Teaching as Testing

Folks are always surprised that I let my students take most of their tests online.

Colleagues ask me . . .

• Aren't you afraid that your students will simply look up the answers without studying?
  
  
  
  • Afraid? NO! I'm hopeful that they will look up the answers they don't know. By doing so, they are studying!
  
  
  • Students quickly find out that it is far more efficient for them to study first, then take their online test. Before long, they are only "looking up" the really tough ones that are helping them discover what they forgot to study . . . or didn't really learn well when they studied.
  
  
  • Students also find that many of the items on the test are not simple regurgitation of facts that can simply be "looked up." Perhaps they do need to look up some facts, but also have to process the information before the items can be answered successfully.
  
  


• Aren't you afraid that students will get help from other students? Or even other experts in the field?
  
  
  
  
  • Afraid? There's that word again. Hmmm, are my colleagues really that fearful themselves?
  
  
  • No! If they are really stuck, I want them to collaborate to find the answer. We all know that collaborative learning is highly efficient, right?
  
  
  • Experts? Do they really know that many experts who are willing to help them answer every item on every test? Even if they do, isn't that also collaborative learning? Isn't this how I want them to solve problems in the "real world" in their careers? Aren't professionals expected to consult with experts to solve a problem? What a great skill they are learning!
  
  


• Aren't you afraid that your students will print and save their tests and pass them along to other students?
  
  
  
  • Oh, I'm so so afraid. Not. I want them to print out their tests! My tests come up one question at a time and when all the questions are finished, each test is automatically graded (correct or incorrect; the correct answer is not given to them), and may be printed out. They are then advised to find the correct answer to the questions they missed.
  
  
  • The reason they want to find the correct answers, besides their love of learning, is that they get three attempts at each test (the highest score of the three "counts").
  
  
  • In correcting their answers, they again have the opportunity to collaborate with their peers in the course, outside experts, or . . . heaven forbid . . . me!
  
  
  • I use huge test banks of items, of which only a few are randomly selected to appear on each attempt of a test. It would be a major effort to reconstruct the whole test bank . . . so much time and effort that it's far easier to simply learn the concepts. Besides that for most items, the elements (choices, matches, examples) also appear randomly within each item, so memorizing items wouldn't help them as much as you'd think. If they did find and memorize all the items? I think that in the process, they'd be actively studying the concepts . . . right? Why would I be afraid of that?
  
  
  • I do delete, alter, and add to the test banks on a regular basis so even passing an accumulated set on from year to year wouldn't help them much, either. (This sounds harder to do than it is.)
  
  


• Three attempts? Won't they just "wing it" the first time, then use the first attempt print-out to ace the second attempt?
  
  
  
  • Yes and no.
  
  
  • Some of my students do strategically wing it for the first attempt in order to assess the type and quality of test items I'm shooting at them. Then they use that knowledge to help them figure out how I'm approaching the course objectives and what kind and depth of understanding they need to be successful. Good for them!
  
  
  • As I stated earlier, each attempt is very different from any other attempt that they or their peers will be taking. I'll explain more about how randomization of test items and choices works (and how to do it yourself) in a later post to this blog. But with relatively few alternate forms of items in the test banks, I can produce literally millions of different tests. So they cannot really simply memorize answers. All they can do, alas, is learn the concepts if they want to prepare for their next attempt.
  
  
  • Remember, I want them to use previous attempts of a test to prepare themselves for the next try. That's called studying.
  
  
  • By providing three attempts and accepting their highest score, I can see what their best effort really is.
  
  


• So the students cannot possibly cheat?
  
  
  
  • Not easily! What many folks would call cheating, I call collaborating and learning from your past mistakes. Like in the real world, eh?
  
  
  • There is one type of cheating I do recognize: they cannot have another person take their test for them.
    
    
    
    • If I have 9 online tests, each with up to 50 (really tough) items, and each with up to three attempts possible (a total of up to 1350 difficult test items) . . . who in the world are they going to get to do that for them and do it successfully?!
    
    
    • My course management system does have some information (time-stamping, etc.) that can give me hints about possible cheating.
    
    
    • Their peers (and my colleagues who supervise the on-campus computers) are watchful for this behavior.
  
  
  • We talk together about academic integrity and about how we all need to prevent academic dishonesty in our course (this helps more that you might think). We even run through case studies regarding academic integrity as a small-group discussion exercise.
  
  
  • According to my new favorite book about teaching (and I have few favorites in that genre), What the Best College Teachers Do
    , the best teachers don't worry so much about cheating. Instead, they work hard to foster a trusting relationship with students . . . one that students will be unlikely to violate trust by cheating. And recognize that one cannot completely erase dishonesty from any group of humans, no matter how well you foster trust.
  
  

They do get some in-class exams, as well. So there are points at which I do a summative evaluation of their learning in a more traditional manner. They do much better at those now than they ever did before I started using online testing. The proof is in the pudding, as they say!

OK, this is just to the the discussion going! Comments, anyone?

(Remember, I'll have more details on how I do some of this in later blog postings.)

Want to know more about online assessment of learning? Check this out:

Assessing the Online Learner:
Resources and Strategies for Faculty
(Online Teaching and Learning Series (OTL))