coleshaw
commented
8 years ago I don’t quite recall us talking about unique LOs…I thought we agreed that the wrong answer LOs cannot match the correct answer LO.
From an authoring standpoint, you should currently be able to assign the same LO to multiple wrong answers. But my question then becomes — should p1 + p2 be a separate outcome, or are we actually talking about assigning multiple LOs to a single wrong answer? I feel like the issue is the latter, not just linking multiple answers to the same LO.
If we think generally about multiple wrong answers having the same LO, I think the issue actually moves to the sequencing / student app — what does the app do when they pick one? The logic could be the same (pull from the pool of LOs) and is ignorant of the fact that other wrong answer choices also point to the same LO.
If there are multiple LOs assigned, I can imagine a scenario like you suggested, perhaps first a p1 then a p2 question get inserted. But that is in the logic, so what do we need in the authoring tool to support these things?
El jun 1, 2016, a las 12:54 PM, Chad Lieberman notifications@github.com escribió:
I think we had a brief discussion about this before and I said something like "it would be good to make sure the LOs corresponding to wrong answer choices are unique because that is the best way to differentiate based on misunderstanding", i.e., it is good question design.
I think we need to discuss this issue further with the faculty members probably because there are questions they use for which this doesn't work. In fact, writing a question with a different distractor would give away some information about the correct answer. To demonstrate the issue, consider the following question:
What is the equation of the line shown below, written in slope-intercept form? https://cloud.githubusercontent.com/assets/1149890/15717935/1a8a18dc-27f7-11e6-9af7-d7d192a1dee2.png The correct answer is y=6x+4, and the faculty have selected the distractors y=6x-4, y=-6x+4, and y=-6x-4. (That is, the answer choices are all of the combinations of m={-6,6} and b={-4,4}.)
Basically, the question relies upon two prerequisite outcomes: (p1) determine the slope of a line from its graph and (p2) determine the y-intercept of a line from its graph. The distractors correspond to issues with (p1), (p2), or (p1) & (p2) together.
In the context of remediation, those selecting the distractor corresponding to (p1) & (p2) could get scaffolding for just p1 (e.g.) and then try the target question again (at which point remediation of p2 might be performed). In order for this to happen, I suppose we need to be able to assign the outcome p1 to multiple wrong answer choices. =/
Unfortunately, designing the distractor differently necessarily leaks information about the solution to the problem for those who are trying to game the system a bit --- the option could be eliminated because it doesn't share the common features of the other choices.
I'm open to discussion about this. We should pose the question to the faculty as well probably.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/wombats-writing-code/fbw-components/issues/11, or mute the thread https://github.com/notifications/unsubscribe/AEs6UTK9ImttwNR36Pe1OSJN-gFbE50zks5qHbk1gaJpZM4Irwh4.
chadlieberman
I think we had a brief discussion about this before and I said something like "it would be good to make sure the LOs corresponding to wrong answer choices are unique because that is the best way to differentiate based on misunderstanding", i.e., it is good question design.
I think we need to discuss this issue further with the faculty members probably because there are questions they use for which this doesn't work. In fact, writing a question with a different distractor would give away some information about the correct answer. To demonstrate the issue, consider the following question:
What is the equation of the line shown below, written in slope-intercept form?
The correct answer is y=6x+4, and the faculty have selected the distractors y=6x-4, y=-6x+4, and y=-6x-4. (That is, the answer choices are all of the combinations of m={-6,6} and b={-4,4}.)
Basically, the question relies upon two prerequisite outcomes: (p1) determine the slope of a line from its graph and (p2) determine the y-intercept of a line from its graph. The distractors correspond to issues with (p1), (p2), or (p1) & (p2) together.
In the context of remediation, those selecting the distractor corresponding to (p1) & (p2) could get scaffolding for just p1 (e.g.) and then try the target question again (at which point remediation of p2 might be performed). In order for this to happen, I suppose we need to be able to assign the outcome p1 to multiple wrong answer choices. =/
Unfortunately, designing the distractor differently necessarily leaks information about the solution to the problem for those who are trying to game the system a bit --- the option could be eliminated because it doesn't share the common features of the other choices.
I'm open to discussion about this. We should pose the question to the faculty as well probably.