Is using voiceover adding any value to e-learning?

Here's my understanding of Clark & Mayer's research. It showed improvement with audio when the narration was accompanying graphical content, for example, describing parts of a machine as the slide highlights them on an illustration. In such cases, the combination of audio and visual helps because it uses two input channels (ears and eyes) that each take in different but related info. In other words, narration is great when you've got something besides text to show on screen. Of course, the transcript (that is, Notes, closed captioning, or other text that can be turned on and off by the user) should also be available to accommodate those who can't hear.

However, when the audio is merely reading the text on screen verbatim or almost verbatim, that actually increases cognitive load and, therefore, decreases learning. Not to mention that it's frustrating for users if you force them to stay on a slide until the audio finishes, when reading is faster.

In my experience, it seems some people want audio mostly to enforce a given seat time, with the hope that if they make users wait for the narration to play, the users will have to learn something. But if you don't trust your learners to read what they're supposed to, you can't trust that they'll listen to what they're supposed to. And that's a motivational issue, not an eLearning issue.