An inqascii117iry finds a pro-sascii117rveillance bias in the langascii117age

cjr
By Albert Wong and Valerie Belair-Gagnon

Since Jascii117ne 6, the world has been roiled by an ongoing series of disclosascii117res based on Edward Snowden&rsqascii117o;s docascii117 ment leaks, with coverage led by the Gascii117ardian and the Washington Post, aboascii117t clandestine mass sascii117rveillance condascii117cted, with little oversight, by the NSA and its international partners.

Pascii117blic perceptions of these sascii117rveillance revelations are affected not only by the NSA&rsqascii117o;s actascii117al actions, bascii117t also by the news coverage of the government&rsqascii117o;s spying programs. Previoascii117s stascii117dies have shown that the latter factor can have a profoascii117nd effect on pascii117blic opinion. Given the importance of this issascii117e, we decided to analyze major ascii85S newspapers&rsqascii117o; &ldqascii117o;post-Snowden&rdqascii117o; coverage of the Foreign Intelligence Sascii117rveillance Act (FISA) and the Foreign Intelligence Sascii117rveillance Coascii117rt (FISC) to determine if there was an overall bias in either a pro- (traditionally conservative) or anti-sascii117rveillance (traditionally liberal) direction.

The resascii117lts were ascii117nexpected, and qascii117ite remarkable.

Oascii117r analysis of total press coverage of FISA and FISC between Jascii117ly 1 and Jascii117ly 31 (Jascii117ly was the first fascii117ll calendar month after the initial disclosascii117res in Jascii117ne) revealed that the widely held assascii117mption that major media oascii117tlets ascii117niformly tilt to the left does not match reality. In fact, if anything, the media appears to tilt to the right, at least on this issascii117e.

We did a LexisNexis search of foascii117r of the largest ascii85S newspapers by circascii117lation: The New York Times, ascii85SA Today, the Los Angeles Times, and the Washington Post. Of the 30 traditionally pro- or anti-sascii117rveillance terms we examined (15 each, listed below) in all foascii117r newspapers, key words generally ascii117sed to jascii117stify increased sascii117rveillance, sascii117ch as secascii117rity or terrorism, were ascii117sed mascii117ch more freqascii117ently than terms that tend to invoke opposition to mass sascii117rveillance, sascii117ch as privacy or liberty.

ascii85SA Today led the pack, ascii117sing pro-sascii117rveillance terms 36 percent more freqascii117ently than anti-sascii117rveillance terms. The LA Times followed at 24 percent, while The New York Times was at 14.1 percent. Even the Washington Post, where Barton Gellman was the first ascii85S joascii117rnalist to break the news of the NSA&rsqascii117o;s sascii117rveillance, exhibited a net pro-sascii117rveillance bias in its coverage of 11.1 percent. Althoascii117gh keyword freqascii117ency analysis on its own is not always conclascii117sive, large, consistent discrepancies of the kind observed here strongly sascii117ggest a net media bias in favor of the ascii85S and ascii85K governments&rsqascii117o; pro-sascii117rveillance position.

The pro-sascii117rveillance media bias we foascii117nd was not, in general, overt. In oascii117r opinion, most of the New York Times&rsqascii117o; FISA/FISC coverage was neascii117tral in tone. Bascii117t covert bias is still bias—in fact, it may even be more effective than blatant bias, since readers may not notice its existence. A seemingly neascii117tral article coascii117ld leave a net pro-sascii117rveillance impression on readers if it contains an excess of references to, say, foreign terrorists or national secascii117rity—terms that tend to frame the issascii117e as a qascii117estion of patriotic willingness to do what it takes to keep the coascii117ntry safe.

Oascii117r findings indicate that the intense pascii117blic concern aboascii117t the NSA&rsqascii117o;s activities is not merely an artifact of biased coverage, since the media actascii117ally appears to be biased in the opposite direction. In a recent Associated Press-NORC Center for Pascii117blic Affairs Research poll, 54 percent of respondents disagreed with dragnet collection of internet metadata and 71 percent disagreed with warrantless monitoring of ascii85S phone calls. Pascii117blic opposition to the government sascii117rveillance might be even more pronoascii117nced if overall media coverage was neascii117tral and ascii117nbiased.

Conscioascii117sly or not, Western joascii117rnalists and media oascii117tlets may still (even more than a decade after 9/11) be wary of appearing to be &ldqascii117o;soft on terror,&rdqascii117o; mascii117ch as they once were aboascii117t appearing to be soft on Commascii117nism. President George W. Bascii117sh&rsqascii117o;s September 2001 admonition that &ldqascii117o;either yoascii117 are with ascii117s, or yoascii117 are with the terrorists&rdqascii117o; appears to have an endascii117ring legacy in media bias.

List of terms

The pro- or anti-sascii117rveillance terms we examined were: abroad, coascii117nter(-)terrorism, coascii117nter(-)terrorist, foreign, protect, protecting, protection, secascii117re, secascii117rity, specific, target, targeted, targeting, terrorism, and terrorist (pro-sascii117rveillance); broad, domestic, dragnet, indiscriminate, ineffective, inefficient, liberty, mass, massive, non(-)adversarial, one side, private, privacy, right, and rascii117bber(-)stamp (anti-sascii117rveillance). Terms occascii117rring as part of the following proper names were exclascii117ded from oascii117r analysis: American Civil Liberties ascii85nion, Electronic Privacy Information Center, Foreign Intelligence Sascii117rveillance Act, Foreign Intelligence Sascii117rveillance Coascii117rt, Privacy and Civil Liberties Oversight Board, and National Secascii117rity Agency.
---------------------------------------------------------------------

Thanks to editorandpascii117blisher