Description:
Question: Question text
A. Text {choose True or False}
B. Text {choose True or False}
C. Text {choose True or False}
D. Text {choose True or False}
Grading method. If a student choose 1 correct answer, they will get 0.1 point. If 2 correct answser: 0.25, if three 0.5 and if all correct will get 1.0 point.
THank you very much.
Hello Rick,
Your solution is interesting. However, it will require a different quiz for each set of four true or false questions, in addition to the calculation items in the gradebook. If Nguyen has multiple sets of four questions, it will result in a lot of quizzes and calculations in his gradebook, to the extent that I wonder if this solution is practical.
But then again, your solution may be better than mine.
You just collided with the elephant in the room! This kind of T/F in multiples is a big thing in certain fields, in medicine for example. The "inventors" must have created clever questions of that type, no doubt. But today it is an alibi cover simple T/F questions because single such questions are too cheap. Bundling them don't increase the quality, they just look impressive. The argument today is that they are related. Being from the same chapter doesn't mean that the questions are related. For example, what is the difference between asking these questions separately or in bundle:
1. Mars is the closest planet to the Sun T/F?
2. A day in Mars is almost three times longer than a day in Mercury. T/F
3. Mercury, Mars and Venus align every 400 million years. T/F
4. Out of the planets other than the Earth, Mars has the highest chance of life. T/F
Visvanath,
I think there's more to it. For instance, considering they are true/false questions, there's an equal chance on average for correct answers to be true as false. Students who haven't studied at all and answer randomly can score 50%, even though they should maybe score 0%. For example, I could easily score 50% by guessing in a quiz with true/false questions about microbiology even if I don't know the subject.
There are various ways to address this issue, such as using negative points for incorrect answers or employing a non-linear marking scheme as mentioned by Nguyen. Such non-linear marking is common, and you can find more details by referring to the literature on the subject.
In any case, the point is that this type of marking is required by the Ministry of Education of Nguyen. Telling him that it's not a good type of marking isn't very helpful. It looks like it trivializes Nguyen's request, which of course, we are not doing here.
Yes, as you say, the point is what the OP needs. I thought you solved it here and here. The only remaining thing is he has to implement one on his own. I think, repeating it in different versions, in different voices will only confuse him. The said Ministry has sent him in enough missions in a short time - the majority still unresolved.
Now that is out of the way, I can't resist talking the pedagogy behind these questions. Will reply to Rick soon in a sub-thread.
P.S. I've noticed that you found Nguyen's mission today: Import File containing Maths into question banks. Let's hope that this yesterday's mission won't stay in yesterday.
And, there is also https://github.com/moodleou/moodle-qtype_oumatrix, which we created recently, but at the moment that is only on github, not in the plugins database.
Tim,
At first glance, adding a bunch of sub-questions might seem like a good idea. However, it doesn't address Nguyen's and his Ministry of Education's main request in any way, which is to have non-linear marking. It also doesn't provide the option for a quick and effective workaround while waiting for a full-fledged plugin to be developed.
I looked at Visvanath's reply to my post above, and I can't answer his question, "What is the difference between asking these questions separately or in bundle" about Mars and planets? Furthermore, questions can be already be grouped in Moodle quizzes, as Visvanath shows. Also, I can't see why getting two questions correct would be worth .25 instead of .5, but this is not to say that someone, or a government ministry, is wrong for wanting a non-linear grading scale. But I will need to think more about it and perhaps develop a better understanding. In my years of using LMSs, I have seen "colleagues," and "bosses," "departments," "schools," and others say, "It must do this or that." I understand the sensitivity associated with questioning people.
You wrote:
> "What is the difference between asking these questions separately or in bundle" about Mars and planets?
I repeat them from this post, for everybody's convenience:
>> 1. Mars is the closest planet to the Sun T/F?
>> 2. A day in Mars is almost three times longer than a day in Mercury. T/F
>> 3. Mercury, Mars and Venus align every 400 million years. T/F
>> 4. Out of the planets other than the Earth, Mars has the highest chance of life. T/F
The answer is, they are unrelated. Why should the student get less marks for Q2 if he answered Q1 wrongly?
Would you happen to have an example of four interrelated questions that are suitable for nonlinear, or non-proportional, grading? It might be challenging to find such examples with just true/false questions.
Dominque, I think I can provide an example addressing your question. This discussion takes me back to a particular scenario that I ran into involving "statistics." I desired to provide a dataset and then ask a half-dozen questions about it, such as the mean, mode, median, standard deviation, 90% confidence interval, etc. In my opinion, this scenario represents a very powerful need for "question grouping." Visvanath's example of "Mars" and your example of "Vietnam" do not seem to have a strong need for grouping (which Visvanath verified). In my example, I can see someone saying it is much more difficult to calculate the 90% confidence interval than to calculate the mean, so that question about the confidence interval should have more "points." This emphasis on "more points" is based on the difficulty of the question and not on the fact that the student happened to get 4 of 12 questions correct.
My example is from "statistics," but I don't doubt that other fields, like medicine, might have their own natural groupings of knowledge.
Returning to the four TF questions, what puzzles me is that a TF question perhaps represents the most "binary" kind of question, true or false. It perhaps provides the least amount of "resolution" about knowledge. Yet, the Ministry desires to provide a more intricate resolution to grading (i.e., a non-linear grading scale) than they desire for the knowledge itself. (I hope that this makes sense to someone.)
Now looking back at the Mars & Co. set of questions, I see that they don't even have to be MTF. It could be a multi-choice MC or even a single-choice Q. For example:
Which one (are) correct?
a) The statement is correct. The explanation is factually correct and is also the correct explanation.
We, Dominique and myself, are still looking for a real example where non-linear marking has a logic. Could you post one example?
In my example at least, if the answers need to be a bunch of T/F questions, they need to be interdependent, since combinations like "The statement is wrong. The explanation is factually correct and supports the statement." don't make sense.
You can also tweak the Cloze question a bit. At first glance, the tweak may seem complicated, but it is actually very simple. I will also eventually improve it.
You can try the following example at https://dynamiccourseware.org/course/view.php?id=89§ion=27➚.
Hello Nguyen,
I have greatly improved the question, and the script it contains is now much simpler.
The question is of Cloze type. It consists of four sub-questions, the only special characteristic of which is that they each count for zero points. The question also includes a hidden sub-question that calculates the grade of the question based on the given scoring scale. This sub-question is somewhat difficult to develop, so I am providing you with an Excel file to generate it automatically.
The question is secure, meaning that correct answers are not identifiable even by inspecting the DOM.
Building a new plugin is a daunting task, and I will not have time to do it soon. However, in the meantime, the solution I propose is already available and should work very satisfactorily.
For now, you can only use one question per quiz page, but I can arrange that after the spring break. Also, the order of sub-questions is fixed, but it may be possible to make it random.
You can try the following example at https://dynamiccourseware.org/course/view.php?id=89§ion=16➚.
Note that you can have as many such questions in a single quiz (with the only restriction, for the moment, of having each question on a separate page).
I just read the discussion on that topic.
I just want to add a rather simple solution using FORMULAS qt (what else).
FORMULAS has to my knowledge only one real global object and that is an array, which can be read and rewritten in all parts of the question, especially in the grading section.
Using the global array kor I count the number of right answers (in the grading variables part).
Depending on that number I assign the marks to kor[0] (using pick) and that is the grading criterion!
See attached question!
Cheers
Joachim
Thanks a lot for joining the conversation and providing a Formulas solution, which of course is the most straightforward one.
I don't remember why I left the Formulas solution aside, but since Nguyen mentioned using a plugin, I think the Formulas plugin should suit him.
I'm out of the office until Thursday and I'm supposed to talk to Nguyen when I return but, if he contacts you in the meantime, please give him all the necessary explanations.