I'm at the annual conference of the American Society for Indexing, and it's amazing how much time indexers spend thinking and talking about what readers (indexers users) want and how they use an index. Yet indexers have absolutely zero time to do user research. In fact, there have been very few usability studies at all.

A study by BNA a few years ago did show that legal indexes significantly improved performance of individuals asked to find relatively simple legal information in a text compared to searching. We also know that increased use of cross-references (see and see also) improve performance in using indexes.

But no individual indexers facing the problems of indexing a particular text really know how what the users for that text would like to find or how they would like to find it. All we can do is guess. And commiserate. And dream of being able to do real usability testing of any sort.