What is the documetary interview workflow for Story/Premiere CS5.5?

Question

I've looked in vain for the answer to this question, which I'm sure is thoroughly documented SOMEWHERE - I just can't find it.  The few things I HAVE found online including at Lynda.com or other paid services involve CS5, which has a different workflow that involves OnLocation; this changed in CS5.5 but I don't see any on-point articles that incorporate the changes.
We've been using Premiere CS5.5 since its release but have never worked with dialogue metadata.  At the moment we're working on a documentary with a number of lengthy interviews.  In each case the interview subject is talking to an off-camera producer.  Accurate transcriptions of the raw interviews have been prepared.  Only the interview subject's ANSWERS are transcribed; the producer's questions are not (and are also barely audible on the soundtrack).  So, typically, the producer will ask a question and then the subject will give an answer that could be ten seconds or a minute.  Then another question, then another answer, all in the one clip.
So what I am looking for is a step-by-step workflow (for someone who has never so much as TOUCHED Adobe Story before) for importing the transcripts into the raw footage and then ultimately having Premiere match the transcript up with the spoken audio, and then maximizing the usefulness of this with the newly annotated clips as source footage.  Realizing that the interview answers are separated by barely-audible producer's questions, I assume that one challenge will be to let the software know that it should only be paying attention to the answers, and lining the text up appropriately, and not getting confused by the questions.  Perhaps the answers need to be subclipped?
If there's a tutorial on this subject I would very much appreciate being pointed to it.  If not, if someone could do a step-by-step (beginning with an empty Premiere CS5.5 project and nothing in Story) I'm sure there must be a number of other folks in my position who would benefit from it.
As a product suggestion: for documentary applications it seems to me that it would be much better to be able to cut and paste text directly from a text editor into Premiere and then let it fit it to the audio instead of going through Story.  While I understand it was even more cumbersome prior to CS5.5, I don't think interviews in a documentary need most of what Story offers (unless I'm missing something), so it seems like an unnecessary complication that could be eliminated.
Thanks in advance for any help.
Peter

Answer 1

Well, I tried following your steps, and I'm going wrong somewhere.
I created the film script and four scenes (one for each of four clips), giving the speaker's name as a character name (it's actually the same character in a continuous interview broken into several five-to-ten minute clips).  I let Story number the scenes (1 through 4, as expected).  I then exported an .astx file.
Within Premiere, I located the four clips and in the Scene column numbered them 1, 2, 3, and 4 respectively.
When I right-click on one of the clips, select "attach script file," and point to the exported .astx file, I get the following error message: "Failed to attach script file to the following clip: \MyProjectName.prproj\FolderNameWIthinProject\00019.mts [Matching scene number not found]" (of course I have replaced the path and folder names in this example with generic names since the actual names reveal my client company's name).
Opening and inspecting the .astx file in a text editor I confirm that it does in fact contain scene numbers 1, 2, 3, and 4, respectively, for the numbered scenes.  Likewise I confirm the numbers 1, 2, 3, and 4, respectively, in the Scene number column for the four relevant clips.
Any idea what the problem is?
Peter