At this year's BEACON conference, Art Covert and I gave a presentation on how to give engaging presentations. The slides for my part of the presentation are available here. If you browse through the slides, you'll see that some of them are rather text-heavy. I did this on purpose, to demonstrate that text-heavy slides are not the bane of engaging talks, if presented appropriately. By all accounts, despite the amount of text on my slides, my presentation was engaging.
One of the first recommendations that we usually give our students is to put less text on their slides, and instead use images, drawings, and diagrams. I wholeheartedly agree with this recommendation in principle, and for most of my presentations I spend a lot of time developing graphics instead of text. However, I think that there are other recommendations that we should emphasize first; the text-heavyness of slides really is a minor issue. After all, one can absolutely give an excellent presentation using text-heavy slides, and one can also give a terrible presentation using slides with very little text.
Thus, I propose to replace the advice of using less text with the following advice, which I call Claus' 1st law of giving presentations:
All text on slides needs to be read aloud, word for word.
(Similarly, all visual elements need to be pointed out and explained.) The problem with text-heavy slides arises when the presenter paraphrases the text or, even worse, talks about something completely different while displaying the text-heavy slide. The audience members cannot process two different streams of information, one visual and one auditory, at the same time. Hence, they have to make the decision of either listening or reading. Whatever decision they make, they will likely feel that they missed something, and they will be more likely to disengage from the presentation altogether. I believe in particular that when audience members start reading instead of listening to the presenter, the presenter has lost them at that point and will likely not get them back for the remainder of the talk.
By contrast, if the presenter reads the text word for word, the visual and the auditory cues reinforce each other. The audience will be more engaged and less likely to drift mentally. For the same reason, I like text to appear line by line, as I read it, rather than all in one block. Again, this increases the synchronicity between the visual and auditory cues and gives the audience members less opportunity to drift. If you ever get the chance to see a talk by Steven Pinker, observe how he reinforces auditory and visual cues. Nearly his entire talk is spelled out word for word on his slides. His slides basically serve as sub-titles to his voice. (You can probably find an example of this on YouTube. If you do, please post the link in the comments below.)
Note that an additional benefit of my 1st law is that you can never make slides too text-heavy, because there is a limit to how much text you can read during the time alotted for your presentation.
tl;dr: Think of text on your slides as sub-titles to your talk.