the fussy details of survey formatting

I've been in the OI office most of the day, and while I'm waiting for various people to finish meetings so that they can meet with me, I've passed my time by rewriting our survey from it's rough draft in Word into something more formal in Excel. After my experience last year, I'm much more aware of the differences between how something looks on the screen and on the page, not to mention the importance of the order in which questions are organized or how close together they are to one another. I'm trying to make this survey far more user-friendly and easily-read than the one that we used last year in India. For the most part the changes are straightforward (I'm increasing font sizes and leaving larger spaces for answers), but it's sometimes tricky to decide whether or not to create a chart or to break a question up into parts or to just restate it more than once (and I should say that JP, who co-wrote the survey and will be using it in Chennai, has been very good at figuring out ways to state things clearly). Questions that involve a second and third "if yes, then blah blah blah..." question are particularly annoying to me. It's important that the initial question is answered clearly yes or no, because sometimes the enumerator will fill out the answers to subsequent questions even when they aren't applicable, which then can present a whole mess of confusion.

You might be wondering if there isn't some professional standard that I could be following in this situation. Well, sort of. Our survey questions themselves are loosely based on the LSMS, which is the World Bank's Living Standards Measurement Study. This is an exhaustive series of survey modules asking about pretty much everything you can ever imagine wanting to ask a household in a developing country. The thing about the LSMS, however, is that it is insanely long and complicated, and the questionnaires themselves are designed to be administered and filled out by professionals. Also, the LSMS does pretty much all of the coding within the survey (i.e. if the person gives this answer, you put a 1, if they give this answer, you put a 2, etc), which, while making the answers shorter, adds to the complexity of the layout (and the length of the explanatory text). (My survey won't involve very much complicated coding, because it's mostly quantifiables and yes or no questions, and all of the coding will be done by me when I enter the data.) The bottom line is that the LSMS is not even remotely user-friendly, so it's not much of a model for me in that respect.