Monday, August 27, 2007

Views on video in eLearning

If the reader's objective is to tape face-to-face presentations, and then deliver these as "e-learning," I predicts that the reader will be disappointed with the outcome.

Although many people believe video to be 'active' and 'animated,' in reality it is an extremely passive form of training and results in very low knowledge retention. Furthermore, people can read and ingest information about three times more quickly than when hearing information in spoken format. This means that unless the presenter is attractive, professional and smooth, the viewers will quickly disengage. We have found that videos that last more than about two minutes are abandoned quickly; this is even true for YouTube.

If you want to offer employees self-paced training, I suggest that you create Web-like content in lieu of video. This means that the content has tables of contents so that students can establish their own navigation path that takes them to material of interest. This becomes very 'active' learning because the students are in control of the experience."

Thursday, August 23, 2007

eLearning theory and Rapid eLearning Tools:

Theory has to do with figuring out the best course layout (and I don't just mean
graphic design - I mean content organization), the best element types (tables,
do-it-offline exercises, bullet point presentations, etc.) and the best delivery
distribution (face-to-face versus self-paced) for the various learner environments.
For Rapid eLearning tools, have they adopted a "PPT conversion" paradigm, a
"web content and organization" paradigm, a "CBT visually stimulating" approach,
or a "we call ourselves 'rapid' so we can sell more copies of the same old thing" approach.

Once again, it is important to choose the right tool for the job.

For "standards", here is the issue. The standards provide a collection
of data that you can store. It has names like "score", "session_time",
"interaction.score", "interaction.correct_answer", and
"interaction.student_response". You need to understand what you can use
each of these for in the context of what the Learning Management System
or Tracking Engine will store and report. What use is it to capture
every student_response to a question if you can never find this
information out? Therefore, you would need to understand how to use
these variables and the delivery platforms in order to actually get
useful information about the performance of your content (in addition to
the performance of the learners.)

Tuesday, August 14, 2007

What size font works best for eLearning?

If a course designer has established a 10 pt font, they are on their way to serious frustration. The concept of a fixed-size font is related to printed material, so it should go along with establishing a paper size for eLearning. How many points are there on the end-users screens?

Good web practice dictates that you should use a relative size font. The base font size should be the end-user's choice. In MS-IE, you can go under Tools, and find the button for "Size". In Firefox and other Mozilla browsers, the learners can select their base font size much more freely. Why does this matter? For accessibility. A 10 pt font may be fine on an 800x600 pixel screen, but it is too large for a PDA, and it is too small if the end-user has visual disabilities or they have a 1600x1200 19-inch screen. Instead, the font size should be specified as a percentage of the end-user's default (e.g. 100%, 80%, 120%, etc.)

With web pages, it is also important to select:
1. Acceptable combinations of foreground and background color (if there is not enough contrast, the font will be illegible).
2. Acceptable foreground colors: Keep in mind that in many cases the end user may turn off images and background images (like on a Blackberry). If you choose a white foreground font with a dark background image, the text will be invisible because the PDA or other rendering device will provide a white background. (So use dark colors.)
3. Relative font sizes, font families, text decoration, etc. for different elements of the page, e.g. Heading, summary section, test question, test answer, test distractor, Outline numbering, ... I recommend that a read-through of the Cascading Style Sheets specification be made before going much further with this exercise.
4. Define navigation graphics (forward, back, glossary, test, try-this)
5. Define navigation texts ("next", "back", "Grade the Test", "Submit Survey Responses").

It would be worthwhile to review a complete web-based eLearning course (not a PowerPoint presentation displayed through the web) to see all the different kinds of elements that will be used. Otherwise, the specification of a 10 point sans-serif font for "eLearning" is like saying, "I want to use blue paper for Word documents."

Friday, August 10, 2007

ADA 508 conformant courses

If you stick to web standards, like W3C HTML, your content becomes ADA 508 conformant if you follow some simple rules. Beyond that, the benefit of W3C HTML is that your content becomes searchable, deliverable on multiple platforms (e.g. PDAs), deliverable without a lot of plug-ins, and faster to download.

Most plug-in based content is inaccessible to blind
readers. 508 should be about making the content accessible to everyone
at the same time, not about certain pages accessible to blind people
with 3 other pages accessible to hearing impaired only. Also, one of the
best features of Sesame Street (and other shows including my favorite,
"Between the Lions") is that they can offer repetition of presentation
of material without requiring a live instructor. Web based training can
also offer this (if the reader is interested in the topic).

Moving from a trainer to eLearning. Do you try to replace the trainer?

I guess the question is whether the avatar is a good enough replacement
for a live instructor that learners.
1. Can you interrupt it?
2. Can you ask it a question?
3. Can it understand the real reason for the question? Usually people
ask a question that shows a symptom of their misunderstanding, rather
than the fundamental reason for the question. If only the symptom is
answered, the fundamental misunderstanding usually remains.
4. Can it answer your question?

With electronically delivered training the different learning needs
pose challenges and offer possibilities, especially when you remove the
live instructor. I have never believed that a pre-recorded instructor
was a good substitute, however I see many trainers and content
developers sticking to the old paradigms. That is, "lets see how close
we can get to instructor-based training while delivering remotely":
1. Record the instructor on video or audio.
2. Expect that the course will only be accessed once (so provide only
linear navigation)
3. Provide only the instructor's presentation (both what they gave as
material, and only showing their answers to the questions posed during
video taping).
4. If they can't record the instructor, drop back to an avatar.
5. Remain in PPT "instructional design" which means breaking the content
so that it fits 3 points per slide, rather than organizing content to
coherent chunks with natural divisions. Here the fundamental issue is that the trainer is trying to replace the form of face-to-face instruction while delivering electronically.

These short-cuts result in courses with low value and low attention
retention. Some solutions:
1. If you record a video or audio, keep it under 2 minutes. If it is
longer than that, the user will "float" away to check e-mail or play
freecell. Make sure the video downloads and plays in less than 20
seconds. Justify its value: Does it show a process or is it just a
talking head? If a talking head, are there any visual queues that are
absolutely essential that could not be given with lower bandwidth
requirements (e.g. "this is important"). If audio doesn't match the
text, the user will have to follow one or the other, resulting in
cognitive overload.
2. Make the course following web practices: lots of navigation, let the
user choose their path, let the user reach any content within 3 logical
clicks, make it easy to get back to the content when the user needs to
look something up. Many LMSs block the student from viewing content
after they have completed the course. This would remove re-usability
for the end-user.
3. Show the same material in multiple forms. This is where the web (and
creativity) can really shine. Provide a screen snapshot animation, a
broken down step-by-step procedure, a longer article, a try-this
example, a quiz, a tracked test, links to other sites (vendors,
SecondLife, FaceBook, Wikipedia). Users will naturally gravitate to the
presentation forms that are easiest for them. (Don't try to control the
user's will just lose them.)
4. People can read much faster than spoken language. People can also
understand content narrated much faster than we normally speak. Audio
content speed should be adjustable. It is amazing to watch a blind
person using a blind reader. They often set the audio to go really
quickly. Keep in mind that audio delivery is very linear, so it needs
to be sped up for people who can only access from audio.
5. Use the multiple navigation abilities (links to drill-downs) to break
the content based on what will give the best organization for learning,
rather than what will fit on one slide without scroll bars. Remember
scroll bars? They work fine if vertical. Horizontal scrolling,
however, should be avoided.

Thursday, August 9, 2007

What is important for eLearning design?

I keep running into people who
completely discount other people's ideas and abilities simply because
there are a few typos or spelling mistakes in their presentations. (I'm
not referring to Phil as guilty of this, but because he illustrated that
this can be an issue.) I see this as a symptom that also affects course
design. These same people will spend hours agonizing as to whether
their bullet points should do a flip as they fly into the page, and will
choose visually animated content any day over solidly instructionally
designed material.

Yesterday, I was talking to someone (with very strong graphic arts and
engineering capabilities), and they mentioned that they had yet to see
an eLearning course that they found really interesting. It dawned on me
that indeed, eLearning courses will only be as interesting as the content.
Adding visual stimulation is the kind of window-dressing/ distraction that
the "analysts" seem to push to compensate for poorly designed or even just
run-of-the-mill content.

I'd like to get people to move towards the idea that if your content is
average (and 99.9% of content falls into that category), the best way to
compensate is to give the student the ability to control their session.
It is more important to have students who use the content and are not
immediately repulsed.

For example, I have seen PPT presentations that were shipped out with
mandatory (but extremely boring) content. The company didn't want to
spend the extra day of work to port it into a more web-friendly
structure (and I don't mean Flash) because they thought it was a waste
of time. I watched how the audience viewed the content:
1. They opened the PPT (it was a PowerPoint Slide Show)
2. They clicked the "Next" button on every page until they got to the
certification piece. Average time on each page was probably 5 seconds.
3. They completed the certification and closed out.

My observations:
1. When you multiply 5 seconds/page times 20 pages times 1000 employees,
you realize that a lot of time was wasted.
2. If the content was more "friendly" it would have a link to jump
straight to a "What's New" summary, and a link to the "Certification".
This would actually result in employees spending useful time on what is
"new", and then jumping to the certification page, rather than
dismissing every page as quickly as possible (without reading them).
3. The flying bullets only slowed the students down from getting back to
their other daily tasks.

Wednesday, August 8, 2007

What type of course do people want to create?

I believe that IS the primary market in eLearning. Here's what I mean.
Much of the training community seems to still be at the Web 0.5 stage
e.g Web .5 is the technology we used before the web.
Most SME's (subject matter experts) like to give presentations, but now they are realizing that because of financial pressures, they must adopt the new tools. However, and
this is ironic, trainers don't want to learn new tools and approaches. So
the number 1 solution is to outsource. The number 2 solution is to use pre-web
technology e.g. PowerPoint. The end products tend to reflect
the Web 0.5 approach because anything that is Web 1.0 or later is "too

Monday, August 6, 2007

Thoughts on assembling a course for use at delivery time

At McDonalds you can re-use fried potatoes (french fries) and Soda
with any other meal. And the result is a McDonalds "Happy" Meal. This
is "food", but I wouldn't call it nutrition or fine cuisine.

Using Spaghetti Sauce as a trackable re-usable object could make sense;
however what is the point of tracking spaghetti sauce when you want to
track that the user ate pizza, or that the user had a balanced meal?
Or, we could put the spaghetti sauce into an LCMS (Linguine Cannelloni
Management System) and have the LCMS re-use the spaghetti sauce
automatically for us - put it with lobster, put it with ice cream, put
it with peanut butter.

If we require that a SFO cannot lead to another SFO (e.g. the wine
course preparing your mouth for the artichoke course), then we shouldn't
be breaking the objects down so small simply so we can put them into the
blender (LCMS) to re-use them. I don't agree with making each SFO
primarily for the purpose of reusability. I should expand my kitchen so
that what it delivers is complete courses (pun intended) rather than
just building blocks reassembled (because it is easier to let the
blender reassemble the meal). If I want to re-use the components, I
should have the tools that let me blend them correctly before they leave
the kitchen.

In SCO language, what I'm trying to say is that I think the idea of
having an LCMS assemble a course for us at delivery time is a bad
solution. Yes, you should be able to re-use components, but this should
be done while you, the instructional designer, assemble the course for
the specific purpose. If you start mixing and matching content from
different authors simply because SCORM says you can, you will end up
with really bad courses that look disjointed. But this goes contrary to
a common perception that SCORM is about authoring based on reusable
components. I think SCORM should be about having the option of reusing
your courses in different delivery mechanisms.

To blend the idea of a multi-navigation course with a consistent
look and feel, the best solution I see is to make a larger SCO. For
granularity of tracking, the LMSs/courses need to use the interactions
and objectives groups. Then, you can serve a complete course (meal) as
a SCO, and provide sensible instruction with complete tracking so that
you know what each learner answered on every question. We need to have
final objectives that go beyond just creating a SCORM conformant package.

SCORM does not allow for dynamically generated content because of loss of interoperability/portability, but I have seen that you can fool it if your dynamically generated content is an "asset", e.g remote web page or remote graphic. What frightens me is
that there is a SCORM "group think" mentality that has taken over where
the concept of dynamically generated content (at delivery time) has
been replaced with the concept of an LCMS that organizes the chunks
mechanically. The two concepts are quite opposite and should not be
confused. One is a coherent approach that can provide up-to-date
variable content. The other is a solution looking for a problem to
solve based on people who want to use "big iron", resulting in
disjointed ransom note courses.

Thursday, August 2, 2007

Implementing AICC

AICC is actually an easier task to implement than SCORM. Here are the
components you would need:
1. Your database to create a unique session identifier for each
student. Frankly, you could just use the student ID. It all depends on
how detailed you want to be in your reporting.
2. When the student is ready for the courses, you need to create a
launch page - that is a dynamically created web page with a list of the
courses available to the user with URLs pointing to those courses. The
URLs have to include the session identifiers.
3. On your server you need an application that can catch and parse the
data coming from the course. Most AICC courses never use or ask for
previous history data, so you may be able to get away with not providing
the getParam response. The tricky part of this application is that it
has to organize the incoming data (by SessionID, perhaps) and store it
for reports later. When this application receives an exitAU message, it
should end the session (and ideally store the elapsed time, although I
haven't yet seen an LMS that does this.)

For a low price, I can provide a basic catcher routine, but you'd still
need to do a lot of work to store the data. We use this internally to test AICC
courses to ensure that they're reporting the information we intend them
to report.

Alternatively, contact me directly, and I will give you the e-mail
address for someone I know who created an AICC-based LMS. They may
resell it to you for your development. The price would end up being
lower in the long run than developing it yourself.

Wednesday, August 1, 2007

LMS Recommendations

The "back end" is where most LMS problems -- and costs -- tend to be concentrated. Unfortunately, few customers get to see and play with the "back end" of their LMS until after they've paid for and begun using the system in earnest.

With this in mind, I recommend doing the following to save yourself time, money and headaches:
  • Determine whether the LMS provides a way for you to export data and produce your own reports. "We have heard of LMSs that have data stored in such a confused manner that it's impossible to retrieve them from the database," says Moser. Worsening the problem, he says, is the fact that some LMSs charge $25,000 for every custom report that a customer needs. ("This is for a single-event report, not for programming a report that you can re-run. So, it is essential that you are able to get the data out of the LMS (in case the LMS goes broke or doesn't have a canned report that meets your needs).")

  • If you are looking at a learning content management system (LCMS), or an LMS with authoring capabilities, will you be able to take your courses and serve them outside the LMS? "Once again, if the LMS goes broke, gets bought, or does an upgrade that you don't want to pay for, it is essential that you be able to take YOUR content and host it elsewhere."

  • What kind of support does the LMS vendor provide? "We have seen many (some of the best known ones fall into this category) where you are routed to a customer support agent who barely knows how to spell 'AICC.' They can waste weeks of your time giving you the run-around. When negotiating the contract, I recommend a clause that if you don't get a satisfactory support response within two days, you have direct access to the product developers/engineers, rather than being stuck with tech support. If you don't get this support, the LMS should pay a penalty."

  • Be wary of contracts that include off-the-shelf content. "As a thought experiment, consider how many off-the-shelf PowerPoint presentations you currently use internally. How many off-the-shelf face-to-face training courses do you use? Your e-learning needs will end up being about the same -- that is, you will need to develop most of your own training. The extra cost for a library of stock courses might be an inefficient use of resources."