This article discusses some of the fundamental methodological challenges related to doing historical network analyses of the web based on material in web archives. Since the late 1990s many countries have established extensive national web archives, and software supported network analysis of the online web has for a number of years gained currency within Internet studies. However, the combination of these two phenomena—historical network analysis of material in web archives—can at best be characterized as an emerging new area of study. Most of the methodological challenges within this new area revolve around the specific nature of archived web material. On the basis of an introduction to the processes involved in web archiving as well as of the characteristics of archived web material, the article outlines and scrutinizes some of the major challenges which may arise when doing network analysis in web archives, among others such issues as completeness, construction of a corpus, temporal and spatial inconsistencies, and cross-archive analyses. The article uses an ongoing case study to feed the methodological discussion, namely the political network on the web which was available to a voter at the Danish parliamentary elections in 2011, 2007, and 2001. As the Internet grows older historical studies of networks on the web will probably become more widespread and therefore it may be about time to begin debating the methodological challenges within this emerging field.
Social Science Computer Review, 2013, Vol 31, Issue 3, p. 306-321