The Internet Defends Itself: The Network Neutrality Debate on the Web

Abstract

This study examines the network neutrality debate, as represented online. The research begins by conducting network analysis to identify key websites, followed by retrieving the relevant documents and using content analysis. Results demonstrate that the online version of the debate skews heavily toward the pro-network neutrality side. The web debate also includes much higher proportions of voices from nonprofit sectors, especially nongovernmental organizations. Telecommunications companies and trade groups, which anchor the anti-network neutrality coalition, are relatively quiet online. These findings show groups that are less powerful making heavy use of online communication and, in light of the political history of the issue, they also suggest online mobilizing may help reshape the dynamics of issue advocacy.

Keywords::

• Internet advocacy
• lobbying
• network neutrality
• online research methods
• policy studies
• political mobilization

ACKNOWLEDGEMENTS

Research funding for this project included a grant from the Association for Education in Journalism and Mass Communication (AEJMC), as part of the AEJMC Emerging Scholars program.

Notes

Gruenwald (2011) reports that Google spent $2.06 million in the second quarter (Q2), up from $1.48 million in the first quarter (Q1) 2011. In contrast, the Q2 expenditures for AT&T were $4.85 million, down from $6.84 million in Q1, and Verizon spent $4.38 million. The only other tech company reported to have spent similarly to Google was Microsoft ($1.85 million), and they have not been major players in the NN debate. Newer tech companies, which have been more visible NN supporters, such as Facebook ($320,000) and Netflix ($110,000), made negligible investments in lobbying.

We also coded for publication date, primarily as a check against the Web's short memory (Hellsten, Leydesdorff, and Wouters 2006). Nonetheless, the findings illustrate the rise and fall of attention to NN. Of 1,180 documents, 1,050 had clear dates, ranging from 2000 (excluding a speech by Columbia's Eli Noam—from 1991) to 2010. Of these, 38 (3.6%) were from 2005 or earlier. Most were dated 2006 (247, or 23%), 2009 (219, or 21%), or 2010 (291, or 28%); in contrast, 2007 (122, or 11.5%) and 2008 (141, or 13.3%) were quieter years. That 2006 totals far exceeded 2007 and 2008 shows that the Web's memory of the NN debate is surprisingly long. Online and off, attention to NN peaked in 2006 and again in 2009–2010. In November 2005, Ed Whitacre, then chair of SBC (now AT&T), proclaimed that SBC could and would start charging Internet companies like Google and Vonage for access to SBC customers (Whitacre 2005). This sparked an explosion of public and policymaker attention to NN (Hart, 2011). The new Democratic majority in 2007 showed little interest in major telecommunications proposals or stand-alone NN regulations. Yet Barack Obama campaigned in explicit support of NN, so his victory injected new life into the issue in 2009 and 2010, with interest waning since.

Kim et al. (2011) found the following splits between corporate voices and “other” voices (including NGOs, scholars, and ordinary citizens): As newspaper sources, 40% industry and 36% other; in FCC hearings, 35% industry and 58% other; and in Congressional hearings, 55% industry and 32% other. The Web featured 4.84 documents with nonprofit authors for every industry-authored document, a sharp contrast with the other three media. The nonprofit/industry ratio on the Web was 5.26 times that in newspapers, 2.95 times that in FCC hearings, and 8.24 times that in congressional hearings, for d-like effect sizes of .92, .60, and 1.16, respectively.

df = 2, χ² = 391.0, p < .001. Cramer's V, an r-like measure of effect size, is .54, which Cohen (1998) identifies as a large effect size. Among those documents taking a stance, 48 percent of articles were for NN and 52 percent against. This contrasts sharply with the Web, where 83 percent of documents taking a position supported NN and just 17 percent opposed, a 5-to-1 ratio. Thus, among documents that take a position, a given Web document is thus 5.3 times as likely to support NN as is a given newspaper article. This converts to a d­-like measure of effect size (Chinn 2000) of .92, which is also a large effect size (Cohen 1988). This reflects only documents that take a position, which leads to more modest measures. Just 12% of Web documents were neutral, versus 69% of newspaper articles (92 of 134 total) (Kim, Kim, and Chung 2009). Including these, 73% of Web documents (856 of 1180) supported NN, but just 15% of newspaper articles (20 of 134) did. Thus, a given Web document was 27 times more likely to support NN than a given newspaper article, for a d-like effect size of 1.83. Newspapers are better described as evenly split, though, so the more modest in-text figures are more accurate.