Some Israelis have recently expressed concern over the fact that an organization that received funding from the U.S. State Department (among others) is seeking to affect the outcome of their country’s upcoming election. But there has been far less attention paid to the potential for interference of another sort by an arguably less-shadowy American organization: Facebook.
Mark Zuckerberg’s company announced yesterday that on the coming election day, users who list their location as Israel will see an “I Voted” button encouraging them to vote. The button was first introduced during the 2010 midterm elections in the United States, and a 2012 study published in Nature found that it induced more than 300,000 people to vote who might otherwise not have. Sounds great: More votes means more democracy and we like democracy.
Well, not necessarily.
Part of what enabled Nature to do any sort of science using Facebook’s findings is that the network did not show the same button to all of its users in the same way. Instead, it showed different versions of the button, to different people, at different times of day. These variables enabled Nature to draw robust conclusions from the data, but also raised the specter that Facebook might intentionally manipulate who sees the button — and how, and when — in an effort to tilt a close election in the future. A lot of people are understandably uncomfortable with the idea that a private company might affect democratic outcomes without any of the safeguards theoretically provided by election law.
But even if Facebook committed to show exactly the same message to each and everyone of its users at precisely the same time, it might still change the outcome of an election. It should come as no surprise that the demographics of Facebook users do not reflect those of the electorate as a whole. Speaking generally, Facebookers are younger and more female (and more progressive) than the typical American voter. At the very least, it is clear that Facebook’s involvement has the potential to change the outcome of an election, whether the company intends a specific outcome or not.
Meanwhile, Israel’s four million+ Facebook users are not just younger and more female than the country’s electorate as a whole — they are also more secular (religiously) and more Jewish (ethnically). Nearly all Israelis (~90%) who use the internet are on Facebook, but a significant number of Israelis are not on the internet at all. One 2012 study found that fewer than “60 percent of Arabs and 42 percent of ultra-orthodox Jews in Israel have Internet.” To put that in perspective, the ultra-orthodox were “seven times more likely as secular Israelis to lack Internet access.” Internet access also differed by socioeconomic status, as “only 55 percent of Israelis with below-average income have Internet, compared to 85 percent of Israelis with above-average income.”
These numbers have obviously changed in the past two and a half years, and the gaps are probably smaller than they were in 2012, but I would be surprised to learn that such differences have evaporated entirely. Moreover, the impact of these differences is likely to be magnified in Israel (compared to, say, the United States). Not only is Israel much smaller, it apportions seats in the Knesset among several political parties, many of whom draw voters from quite specific constituencies. (Three Arab-Israeli parties recently combined to form a single “Joint List“, while the ultra-orthodox have famously wielded outsize power thanks to their willingness to join any coalition so long as it guaranteed their institutions funding.) While an election in the United States would need to be incredibly close for Facebook to actually swing the outcome, a difference of one or two seats in the Knesset might actually determine which party will form the government (recent polls show Likud in a virtual tie with the Zionist Union), and how much power small parties can demand in exchange for their participation in the governing coalition.
At the end of the day, Facebook is trying to do good. As noted above, high voter turnout is generally considered healthy for democracy, and efforts to get out the vote should be applauded. After all, there will always be factors that make certain classes of people more likely to vote than others, and Facebook can’t actually make someone vote, much less on behalf of any particular party. Moreover, it’s not like this is something Facebook is only doing in Israel — it rolled out a similar button during recent EU and Indian elections. But it’s still worth contemplating the fact that an ostensibly non-partisan company that exhibits many characteristics of a utility might alter the outcome of a democratic election. And at the end of the day, none of these mitigating factors mean we have to like it.