Google News may be swaying the election towards Obama according to an October 9th post by a Wharton Professor in the Huffington Post. Similar to other articles discussing the importance of social media on the election, I question the assertion that search is as powerful a tool to influence an election. Particularly with the advertising dollars that the candidates are spending to do the same.
The authors offer the following arguments to support this viewpoint:
- Google News results show far more positive results about Obama than Romney (they say search but their cited data is of Google News)
- The reason for the skew is that Google CEO Eric Schmidt wants regulatory concessions from Obama for the Android OS (this is from the footnoted article)
- Public opinion can be influenced by content presented by Google and the order that it is presented.
- Google is unbiased most of the time
- Traditional media cannot manipulate readers’ opinions to any large degree
- Google has scale that no other new source enjoys
- Manipulating search results is more insidious than a corporation donating millions of dollars into a super-PAC, because corporations are people.
Before digging into the argument, it’s probably worthwhile to debunk the conspiracy theory. Despite any relationship between Obama and Schmidt, there is a stronger relationship between political donations and legislation. Since Republicans support deregulation, Schmidt should be able to achieve the outcome he wants agnostic of who is President. Schmidt would be very naive not to understand the shelf-life of political candidates.
Also, to skew search in that way would take a coordinated effort. There are enough Republican-leaning employees at Google (based upon this year’s political donations) that one imagines it would be very hard to make such a coordinated effort to consistently skew Google News results in favor of Obama.
Is Google News influencing the election?
In a word, maybe. But probably not to the extent that the authors suggest. We live with a unique kind of information poverty: there is so much out there to consume that we need a vehicle to bring order to it. Google and Bing do that. But how do they? We have a general idea, but their algorithms are proprietary and constantly in flux. SEO exists as a discipline to convey a deeper understanding of these secret rules, an acknowledgement of search as a sort of game.
Mitt Romney is a horrible political candidate. Bruce Willis called him the “Dash Riprock of the Republican Party,” and save one debate he has lived up to that billing. To say that Google News is unfairly skewed to favor Obama is to ignore the fact that until last week, Obama was running a much better campaign. From what I can tell the narrative on Google has since shifted this week, which seems to dispute the bias allegation.
Let’s take a look at the arguments that were cited as evidence of the Google News bias towards Obama:
Do Google News results show far more positive results about Obama than Romney? Not this week, which seems to indicate that the content presented is influenced more by the campaigns than by a skewing of search.
Can public opinion be influenced by content presented by Google and the order that it is presented. I don’t think this is disputable. But Google organic results aren’t the only results on any given SERP. Both candidates invest heavily in digital advertising through Google, with Romney’s digital strategist pointing out that he believes his team is more adept than Obama in this area. I suspect that the Romney team probably doesn’t buy into the Google conspiracy either, or they would probably make a bigger deal of it.
Is Google unbiased most of the time. Absolutely not. However, they most likely are not specifically biased towards any given search result because of the resources and secrecy that would be required to do so. Their bias is a by-product of their algorithm.
Can traditional media manipulate readers’ opinions to any large degree? Probably. If you dispute traditional publisher’s influence on readers then the argument that Google News is distorted is irrelevant as well. Google News presents their stories.
Does Google have scale that no other new source enjoys? By the numbers that’s accurate, although it’s unclear how many people use Google News as their news source. Reuters published a report suggesting that less than 10% of people get their news from Google News, so the extent of their influence is comparatively minute. Additionally, Project New America published a report detailing that the largest undecided voting demographic are caucasian women 18-49. This is a demographic that is less likely to get their news online than their male counterparts. So it appears that the people that the candidates most need to influence are unlikely to be influenced by Google News.
Is manipulating search results more insidious than a corporation donating millions of dollars into a super-PAC, because corporations are people? I don’t know how to respond to that except that by that definition Google is a person and probably shouldn’t be held to any higher standard than any other corporate personage.
I don’t think anyone disputes that Google has a unique capability to influence us by how they present us information. By definition it will not be fair, because fair is highly subjective.
That said, the argument that Google News is giving Obama a pass is over-the-top. The biggest evidence to support the fact that Google News cannot influence an election would be that Romney would win if the election was held today.
There may be instances where Google is able to wield more influence than others, but the diaspora of online media properties that shape collective opinion is vast. And the audience that will determine the outcome of this election is probably not using Google to inform their decision.
Post-script: It turns out that Romney’s picture shows up in Google Images if you search for “Completely Wrong.” The reason? When Romney apologized for his 47% remark he called that assertion “completely wrong.” So it’s important to consider the context of what Google is indexing before jumping to conclusions about causation. Anyone remember “Santorum?”