You are mistaking the baked result of the Audio to Keyframes keyframe assistant for the actual waveform display. It works as it should. hit rewind and watch the tutorial again. You skipped over some important detail.
I'm not so sure I am. See, I've done this once before through a tutorial from this book, following their exercise with their provided media files. I had a bit of trouble translating their exercise to my own media due to their wording, so I hunteda around on Youtube and found that tutorial I've already linked to, which was a bit easier to follow when not using the books files.
ANYwho, I think the problem might be the media. It was recorded using a wireless lav mic and I'm thinking that perhaps it recorded in mono, hence the results I'm seeing in AE CS5. Re-recording isn't really an option, so I'm wondering now if maybe the right thing to do is take the audio file into Soundbooth and split it somehow?
Edit: I've put up everything in this post's screenshot, in case that helps clarify.
Craps, I WAS forgetting a step. Converting autio to keyframes magicked the wavform into left, right and 'both' channels.