The appropriate unit of analysis is not related to the possible world of exposure for a model reader (or I should say "need not be"), but is instead related to your research question. You are correct in noting that the web, as an object of study, requires more care be taken in defining the unit of analysis, if only because there is not a deep tradition of research or of information design to fall back on. In analyzing newspaper text, for example, a researcher can refer to an article or paragraph and be assured that others will comprehend what he or she is measuring. Perhaps the only similar measure for the web is the "page," though even this can be problematic given the dynamic nature of the web. It is therefore vital that any examination of web content explicitly define the unit of analysis. That the web is assembled and structured through hyperlinks is, I think, not particularly important; interesting, yes, but not a limiting factor. Books, newspapers, television and conversation are also structured in various ways and bound together with glue and citations and other structural stuff. The question comes first, and both the unit and the method of analysis follows that question, I think. Alex -- // // Alexander Halavais // Graduate Director of Informatics // University at Buffalo School of Informatics // contact info: http://alex.halavais.net //