Note: You are currently viewing documentation for Moodle 3.11. Up-to-date documentation for the latest stable version of Moodle may be available here: Effective quiz practices.

Effective quiz practices: Difference between revisions

From MoodleDocs
(Added see also * [https://www.unmc.edu/facdev/_documents/ConstructingWrittenTestQuestions_WritingManual.pdf)
(Added section on ==Proctored exams==)
Line 138: Line 138:


See [[Using certainty-based marking]].
See [[Using certainty-based marking]].
==Proctored exams==
* As a start, check the info in this discussion: https://moodle.org/mod/forum/discuss.php?d=399255#p1610674. Use advance search for the forums, as there has been a bunch of discussion in 2020 about this subject.
* One of the better solutions seen recently for lockdown testing, was a combination of things:
**Use of the safe browser and have the proctor call the student via some form of video call.
**Have the student show the proctor their computer setup and the room the test will be taken in.
**Then, have the student set up the phone in a position showing the student and their computer screen, as best they can from across the room.
**This allows the proctor to watch and hopefully, see any attempts to cheat.


==See also==
==See also==

Revision as of 14:19, 11 May 2020

As we’ve seen, Moodle quiz engine is a powerful, flexible tool for monitoring and diagnosing student performance with certain types of knowledge. Using this tool effectively can boost your course’s effectiveness, and promote student performance. While a computer-scored quiz is a different performance than more open-ended assessments, it does give a valuable window onto student thinking, especially when you use good strategies, and a little creativity.

Quiz strategies

Of course, using the quiz engine effectively takes some work and practice. The first thing to do is to use effective question design strategies. If you ask good questions, you’ll get useful data about your students’ performance and understanding of the material. Of course, the converse is also true. There is a ton of literature about effective assessment design available. I’ll just highlight a few of the most important ideas.

  • Tie each question to a course goal. After all, you want to know whether your students are achieving the goals of the course, so why not ask them directly?
  • Try to ask multiple questions about each important idea in the class. This gives you more data points about student understanding.
  • When writing a multiple-choice question, be sure each wrong answer represents a common mis-conception. This will help you diagnose student thinking and eliminate easy guessing.
  • Write questions requiring your students to think at different levels. Include some recall questions, some comprehension questions and some application and analysis questions. You can determine where students are having problems in their thinking. Can they recall the material, but not apply it?
  • Test your questions. After you’ve established an initial question bank, use the system reports to determine which questions are useful, and which aren’t. As you write new questions, give them a lower point value and throw in a few to establish their reliability.

Once you’ve got a few well-written test banks, be sure to use the quiz reports and statistics to monitor your classes performance. The detailed reports and statistics available to you are valuable tools for understanding student understanding of the material.

Creative quiz uses

With the Moodle quiz engine, it’s easier to utilize educationally sound assessment strategies which would be too difficult to implement with paper and pencil. Most people think of tests as an infrequent, high-stakes activity, like mid-terms and finals. Better strategies involve frequent, low-stakes assessments you and your students can use to guide their performance during the course of the semester.

Creating a series of small mini-tests gives you a very flexible system for gauging performance and keeping students engaged in the class. Here are a few ideas for quick quizzes you can use as part of a larger assessment strategy.

Chapter checks

Getting students to complete reading assignments has to be one of the hardest motivational tasks in education. Reading is critical to understanding most material, and fundamental to success in many classes. The problem for most students is there is no immediate reward or punishment for procrastinating on a reading assignment. If you haven’t done the reading for a class discussion, you can either keep quiet, or, as I used to do occasionally, wing it by skimming in class. If you have a lecture course, there’s almost no need to do the reading as the lecturer usually covers most of the material in class anyway.

Creating a little mini-test for each reading assignment solves a number of problems. First, it encourages students to do the reading so they can do well on the quiz. Second, it gives the students feedback on how well they understood the reading assignment. Third, it gives you data about what aspects of the reading students found confusing, and which they have already mastered so you can focus your class activities.

For a reading mini-test, I would recommend setting a limited time quiz students can only take once. Because it’s a low-stakes activity you want students to use for self-assessment, I would also display feedback and correct answers. If you’re concerned about students sharing answers after they’ve taken the quiz, randomize the question and answer order. If you have a test bank, make some of the questions random as well. As an additional assignment, students should write down one question about a question they got wrong, and bring it to class.

Test practice

The key to effective practice is to have a realistic practice environment. Many students worry about tests, especially high-stakes tests, because they have no idea what to expect. What question format will you use? How detailed will the questions be? What should they study?

You can help alleviate test anxiety by creating a practice test students can take to help answer these questions. These tests are usually based on old questions similar to the current test questions. Use last year's final as an example test, which will force you into the practice of writing new questions every year. This is a good idea anyway, as you can be sure someone has a copy of last year’s test they are sharing with others.

To set up a practice test, I’d create a zero point test with questions from the year before in random order with random answers. I would also allow students to take the test as many times as they’d like so they can test themselves as much as they need. Display feedback, but not correct answers so it presents more of a challenge.

Gamification may be used for quiz practice. See the Quizventure activity.

Data gathering

As an expert, you know a lot about your field. Your challenge as a teacher is to translate your knowledge for a novice who doesn’t share your conceptual structure or experience. An example or lecture you think is brilliant may leave your students completely confused. It can be hard to tell what students really understand and what’s leaving them baffled.

A data-gathering quiz is similar to a chapter check, but it takes place after a class meeting or lecture. Your goal is to quickly get some feedback on student understanding of a lecture. What did they really understand? What do you need to spend more time on? I’ve found many instructors have trouble gauging what students find difficult, and what the students find so easy they are bored.

Setting up a post-class data-gathering quiz is similar to creating a chapter check. Set the quiz for a limited time, like a day or two before the next meeting. Allow them to take it once and display feedback and correct answers.

Quiz security and cheating

Of course, online testing also presents another chance for the cheaters in your classes to try to game the system. Most online quizzes are meant to be taken at home, or at least outside of class. Students can download the questions and print them out. They can take the tests with other students, or while reading their textbooks.

Fortunately, you can counter many of these strategies, making them more trouble than they are worth to the students. Let’s look at a few strategies for countering most cheating schemes

Printing and sharing questions

If you display feedback and correct answers, students can print the results page and share it with their friends. Or they can simply print the questions themselves directly from the quiz. The key to discouraging this behavior is to randomize the question order and the answer order. It makes the printouts a lot less useful. Creating larger question banks and giving tests with random subsets is also an effective strategy. If students can only print a small number of questions at a time, they will need to view the test again and again, then sort the questions to eliminate duplicates.

Warning: Assume there will be printed copies of your questions available to students who want them. Most instructors don’t realize students frequently have copies of old paper based tests, and electronic test delivery is another way for students to get copies of the questions. I know one professor who had over 1100 questions in his online test bank. At the end of the semester, he confiscated a printout from a student. It had every question with the correct answer, neatly formatted and divided by textbook chapter. We decided if students wanted to memorize 1100 questions and answers to the level where they could answer a small number of them displayed at random, then they would have learned more than if they had just studied. Of course, we used timed quizzes and other strategies to minimize using the print-out as a reference manual.

Using the textbook

Students will frequently look up the answer to questions in the textbook or a reading. If you are giving a chapter check quiz, then this is what you want them to do. Otherwise, you need to come up with creative ways of making the textbook less directly useful. Timed quizzes are the single most effective tool for eliminating this strategy. A timed quiz requires the students answer the questions in a certain amount of time. If you give enough questions and make the time short enough, they won’t have time to look up all the answers. I usually give about 30 seconds per multiple-choice question. If they answer them faster and have time to look up some answers afterward, I figure they knew enough to deserve to look up an answer or two.

Asking students to apply their knowledge to novel situation can also make a difference. Synthesis and application questions can’t be looked up. Students have to understand the material and apply it creatively to answer the questions. So while they may take the time to review the text, they will still need to try to understand what they’ve read to successfully answer the question.

Working with friends

If your students are on the same campus, they may get together in a lab and try to take the quiz together. This is an easy strategy to thwart with random question order, random answer order and questions randomly pulled from a test bank. If my screen doesn’t look like yours, then it’s harder for us to quickly answer all of the questions. A timed quiz also makes it harder for the two of us to cheat if we have different questions and we only have a short amount of time to answer.

Have someone else take the test

The old adage goes “On the Internet, no one knows you’re a dog”, and no one knows who is actually taking the test. Students will sometimes pay classmates, or others who have taken the course in the past, to take online quizzes for them. There are two ways to counter this strategy. One, have an occasional proctored exam where students need to show ID. If they haven’t taken the quizzes or done the work until then, they will do poorly on the proctored exam. To eliminate current classmates from taking each others quizzes, only make them available for a short time. You could require everyone take the test within a 2- or 4-hour block. If the test is properly randomized, it will be very difficult to take it more than once during the testing period. The test taker will worry about their own grade first, then about their employer's grade.

Obviously, there are many strategies students can use to cheat. While it would be naïve to assume there isn’t cheating, the vast majority of your students want to succeed on their own merits. The anonymity of the online environment may open up new avenues for the cheaters, but it’s not really much different from your face-to-face classes. A few people will go to great lengths to cheat, but most will be honest as long as it’s not too easy to get away with it. A few precautions will eliminate most of the easy cheats, and the classic strategies will work for the others.

Robust testing with random variants

This section describes a good way to help minimise the potential for cheating, and increase the opportunity for students to learn from the feedback by repeated attempts at the quiz. The basic idea is to take each particular question that you were thinking of, and make several slight variants of it. Then use Moodle's random question feature, so that each student gets one of the variants picked at random.

An example

A good example of this (although not in Moodle) can be seen at https://students.open.ac.uk/openmark/mu120.m5omdemo/. Take that test once, making a rough note of the questions you are asked. Then after you have done 'End test', do 'Restart entire test' and see that you are asked a different set of questions that have different answers, although they test the same knowledge. This sort of strategy is easier to implement in some subjects than in others.

How to set this up in Moodle

Suppose we are going to create a quiz with 6 questions about interpreting diagrams (that is, we are going to try to clone the OpenMark example above). For the fourth question, the closest we will be able to get would be the Image target question type from the Modules and plugins database.

1. Create a category for each 'question' in the quiz

As you can see from the screen shot, I have created six appropriately named categories, all neatly grouped inside a parent category. You do this on the 'Categories' tab of the question bank interface.

Variants categories.png

2. Create the first variant of the first question

Create the first variant of the first question, just like you would create any other Moodle question.

In our example, this might be a Embedded answers (Cloze) question type. The question text might be:

Below is a plan of a proposed garden. The scale is that each division in the plan represents a length in the garden of 0.5 metres. What is the proposed length and width of the Patio in the garden?
Variants flowerbed.gif
The Patio is {CLOZE syntax} metres by {CLOZE syntax} metres.

3. Create the other variants of the first question

To easily create a variant, click the duplicate icon next to the first question, then make the changes you need to turn it into the second variant and save it. Repeat this process to create as many variants as you want.

In our example, we might change the word Patio, and the scale factor each division represents 0.5 metres. We would also need to change the answers and the associated feedback in the {CLOZE syntax} bits.

4. Repeat 2. and 3. for the other questions

The screen shots show the variants of the third question. This one is a bit more of a pain to set up, because each variant will use a different image of a pie chart, so there is a bit more editing to do, and more files to upload to the course files area.

Variants questionsincat.png

5. Add the questions to the quiz

Once you have created all the questions, add them to the quiz using the 'Add random question' feature. Select the first category (Reading a plan variants). Ensure 'Display questions from sub-categories too' is off. Use the controls at the bottom to Add 1 random question to the quiz.

Repeat for each of the other categories in order.

Variants quiz.png

Creating variant questions in Atto

In 2016, a new additional plugin, the Cloze editor for Atto has a 'Duplicate' button, that allows you to easily and quickly create many question variants.

Comments

Obviously this is more work to set up (although not three times as much work as creating one quiz). It is up to you to do the cost benefit analysis for your particular quiz. Note that once you have set this up, you are more likely to be able to reuse quizzes in future, because you have reduced the potential for simple copying of answers.

As an alternative to 'Save as new question', you can use Moodle's import and export formats, and copy and paste in your text editor to create variants.

One issue you have to worry about is, are all the variants you have made of each question really equally difficult? Moodle 2.0 will feature a new Statistics report which should help you analyse your quiz results to see how difficult each variant is.

Experience shows that 'a few variants' can normally be taken to be 3 variants. This is enough to ensure that two students working at neighbouring computers will mostly get different questions to each other. More is better (providing you can ensure equal difficulty) but is more work, so you get diminishing returns.

(This section expands some of the advice above under Printing and sharing questions. It also describes how most online assessments at the Open University are constructed. The calculated question type is sometimes another way to implement quizzes like this.)

Certainty Based Marking

  • To make students think about how reliable their answer is.
  • To encourage students to try to understand the issues, not just react immediately to a question.
  • To challenge: if a student won't risk losing marks if wrong then they don't really know the answer.
  • If a student is a careful thinker but not very confident. they will gain in confidence.
  • It is more fair - a thoughtful and confident correct answer deserves more marks than a lucky hunch.
  • Students need to pay attention if they make confident wrong answers: think, reflect, learn!
  • Efficient study requires constantly questioning how our ideas arise and how reliable they are.

See Using certainty-based marking.

Proctored exams

  • One of the better solutions seen recently for lockdown testing, was a combination of things:
    • Use of the safe browser and have the proctor call the student via some form of video call.
    • Have the student show the proctor their computer setup and the room the test will be taken in.
    • Then, have the student set up the phone in a position showing the student and their computer screen, as best they can from across the room.
    • This allows the proctor to watch and hopefully, see any attempts to cheat.

See also

  • Safe exam browser The Safe Exam Browser can work with Moodle to control what a student can do when in Moodle. To use it, it must be enabled in Settings > Site administration > Development > Experimental > Experimental settings. This adds the choice 'Require Safe Exam Browser' to the 'Browser security' field on the quiz settings form.

Forum discussions:

External links: