As someone who has not yet had the honor/pleasure/ to teach a large lecture class, I have had the luxury of deciding quickly and without much agony to just drop the textbook. Instead I created my own coursepackages consisting of primary sources paired with articles that I like. The thinking behind it was to make students realize that history is not a presentation of facts but an ongoing argument.
Did it work? Well, nope. Most students are so wedded to the notion that they need to memorize dates and events to succeed in a history class, that they read the articles and primary sources just for the information about facts/dates/events, and miss the whole "argument" part. I have learned to devote much more time to discuss with students "how" historians actually work. This, of course, comes at the cost of content delivery. But I am honestly starting to question whether this is really such a big deal. It would be different if we taught high school. But we don't.
I mean, why are we doing what we are doing? And to clarify - I am talking about teaching undergraduates here. Most of them (even the history majors) will not, I repeat NOT become historians. So what do we want to get across? Will there be a terrible void in our students' lives if they leave college without knowing about the Encumbered Estates Act? Probably not. I think the loss would be if they left college knowing every single detail about the Encumbered Estates Act but still thought that they understand the world; different cultures; different people; if they just cram dates and names into their heads and by being able to recite "a" narrative about the way things were/are.
There have recently been more and more discussions among our colleagues about whether the sciences can offer a useful model for the teaching of history. In the natural sciences you have the lab. Students learn-by-doing. In history, learning-by-doing often comes too short, I think. We teach a narrative first. Then we raise the issue that this narrative can be contested. We start with textbook readings. We quiz students on the textbook. And only then do we give them primary sources to analyze and to discuss. Maybe this is the wrong order? Could a survey class be like a wiki? Imagine this: your students read primary sources and articles. Then they write the textbook themselves: the weekly assignment could be to go online and write a paragraph of narrative in a wiki format. Many universities provide the necessary webspace. Mine for instance (big public northeastern U) provides instructors with blog space and wiki space. If you utilize a wiki, students can edit what's already being written. Every student builds the mosaic. It will be chaotic. It will be a mess. But imagine how much fun it might be.
Actually, if truth be told, I am afraid my main teaching goal is to have fun while doing it. I am a hopelessly frivolous teacher, I fear. But back to the textbook and back to serious: I think it already helps if you teach students to refer to the textbook like they would to a monograph. Don't let them get away by writing "the book says." Make them say, "As Hewitt and Goldfield argue..." That way, it's easier for them to wrap their minds around the fact that just because something was put in print it is not truth written in stone.
These are just my five cents. And bear in mind: I have not taught the survey, yet. I am spoiled rotten. I live in Disney land.