Over time, people start to develop all sorts of etiquette for these type of meetings (it's roughly the same for people who go to the room and just work on a laptop) - if their attention is required, their name is called out, and the context for the request is spelled out - the assumption is that the person has been working on something else.
Microsoft Exchange (or iCal, or whatever mechanism that feeds the various calendar apps) helps a lot - everyone has the meeting in their calendar, and so various PDAs, Smart Phones, and Calendars of each individual light people up and most people come online within 60-90 seconds of each other.
This is actually a really good solution for the downside of multi-tasking - I can focus 100% on something else, and then, when I'm actually required for the 5-10 minutes of a 90 minute call, I can pay attention at that point.
In effect, you can get 20 people on a call, and instead of taking up 30 person hours (20 * 1.5 hours) - it may cost the organization a small fraction of that, with pretty much the same results.
Clearly for those meetings in which full participation is required, other tactics are required - and I really do try to close my laptop and pay 100% attention to what is going on.
I haven't noticed much difficulty, btw, in working by teleconference - I think people just learn to develop new techniques to optimize the medium - Learning to Mute, call from a good line, judge pauses appropriately to break in, etc..
Another cool solution is the recent feature 37Signals added to Campfire for conference calling, which posts the call transcript back into the chat - nice for people who can't be on the call, or couldn't hear very well.
I think that it could become a liability to do this at some point, though. When everybody is calling in and people have flaky quality, you are wasting at least some time. If a meeting is more effective in person, then have that meeting in person! You shouldn't be afraid to exclude one or two people if they're traveling; they're already excluded anyway. Why provide an illusion of close involvement?
I've never noticed flaky microphones or audio quality from remote participants. Ever. But I've often found that local participants are completely inaudible. If every local participant just used a bluetooth microphone (or really anything dedicated but the shared conference call microphones), things would be so much better.
For example: Most of the meetings I attend are largely ad-hoc with printed material being handed out at the time of the meeting - usually not available on the company intranet (or emailed) until after the meeting. It makes it immensely difficult to follow along and contribute in a meaningful manner.
Honestly, there are many benefits to having everyone "remote" for a group meeting. It focuses the meeting. After the meeting you still have the luxury of walking over to your co-workers's desk and chatting in detail about the problem.
I think for larger meetings this "level playing field" solution would probably work much better for video conferencing or combined text/audio conversations such as on Skype. This way simple cues such as "I've got something to say when you are finished" can be used without interrupting.
However, this does require that everyone be "text-fluent", i.e. able to effectively produce and understand communications in a text-chat medium, in a way that feels like conversation rather than people sending short memos to each other. True of most people <30; intermittently true of people above that age, depending on the person.
The other thing I've noticed with a large number of dial-in participants is that there can be a lot of silent attendees and very few people talking. YMMV
i'd also like to it possible for only one person to speak at the same time.
and keep track of who's talking the most. in some contexts that can be helpful. it might be useful too if each partipant only had the floor for so long.