carpentries / instructor-training

Instructor Training
https://carpentries.github.io/instructor-training/
Other
175 stars 289 forks source link

Different exercise for Faded Examples #68

Closed apawlik closed 8 years ago

apawlik commented 8 years ago

Coming up with a faded example during the session is extremely difficult for the trainee instructors. I suggest a different exercises: give them 2 examples of "broken" faded examples to fix. That is, give them 2 examples with description what problem they are teaching student to solve and ask the trainee instructors to fix/improve these examples. The examples should be "broken" in a way that the faded/taken out bits will be completely random, the teaching would not be effective. This would also allow the trainee instructors to better grasp the idea of what intrinsic, germane and extraneous load. I need some time to think about this but will try to submit something soon.

apawlik commented 8 years ago

Another/additional idea - ask the students to identify which could be intrinsic, germane and extraneous bit in the faded examples presented. For example, if it's a coding faded example (currently in the materials):

ChristinaLK commented 8 years ago

My two cents: I didn't have students create a faded example - just walked them through it.

I like the idea of having participants identify types of knowledge in either the faded example, or some other teaching/lesson example.

apawlik commented 8 years ago

Tried out this at Melbourne IT January 2016

This worked pretty well I think, they seemed to be very engaged and liked it. Though there were lots of questions.

gvwilson commented 8 years ago

@apawlik can you please PR this?

apawlik commented 8 years ago

An attempt https://github.com/swcarpentry/instructor-training/pull/153

One thing we need to note in the materials or Trainer's Guide is that Faded examples work well in the type of knowledge which is algorithm based. I think that faded examples used for example in history may actually only check factual knowledge, rather than teach problem-solving.

In general, I think we need someone who understands congnitive load stuff well to improve the practical side of this lesson At most Instructor Training sessions I ran, people were pretty confused with this and struggled to relate this to teaching (SWC/DC workshops and teaching in general).

lexnederbragt commented 8 years ago

I wasn't aware of faded examples until I saw the disussion of them in the instructor training material a couple of weeks ago. The lessons I know well (shell, git unix) don't (or didn't) have faded examples. Is this kind of exercise 'the new thing' now, and should we (or have we already) add(ed) them to the core lessons? Pardon my ignorance, I may have missed a discussion on this...

gvwilson commented 8 years ago

We don't have any in the current lessons, but we'd like to encourage people to start submitting them.

apawlik commented 8 years ago

Yes, definitely it would be great to have a selection of faded examples in our materials. Also because we could refer back to the SWC/DC materials showing that we really use what we preach for. More examples would also be very helpful for the Trainees to understand what faded examples are (teaching strategy to solving problems) and provide an opportunity to think about different kinds of cognitive load that occur.

Cognitive load theory is one of the most challenging bits of the Instructor Training at the moment and I feel it is because we don't have good enough exercises. More examples in our lessons could help with developing exercises for the IT.

lexnederbragt commented 8 years ago

Would an exercise like 'list all the things a learner needs to know in order to write a for loop in bash (or python)' help?. Followed by part 2: 'what would be added to this list if the exercise were to be done in a Jupyter Notebook?'. Followed by the explanation that, for example, this is one of the reasons we don't use the notebook for teaching the unix shell?

gvwilson commented 8 years ago

An example from today's instructor training is a few lines of R to draw a simple plot, and then each fade after that (a) changes something about the plot being drawn (e.g., show max instead of ave, or scale, etc.) and (b) has more blanks.