Relationship to Web Pages, Blog Posts, and Google Scholar

Back to Table of Contents | Back to Findings

Key Points

• Web, blog, and Google Scholar volume is highly correlated with New York Times coverage, suggesting universities with high print visibility also have an elevated web discourse.
• Overall, institutional characteristics correlated with web, blog, and Google Scholar coverage mirror those associated with New York Times coverage, suggesting these attributes underlie media visibility overall, rather than being unique to the New York Times.
• Institutions with higher enrollments, larger budgets, and higher research output all have higher web coverage. Higher public service expenditures as a percent of overall budget is correlated with decreased web, blog, and Google Scholar coverage, similar to the New York Times.

Discussion

This study has focused primarily on the New York Times in order to examine a historical perspective on how coverage of higher education has changed over the past half-century. Today, however, the majority of discourse surrounding an institution occurs not in the print news media, but online. Web sites, both external and those created by an institution, combined with blogs and other social media, all contribute to the narrative encountered by those seeking information on an institution.

Unlike traditional mainstream media, it is difficult to separate those portions of the online discourse generated by an institution in the form of press releases, web sites, and other outlets under its control, from those created by the news media or private citizens. An institution with a large number of web pages mentioning its name could have strong resonance in the online sphere as external sites discuss it, or might simply generate a large amount of content for its own web sites. Most universities have a decentralized approach to web content, with a myriad domains and web outlets in addition to their primary domain name, making it very difficult to use a domain-based search to exclude institutionally-created content.

One of the benefits of mainstream media coverage is that individuals not actively seeking information about the institution will be exposed to it, increasingly its external visibility. Every reader of the New York Times will be exposed to a front page article mentioning the University of Illinois, even if they had no previous interest in the institution. On the other hand, the University's webpage will only be found by those specifically searching for the university or its research. Thus, it is unclear at this time how the ultimate impact of web coverage of an institution compares with mainstream media coverage in terms of increasingly its external visibility.

Approach

To assess the similarity of web and print coverage of research universities, searches on Google, Google Blogsearch, and Google Scholar were used to measure the total volume of web pages, blog postings, and scholarly works that mentioned each institution. The search criteria used was the same as for the New York Times, which may make these results slightly less representative, especially blog volume, in that open web content is more likely to use abbreviations and shorthand names for institutions compared with the central editorially-controlled New York Times. Also, despite the considerable size of its index, Google monitors only a portion of the entire web and blogosphere, and its result counts are only rough approximations, so size estimates based on such searches are far less accurate than newspaper counts. Google Scholar also has very uneven coverage of scholarly publications and cannot take the place of a true citation index, but offers at least a rough approximation of the scholarly publication output of each institution.

Compared with New York Times

Comparing New York Times coverage volume with online mentions, blog posts are the most highly correlated, at r=0.70, with web pages r=0.66 and Google Scholar entries r=0.60. Measuring correlation of articles about an institution, blogs are r=0.59, Google Scholar r=0.54, and web pages r=0.50.

Geography

As expected, distance to New York City has no measurable relationship with web volume, though distance to the nearest PMSA is inversely correlated at r=-0.17 for blog posts, r=-0.16 for web pages, and r=-0.14 for Scholar entries. For both traditional and electronic media, institutions closer to major cities tend to have a slightly larger volume of coverage, though the correlation is weak.

Enrollment

All three categories of electronic content are strongly correlated with enrollment. For web pages, correlations are r=0.48 for total enrollment, r=0.60 for graduate enrollment, and r=0.43 for undergraduate enrollment. Blogs are correlated at r=0.35 for total enrollment, r=0.55 graduate, and r=0.29 undergraduate. Scholar entries have correlations of r=0.40 for total enrollment, r=0.59 graduate, and r=0.33 undergraduate. Thus, higher graduate enrollments, which attract higher-caliber faculty, are strongly correlated with increased media attention in both print and electronic venues.

Budget

Detailed budgetary information is only available for public institutions, but in all categories budget is strongly correlated with electronic coverage volume. Total assets are correlated at r=0.76 for web pages, r=0.83 for blog posts, and r=0.66 for Scholar entries. Expenditures are positively correlated at r=0.81 for web pages and blog posts and r=0.70 for Scholar entries, while percent of budget spent on public service activities is negatively correlated at r=-0.11 for web pages and r=-0.20 for blog postings. As would be expected, there is no measurable correlation between Google scholar entries and public service expenditures. These correlations are very similar to those for New York Times coverage, suggesting an underlying connection between higher visibility and institutions with larger budgets focused primarily on research.

Research Output

Number of NSF grants is extremely highly correlated with web volume, at r=0.79, with r=0.78 for NSF dollar amount, and r=0.71 for patents. Blog posts have slightly lower correlations, at r=0.72 for total grants, r=0.69 dollar amount, and r=0.61 number of patents. Scholar entries are correlated at r=0.62 with grants, r=0.59 with dollar amount, and r=0.59 with patents. When examining faculty size, web pages are correlated at r=0.73, blogs at r=0.64, and Scholar entries at r=0.68.