As the Election campaign draws to a close, most papers in the United States have made their Presidential endorsements for another cycle. The numbers look good for Obama, who was leading the count 160 to 59 (as at October 26), but do newspaper endorsements really matter?
Matter can be considered in different ways. For the purpose of this post we will consider matter as swinging votes behind one candidate or another. There is another interpretation, and that argues that newspaper endorsements do matter as an accurate reflection of voting intention within each state or voting district, and likewise can be used to predict the outcome of an election. Greg Mitchell at E&P thinks the answer is yes to that argument, and based on recent elections we don't disagree with him. Read his full argument here.
Endorsements and voter recognition
Tim Porter at the American Journalism Review noted in 2004 that "research on the electoral influence of newspaper endorsements is scarcer than a liberal at a Wall Street Journal editorial board meeting," and the figures haven't changed that much since. There are studies though quoted by Porter and more recently Nick Sloan that consider some parts of the puzzle. The following facts pulled from both, where not referenced.
Newspapers may reflect the voting intentions of their readers, without changing them
For the 1980 presidential election between President Jimmy Carter and Gov. Ronald Reagan, Fred Fedler, Tim Counts and Lowndes Stephens (1982) concluded that while a majority of the newspapers endorsed Reagan, many of the newspapers endorsements reflected the political behavior of the reader, which would marginalize any direct effects (p. 10). Fedler, Counts, and Stephens explain that many newspaper readers selected the papers due to their perceived political bias or ideology. In other words, papers which endorsed Reagan had more of a conservative audience while the same was true for Carter and a liberal audience (p. 11).
The effect of advertising in each newspaper
Kathleen Hall Jamieson, director of the Annenberg Public Policy Center at the University of Pennsylvania, wrote about newspaper endorsements in her 2000 book, "Everything You Think You Know About Politics and Why You're Wrong."
"The direct effect of editorials does not appear to be significant enough to find," Jamieson said in an interview. "The effect of newspaper endorsements is largely created through advertising about them that is sponsored by the candidate."
Even then, Jamieson and others interviewed for this article agree, the impact of endorsements on national or even regional elections – contests in which candidates are well-known among voters – is negligible.
Recognition of endorsements (1996)
Among readers of papers that had endorsed President Clinton, "three-quarters reported that fact; 11 percent reported their paper had endorsed Bob Dole; and 14 percent reported their paper had endorsed no one."
Among readers of papers that had endorsed Dole, "less than one-half" knew that, while one-third thought their paper had endorsed Clinton.
Change of Vote (1996)
Of those who knew their newspaper's endorsement, 1 percent said it played a "great deal" and 10 percent said it played "somewhat" of a role in their voting decision. "Of that 11 percent, about a quarter had the endorsement wrong."
Switch of votes may be neutral
Pew Center for the People & the Press study released in January (2004), which measured media influences on voters during the 2004 presidential campaign, concluded that "newspaper endorsements are also less influential than four years ago, and dissuade as many Americans as they persuade."
Long term trends
Stephen Ansolabehere, Rebecca Lessem, and James Snyder, Jr., (2006) report that newspaper endorsements shifted 1-3 percent of the vote throughout the combined presidential elections from 1940-2002 (p. 395).
Nick Sloan conclusion (early 2008)
Although the public may desire an endorsement ¬– or the newspaper believes it does – endorsements matter very little when it comes to party politics and voting behavior. While Sen. Obama, Sen. Clinton, or Sen. McCain may lobby for endorsements, neither should expect them to carry them to victory on Election Day. Voters will also have the tendency to select certain media that will not allow their own views to be shifted.
Knowledge of a candidate (and how newspapers have a bigger role when less is known)
Rocky Mountain News November 2007
"Newspaper endorsements tend to count for lower offices and offices people don't know much about," argues Sherry Bebitch Jeffe, of the University of Southern California.
After all, it's hard for voters to get in-depth information about candidates for smaller races, such as city councils. Highly motivated voters can attend debates, or watch debate replays on the local government television channels. But the vast majority of voters have little information about candidates. Accordingly, a newspaper endorsement may be able to sway a close race.
The death of newspapers, and the substitution of the Internet
What we don't have research on, as noted by Porter in 2004, is the role the internet age has had on newspaper endorsements (from what I could find). We know that newspaper circulation is declining, and that people are reading less newspapers. The decline as a percentage of adults dipped below 50% in 2006 to 48.4% today vs 80.4% in 1964, the first year readership figures per population are available from the American Association of Newspapers.
The Internet may actually offer some additional benefits to newspaper endorsements in terms of readership: endorsements are no longer restricted to geographic boundary, and a endorsement by Seattle PI can just as easily be read by a retired couple in Florida or a coal worker in the Appalachians as it may be by a store owner in the Farmers Market is Seattle.
But does readership equal influence for an endorsement? Given the low figures historically for print endorsements, I would argue no. Readers of such endorsements would more likely have a political bias reading it as well; for example the Republican audience at Drudge or HotAir reading a McCain endoresement, or Kos or The Huffington Post reading an Obama endorsement.
Cumulative endorsements driving positive press
The slew of pro-Obama endorsements, most recently from The Anchorage Daily News is clearly having a cumilative effect on national coverage. According to Google News, there are 3,103 references for "obama newspaper endorses." Google Blog Search returns 17,830 references (switching the word to endorsement gives 2,529 and 15,330 respectively). The news media (including newspapers) are covering endorsements from newspapers, creating a cumulative effect that extends a single endorsement beyond the individual newspaper into the wide scope. We don't have numbers for how many votes this might change, but the positive coverage, and likewise the negative coverage (The Anchorage Daily News coming out for Obama in Palin's home state) may matter more. Conversely they may just simply reflect the undercurrent of voting intentions anyway.
Newspaper endorsement don't matter, at least not that much, and not in a statistical way that would affect the outcome of a national election. They may still count for more in local communities where candidates are much more unknown than Presidential Candidates. Newspaper endorsements are reflective of the current mood of the people in each community, and they give a voice to those preferences, rather then ultimately changing how people may vote.