Why study saying social media will decide 2014 polls is wrong

In the early 1990s, when the Indian economy was opening up to the world, foreign consumer brands came tripping over themselves to sell an assortment of goods and services to “one billion” Indian customers. But after all the low hanging fruit had been plucked, they had to work hard to ferret out ‘niche’ markets that they could sell to: and one of those hitherto-unexplored markets in India, which had remained outside their reach, was the market for intimate women’s apparel.

At that time, a market research agency came out with a well-padded (and, perhaps, underwired) report that claimed – presumably after surveying women in the most remote regions – that nearly 98 per cent of women in India did not wear any kind of undergarments. It then claimed, on the basis of this titillating bit of statistic, that there was clearly a vast and unfulfilled demand  for women’s innerwear in India.

 Why study saying social media will decide 2014 polls is wrong


I cite that research only to say that in a country as complex and as populous as India, you can establish the case for a market for just about any goods or service, and back it up with surveys and studies.  In a country of one billion-plus, the social strata can be sliced and diced in myriad ways to conjure up new markets.

A  recent study on how social media platforms like Facebook can influence elections in India makes the somewhat audacious claim that   Facebook users in India today have the capacity to determine the results in upto 160 Lok Sabha constituencies - and thereby decisively impact the outcome of the 2014 elections. And that Facebook users may be the "new votebank" that Indian politicians need to worry about. (Full report of the study here.)

The study, conducted by IRIS Knowledge Foundation (more on them here) and the Internet and Mobile Association of India (more on them here), suggests that social media usage in India "is now sufficiently widespread to have the power to influence the outcome of the next elections to the Lok Sabha and, consequently, government formation." More than any other social media platform, it claims, Facebook has emerged as the "gorilla in the social media space". The study, therefore, holds serious implications - all of it positive - for Indian democracy for the way it empowers citizens, and politicians would be better served by having a well-thought-out social media strategy, it adds.

If that sounds like a bit of a tall claim in a digitally divided India, it's probably because it is.

But precisely how did the study arrive at these broad-sweep conclusions?

Consider the example that the report cites of its study of the Thane parliamentary constituency in Maharashtra. The constituency has 18 lakh registered voters, of whom some 7.5 lakh voted in 2009. The NCP candidate, Sanjeev Ganesh Naik, polled some 3.01 lakh votes, but won the seat by a relatively narrow margin of 49,000  votes ahead of  the Shiv Sena candidate Vijay Chaugule (details here).

But the study points out, citing data from Facebook, that there are some 4.19 lakh Facebook users in just this constituency: that's nearly ten times the slender victory margin of the winning candidate. "It is self-evident," the report states, "that Facebook users wield a tremendous influence over the fortunes of the candidates in this constituency, especially in a tight race."

On such shallow foundations are high-rise theories built.

Contrary to what the report claims, the only thing that's "self-evident" is that those who carried out the study have a great capacity to come to facile conclusions based on completely irrelevant data sets, and without making even a faint attempt at establishing a correlation between the raw data and their final outcome.  It's a case of Excel sheets gone wild.

The study doesn't stop there. It  rigorously carries out a similar exercise - toting up the number of Facebook users and the victory margin and voting population - for all 543 Lok Sabha constituencies in the country. It then concludes that there are some 160 'high-impact' constituencies where the number of Facebook users exceeds the victory margin in the 2009 election - or where the number of Facebook users makes up 10 percent or more of the voting population.

In the study's estimation, these are the constituencies that are ripe for a vote swing and could determine the electoral outcome of the election - based on Facebook usage.

"When one sees the wins of the Congress and the BJP in the last Lok Sabha elections, we find that 75 of the 206 seats won (36 percent) by the Congress (in 2009) and 43 of the 144 seats (30 percent) where the party finished second are counted among the 'high-impact" constituencies," it notes. Likewise, 44 of the 116 seats that the BJP won in 2009, and 50 of  the 110 seats where the party lost fall in the "high-impact' category.

"A well thought-out social media campaign will no only be about retaining previously won seats, but also about wresting new seats," the study claims.

Proceeding thus, the study identifies an additional 67 'medium-impact' constituencies, where the number of Facebook users exceeds 5 per cent of the voting population and where it assumes that every  Facebook user can influence one other voter who may not be on Facebook.  (The rest of the constituencies are categorised as being "low-impact" or "no impact" constituencies, where the charms of Facebook won't transmit into votes.)

Now, we're all very fond of our social media platforms, but this finding to my mind appears to be a hyperbole that is perhaps aimed at drumming up a niche business on Facebook ads - of the sorts that this illustration suggests. And in much the same way that the market research agency I alluded to hyped up the "women's intimate apparel" market in India, the study perhaps overstates the case by reading too much into an admittedly perceptible trend of greater embrace of social media platforms in India.

As one of the factors that had led them to carry out such a study, the report cites the impact of the social media on the US Presidential elections, particularly the most recent one last year. Citing this study in MIT Technology Review, the report claims that Obama's victory is attributed less to any articulation of a "grand vision" and rather more to "online data analytics and peer-to-peer voter targeting."

This too is a colossal oversimplication and an effort to graft an external solution into an entirely different political ecosystem. As this Firstpost report, on the voter mobilisation effort in the US last year, testifies, both the leading parties in the US do use the social media effectively, but what they're best at is in profiling virtually every single voter.

Party workers "know who the voter's grandfather and father voted for, what is the value of the house he lives in, which church he goes to, how much mortgage he has, his position on various issues, social habits - and so on. This is information built over a significant amount of time, and it helps at election time to 'micromanage' that voter."  It also makes it easier to 'personalise' campaign material for every individual voter. Each member of say, a 5-member voter family could get a different email from the same party depending on what each individual’s concerns are. "A very different world from the poster culture of Indian elections, where one slogan, like one size, fits all," the Firstpost report noted.

As this report, also on Firstpost, observed, it is true that we are seeing a greater embrace of the social media platforms my political parties in India, "come 2014, the parties will still have to win the gaddi the old-fashioned way: wooing allies, making caste and sectarian calculations, trading political favours, hustling for seats..."

In any case, as Malcolm Gladwell observed in his provocative essay in New Yorker magazine a while ago, "networks" of the sort that social media platforms bring together are systems of "weak ties". While they are great for disseminating information, they are not so effective at calling people to true action. As this commenatary noted: "While neworks are awesome at asking people to raise their hand to say they care, they don't have the offline, reality check of asking people to take a bullet."

The study that claims that Facebook users, who are probably doing nothing more political on Facebook than throwing sheep at their peers or playing Farmville games, are the new "swing voters" who will determine the electoral outcome of 2014 is guilty of statistical mubo-jumbo, and excessive hyperbole. The study is, as it itself admits, overly simplistic: the 2014 elections may see a greater embrace of social media platforms, but the battle won't be won there...

Find latest and upcoming tech gadgets online on Tech2 Gadgets. Get technology news, gadgets reviews & ratings. Popular gadgets including laptop, tablet and mobile specifications, features, prices, comparison.

Updated Date: Apr 12, 2013 14:09:13 IST