Newspapers' Use of URLs in Daily Content Has Grown Steadily since Mid-1990s

Newspapers have recently adopted Web URLs (Universal Resource Locators) into their content. Herein we briefly report an exploratory investigation of this phenomenon. There are a variety of published studies on the role of the Web within news organizations.1 However, we found no empirical work that addresses the specific area of this study - the way in which newspapers refer to and cite Web addresses. This is not a trivial question. An annual survey conducted by the Georgia Institute of Technology offers some insight into the role that newspapers play in referring individuals to the Web. When respondents were asked, "How do you find out about new WWW pages/sites?", 63 percent mentioned print media - magazines or newspapers.2 One of the potential problems with citing URLs in newspapers may be the instability of Internet addresses. Several scholars caution that there is no guarantee that a Web site will remain stationary or that its content will remain the same for any length of time.3 One study of URL stability in scholarly scientific journals noted an 18 percent access failure rate over a six-month period. The authors describe these failures as "site instability" with multiple changes over the time frame of the study or URLs' disappearing completely.4 The use of URLs has been a concern of editors. American Journalism Review reported comments by a number of editors and publishers about the practice of including URLs in stories, suggesting that privacy issues, the motivation of the Web author, content accuracy and decency issues were important considerations for editors.5 The decision of whether and when to include URLs is still an imprecise standard with style and usage rules evolving for newspapers.6 The Associated Press is also engaged in the evolution of URL standards. AP published its guidelines when it announced in early March 2000 that it would begin including URLs with national news, sports and financial wire stories.7 Methods Researchers wishing to receive a detailed report of our methods may contact the lead author. Data were acquired from the Lexis/Nexis "General News, Major Papers" resource for the years 1994-1999, with a randomly constructed week sampled to represent each year. An ideal search strategy would have been to search for the appearance of "www." in all stories ("http://" is not placed before many address listings). However, "WWW" is a key word in Lexis/ Nexis, which complicated the preferred strategy. Therefore, the search was limited to stories in which the terms "Internet" or "World Wide Web" appeared. Observation of URLs in this sample is obviously not an ideal means by which to generalize to all newspaper content. But, in the absence of other studies of this kind, we feel these data do have some value, and this approach did allow the examination of attention to the Internet and the Web by newspapers. For most of our work here, the URL was the unit of analysis. This sample yielded 1554 URLs. Another characteristic we wished to observe about the Web-site references in newspapers was the shelf-life of the listings. To do this we converted the data set to a monthly time series, each month represented by a single day. For this analysis, we checked 2,058 URLs, noting yes or no as to whether the link was still active and arriving at a "dead link ratio" for each month in the series. The date of publication was then compared to the date we checked the link in order to plot changes in the dead link ratio by the age of the link. All links were checked over a period of five days. Findings First, we describe the results of the simple count of stories involving the Web or the Internet. Figure 1 shows the number of stories located in the search, plotted by year with the subset that also contain specific URLs indicated. An expected growth function is demonstrated here, but we also see a hidden trend within these data. Over the five years, observed stories including URLs have gone from 2 percent of sampled stories to 25 percent. …