour, and still had not completed 5 questions. Is this an indicator that the test is too long or too hard?
At this point it's too early to say. She may not have been prepared. I multiplied how long it took me by 5, which is quite generous. This has worked well in the past. I'm a bit worried, but I think in this situation she is just not prepared.
Test 2 is college algebra. I gave my students a study guide today which is the exact same length, just different numbers. The book has an issue with chapters being too long and not coherent, so I felt a study guide is the best way to compensate for this. Judging by how long it was taking the students, I am worried that there are too many questions. I took the test myself and again multiplied by 5. There are some multiple choice which should go quick. I think this is why there are many questions. If each question takes less than a minute, you can have more of them.
I do worry that it's too long, even with my multiplier. I think this is because of the lack of flow in the chapter, and because of the many questions. Tomorrow we shall see, but I need to prepare myself with a resolution if it does end up being an unfair test. Do I curve? Do I let them work on it more in class on Monday?
What are your thoughts?
MyMathLab: Extra Credit
I use MyMathLab for my algebra classes, and I wanted a way to give them extra credit. Ideally I wanted to do it via the practice tests: if they do the, give them extra credit for the actual test. However, there is not automated way to do this. It can be done manually though.
How to Add Extra Credit to a Single AssignmentThe downside to this option is that anything not worked on in the study plan is considered 0, so students who work on only a chapter may have only 4% of the study plan done, and then get 2% of that 4%, or 0.08% added to their grade.
Subscribe to:
Posts (Atom)