The science of using audio in e-learning

Next week I’ll be presenting at the eLearning Network on best practices for using audio in e-learning. One reason I wanted to speak was that this would give me the excuse to spend some focused time doing research: What scientific evidence can be used to ground day to day instructional design decisions?  It’s many years since I was a Psychology undergraduate, so I relished the chance to see how science has moved on.  

One of the greatest advances since I was at university is Google,  so I started by trying to answer three questions:

1. Is e-learning with audio more effective than e-learning without audio?
2. What’s the level of recall from reading vs. hearing?
3. Are there certain types of audio that are more effective than others?

Apart from a nostalgic reunion with Craik & Lockhart (1972), there was little out there where science really informed the debate. Cognitive Load Theory gives us some insights, but it’s not really answering the questions at the level of granularity that is really useful to an instructional designer e.g. if I add W type of audio to X type of screen with Y type of audience learning Z type of content, what should I expect the impact on recall/ behaviour change to be?

And possibly this is the heart of the issue: corporate learning and academia don’t mix…or don’t mix enough. There’s some cultural chasms between these two worlds:

• Most Psychology experiments use Psychology undergraduates as they’re free – are they representative enough of people at work to draw valid inferences?
• If a corporate L&D manager thinks something will work, the whole business should have it  – there’s rarely time (or the inclination) to run a control group vs. test group analysis
• Whilst L&D managers have access to data about people at work, they do not have the science/ statistics knowhow to frame a hypothesis and experiment to derive statistically significant results
• If as an L&D manager I do find that something works really well, will my company be happy for me to share this information with competing companies?
• There is also the challenge that the ideal experiment would be measuring the performance change for different versions of e-learning with different types of audio, rather than just seeing how well people answer quiz questions at the end of an e-learning module

I don’t think I can bridge these chasms by next week, but I trust that those attending eLN will take away some useful insights into how to use audio to enhance the quality of e-learning for learners everywhere.