Linked title mentions: a new automated link search candidate

Abstract

Many webometric studies have used hyperlinks to investigate links to or between specific collections of websites to estimate their impact or identify connectivity patterns. Whilst major commercial search engines have previously been used to identify hyperlinks for these purposes, their hyperlink search facilities have now been shut down. In response, a range of alternative sources of link data have been suggested, but all have limitations. This article introduces a new type of link that can be identified from commercial search engines, linked title mentions. These can be found by querying title mentions in a search engine and then removing those not associated with a relevant hyperlink. Results of a proof of concept test on 51 U.S. library and information science schools and four other sets of schools suggest that linked title mentions may tend to give better results than title mentions in some cases when used for site inlinks but may not always be an improvement on URL citations. For links between or co-inlinks to specified pairs of academic websites, linked title mentions do not generally provide an improvement over title mentions, but they do over URL citations in some cases. Linked title mentions may also be useful for sets of non-academic websites when the alternatives give too few or misleading results.