Harold Jarche, saying that learning content should be hackable, included a link to a Ruth Clark article from 2002, Six Principles of Effective e-Learning (PDF). Clark’s widely known for her research-based approach to understanding how people learn. The article’s well worth reading; this post is my summary of his six principles. First, here’s what she means by e-learning:

For the purposes of this discussion, e-Learning is content and instructional methods delivered on a computer (whether on CDROM, the Internet, or an intranet), and designed to build knowledge and skills related to individual or organizational goals. This definition addresses:

• The what: training delivered in digital form,
• The how: content and instructional methods tohelp learn the content, and
• The why: to improve organizational performance bybuilding job-relevant knowledge and skills in workers.

The multimedia principle: Adding graphics to words can improve learning.

Clark cites research showing up to 89% increase in test performance when learners studied text with graphics (compared to text alone). This makes use of the dual-encoding theory. As she points out, though, gratuitous or irrelevant graphics actually detract from learning (as many existing courses demonstrate).

The contiguity principle: placing text near graphics improves learning.

“If the words and the visuals they describe are separate from each other, the learner needs to expend extra cognitive resources to integrate them.” If layout or screen scrolling separates text from visuals, a heavier burden falls to the learner’s working memory.

The modality principle: explaining graphics with audio improves learning.

Here, again, Clark is recommending designing for optimal fit with working memory. By using both the visual and phonetic components of working memory, developers can have a more positive impact on learning. This isn’t to say you should use audio only, but rather strategically.

The redundancy principle: explaining graphics with audio and redundant text can hurt learning.

Clark’s talking about presenting text and reading it aloud at the same time, especially in the presence of visuals. For example explaining a graphic with both text and narration can overload the visual component of working memory.

I also think that most adults don’t like being read to, particularly when they can read the words for themselves. My own rules of thumb in this situation are:

• The less on screen, the more in audio. For example: if you’ve got a process diagram or an animation, don’t clutter it up with a lot of on-screen text.
• The more on screen, the less in audio. For example, say, “here’s a summary of Feist v. Rural Telephone Service.” Then keep quiet. If something was important to know before reading the summary, share that before you get to the summary.

The coherence principle: using gratuitous visuals, text, and sounds can hurt learning.

After my own high-pressure introduction to computer-based training, I decided that the worst course a person would ever develop would be the third. On the first one, you have no idea what you’re doing. On the second, you’re happy to have some idea of how to manage your tools. It’s the third one in which you go crazy.

You’ve seen the same kind of thing in PowerPoint presentations with every conceivable transition or text documents with more fonts than a Bruce Willis film has explosions.

This is a hard one for me, but I think Clark is right: this principle says that less is more when learning’s the primary goal.

[This principle also] suggests that visuals or text that is not essential to the instructional explanation be avoided. It suggests that you not add music to instructional segments. It also suggests that lean text that gets to the point is better than lengthy elaborated text.

The personalization principle: Use conversational tone and pedagogical agents to increase learning.

One intriguing finding that Clark cites: people who reviewed a program on the same computer that presented the program gave it higher ratings than people who made their evaluations on a different computer. “People were unconsciously avoiding giving negative evaluations directly to the source.”

She goes on to discuss the use of agents (e.g., a character offering learning advice). What’s key is not the appearance of the agent so much as an instructionally valid role. Does the agent offer genuine help, or just mindless promotion of the matter at hand?