Winning 2014: By Changing Search Engine Results to get Swing Votes

According to a study, search engine optimisation can alter voter behaviour, as in NDA’s 2014 poll victory.

SEO and Vote Swing

• A new study indicates the way search results rank may actually influence voting behavior
• With SEO algorithms can be manipulated to push positive coverage of parties’ preferred candidate
• The study suggests that a search engine company has the power to influence the results of a substantial number of elections

Did the Google effect contribute to the Modi effect during the 2014 Lok Sabha elections? While that may be unclear, a new study indicates that the way search results rank may actually influence voting behaviour, especially among undecided voters.

The research article, published in the Proceedings of the National Academy of Sciences USA, is authored by Robert Epstein and Ronald E Robertson of the Visa, California-based American Institute for Behavioral Research and Technology, which describes itself as a non-profit, non-partisan organisation.

There are plenty of takeaways from the study and its authors assert that the experiments they conducted demonstrate that

(i) biased search rankings can shift the voting preferences of undecided voters by 20% or more,

(ii) the shift can be much higher in some demographic groups, and

(iii) search ranking bias can be masked so that people show no awareness of the manipulation.

Pushing Positive Coverage

In effect, with search engine optimisation an industry in itself, you could, in theory, have hackers bomb search algorithms to push positive coverage of their preferred candidate to the top of search results, which appears to swing voters.

Of particular interest to those tracking Indian elections is that the researchers actually conducted a real-world run in India during the 2014 elections, to further study what they define as search engine manipulation effect or SEME, and its corollary, vote manipulation power or VMP.

In their India study, they listed three principal candidates – Narendra Modi, Rahul Gandhi and Arvind Kejriwal. The effort was to expose their subjects to various sets of results spewed out by a mock search engine, that were, depending on the group, heavily biased towards one candidate or the other or neutral, alternating the rants and raves.

Swinging the Swing Voters

Science magazine analysed the results thus: “But merely changing which candidate appeared higher in the results still increased the number of undecided Indian voters who would vote for that candidate by 12% or more compared with controls.”  The overall VMP that the sample delivered was close to 25%.

Not that this study is definitive by any means in the Indian context, given the complexity of the country’s politics. Also, the sample comprised of just 2,150 “undecided English-speaking voters throughout India who had not yet voted (recruited through print advertisements, online advertisements, and online subject pools).”

The demographic selected attempted diversity, drawing from 27 of India’s states and Union Territories, and across the political spectrum, with 13.3% identifying themselves as politically right, 43.8% as centrist, 26% liberal and another 16.9% “indifferent”. You could also argue that with poor internet penetration, India isn’t among the best test cases in this instance, but as the authors pointed out, over 500 million Indians will have access by 2018.

Co-author Epstein told Wired magazine that, “I thought this time we’d be lucky if we got 2 or 3%, and my gut said we’re gonna get nothing , because this is an intense, intense election environment.” However, he went on: “In some demographic groups in India we had as high as about 72% (VMP).”

Target Demographics

Among those that were highly vulnerable to persuasion of the online variety, were unemployed moderate males, male Christians, and unemployed females with no political ideology. Among those more difficult to budge were employed females with no political ideology, both unemployed liberals and those on the right, as well as female conservatives.

Among the conclusions reached by the authors is this: “Given that many elections are won by small margins, our results suggest that a search engine company has the power to influence the results of a substantial number of elections with impunity. The impact of such manipulations would be especially large in countries dominated by a single search engine company.”

They go on to remark, “Perhaps worse still, if that company left election-related search rankings to market forces, the search algorithm itself might determine the outcomes of many close elections.”

(Based in Toronto, Anirudh Bhattacharya is a columnist and author of the humorous political novel, The Candidate)