Tuesday, April 13, 2010

Making Open Accessible Content - Some Observations

Starting in January, I've been building a simulation in Excel about the Edgeworth Box, something we torture students with in intermediate microeconomics. Along with the simulation I've made some short screen capture movies with voice over to describe what's going on, using Jing Pro to make the captures. I'm to the point of liking it more than the full fledged Camtasia. Both because we are very conscious of making video accessible on my Campus and because I'm quite enthusiastic about what YouTube has been doing with captions, I committed myself to caption each of these videos. The first video I made the transcript manually. (YouTube puts in the timings if you input the transcript.) That was a pain and probably explains why I was slow to make the next video. Ultimately I figured out how to rip the audio out of the video. I am using the free version of AoA Audio Extractor, which works quite well for my purposes. It produces a very clean MP3 file. I then run that through Dragon Naturally Speaking, which produces a transcript. This I have to edit. It is still a pain but less so. You can see the end of the tunnel when you start the task. And I've gotten better at the editing over time.

I'm now almost done. (The last video still needs to be captioned. I hope to finish that by the end of the day.) You can see the results for yourself and make your own assessment about this approach as a modality for delivering mini-lectures. Before getting to my lessons learned from the experience I do want to mention some of my other motivations for going through this exercise. There has been a lot of mention recently of the Khan Academy and the large volume of mini-lectures to be found on that site. The quantity is quite impressive. However, pedagogically I thought the approach flawed. If one goes so far with the technology to make all that content availabe, why not go a few steps further in producing the content in a way students can better digest it? My experience from when I used to teach with chalk and a blackboard is that a well done lecture was not nearly sufficient for a good number of students to process the content. So the Excel simulations are there for that reason. Whether they achieve that end, I don't know. I do think they help in making the content more penetrable.

Another reason relates to the current fascination with lecture capture. As with many technologies we try to employ for learning, my thinking about the technology has evolved over time. I am much less enthusiastic about lecture capture than I was a few years ago because it is an added expense with an unclear benefit. We seem to be doing it because we can rather than because we should. Lecture capture does economize on faculty time, no doubt. But the product is not so compelling and for an ongoing class probably will be used little expect for exam review or by non-native speakers of English. The mini lectures of which my creations are an example, by being short and to the point, with the captioning and with more activity on the screen, makes for better viewing, I believe. Also, with such mini-lectures in use in-class time can be purposed for activities better done face to face, collaboration and dialog, which is where we should be headed. In a blended learning approach, especially when the instructor teaches multiple sections of the same class or teaches the class on a repeated basis, the mini-lectures can economize on faculty time, but they certainly take substantial effort to produce the first time around the block. That is probably the reason they aren't spreading more on their own.

Also, I may be teaching an online version of intermediate microeconomics in the not too distant future. I've been thinking of a variety of activities I'd have for the students. For each of these my approach is to give it a try ahead of time and then based on the experiment make a determination of whether it will be useful for the students and if I can keep up doing the activity when it would have to scale to the entire syllabus, not just a particular topic area. I'm still not sure on that last one, though I fear it would be the boredom in editing transcripts rather than the lack of time that would be the real constraint in deploying these mini-lectures.

* * * * *

I suspect that most people who found themselves talking to the computer screen rather than to somebody else on the phone, would feel self-conscious and the discomfort from that might prevent them from continuing with the activity. I've done enough of this sort of thing as to not feel particularly disturbed by recording my own voice, even by listening to it being played back. Any recommendation I make in the rest of this posts assumes that the instructor has gotten past the self-conscious point. The only way I know to get past that point is to do it sufficiently.

It is an eye opener to transcribe your own speaking - particular the way the spoken descriptions depart from how you would describe matters in writing. One reason for this divergence is that with the screen movies you can point to places in the diagram you are talking about and the pointing allows you to be much more colloquial, while in writing there would be a tendency to be more formal, as the diagram is more of an impersonal object. There is another part that the language choice in giving descriptions really is idiosyncratic to me when I speak. A different economist explaining the same stuff would use different language to convey the ideas. In writing, there is a tendency for the language to homogenize. So on this each lecturer gives his own version of the course, but were we to write textbooks, those wouldn't be so different. There is a question how important the idiosyncrasy is to the student's learning and what role the idiosyncrasy plays. I am not sure but I suspect it helps the students to be able to inject some of the instructor's personality into the subject matter.

One of the things that happens when you edit the transcript is that you play short segments of the the audio, then stop, look to see if the printed text corresponds, but then also listen for whether you fully understood what was said in the audio. Often there would be a mumble, or a phrase that could have been different words and still give sensible meaning to the overall. If I, the deliverer of the mini-lecture, am unsure of particular phrases, what must the students feel? Would students listen to the audio to gain meaning from it in the same way, playing short clips over and over again? Were I a student, I wouldn't do that. My approach has always been to try to take it all in and make sense of it in a big picture way. An idle phrase would have little or no effect and would likely be filtered out entirely. But perhaps students would be unlike me in this respect.

The media, therefore, enables a different sort of listening, one that can emphasize the particular as well as the big picture. If a student is stuck understanding then they can start and stop the audio at the relevant juncture, think about what is going on and do other activities to process through their misunderstandings. One of the reasons I embedded an audio only track in addition to the video is that the former takes up a lot less screen real estate, so I thought students could fiddle with the Excel simulation more easily if only using the audio, though perhaps they'd rather have that downloaded to the desktop than embedded. I've got the audio for download at archive.org. It is a separate Web page for each file, which is a little inconvenient, but I hope not too bad.

I had thought ahead of time about farming out the editing of the transcripts to students in the class. Each would volunteer (via my coercion if necessary) to do their fair share in exchange for me not requiring a textbook, a reasonable quid pro quo in my estimation. I came to conclude this wouldn't work well. The students, even if well intentioned, wouldn't produce high caliber edits, because they wouldn't have a prior view of what the resulting transcript should say. Apart from some rather obscure jargon I'm prone to use, my "ums" (plentiful) and "ahs" (not too bad on that score) need to be excised from the document. The real problem is where I choose to pause when I'm speaking. Mostly I pause in the middle of a sentence, a momentary gathering of my wits but not the conclusion of the thought. I actually speed up as I come to the end of a sentence, something I'd never have known about my speaking without going through this exercise, as if to assert through the force of my voice that the previous idea is connected to the next one. I can punctuate the mess that is produced, although too many sentences end up starting with "So" or "And." For students, having to put in such punctuation would be a nightmare.

I've also got a few thoughts on repositories based on this experience. Winning factors include convenience and that the service was free. When I started playing with hosting at archive.org, I didn't yet know that Google Docs was allowing any file type. But my guess is that Google doesn't want to have Google Docs used as an mp3 download site. This is why they have a fairly modest quota - 1 Gig. Beyond that you have to pay. A one time payment wouldn't be so bad. An annual payment though is not good - simply the remembering to keep up the account is not good.

YouTube, as I said, is great for captioning from transcripts so unless my Campus offered a caption service (can't afford it in the foreseeable future) and since my videos are under 5 minutes each, YouTube wins. That YouTube does this is also a reason I'm liking Jing more. Camtasia was pretty good on captions, but really didn't help much producing the transcript. I do think that if I were to use this stuff in a course then the Campus might want to be able to brand that site, but why there should be a repository for the content that the Campus controls, I don't really get. These other external hosts are convenient and very functional. If I were making videos with material copyrighted by others, that would be a different matter. For my own stuff, external hosting of the content seems best.

There is a problem with making too much video content in advance of student use. There is no way to gauge student reaction to it and thus the design might occur with many false assumptions in place.

I've given a fair consideration to the content I'd teach if doing this course online. It's a non-trivial consideration. The type of videos featured at the link above are for analytical content. Fifteen years ago, that would have been the bulk of the course. Now I think it would be no more than a third. For narrative content, I'd like the students to do a fair amount of reflective writing. But I've no sense of how that would work with larger enrollments. (My recent teaching has been seminar classes.) One can do automated assessment of the analytic content to determine at least rudimentary comprehension. So a thought is to push on the videos and automated assessment as far as possible, to free up instructor time for the narrative stuff. But if one is pre-recorded and the other is live, doesn't that send a message to the students about what is important?

I'm still thinking through a lot of this stuff.

No comments: