Knowledge is influence: How we can measure whose ideas really matter.
by Peter Gloor and Detlef Gürtler
Who is the world’s richest person? It depends, of course, on what we measure and how we measure it, but ultimately no more than a handful of candidates are in line for this title. Who is the world’s most powerful person? An American, Russian and Chinese are certain to be in the race here.
And who is the world’s most influential person? Whose ideas can move mountains, whose mind can change the world? Who is most likely to influence our thoughts, feelings and actions? Here, the field is wide open. From Vladimir Putin to Lady Gaga, Aristotle to Warren Buffett, little Malala Yousafzai to the towering figure of Elon Musk, any one of them could have an influence on you, and some probably already have. But who has the most?
Influence: bottom-up instead of top-down
The world of ideas in the digital age is too diverse and overcrowded to simply judge from external appearances. Today, no poet or mastermind is really influential and it’s the people themselves – rather than the elites and their media – who have the final say. The influence hierarchy has turned on its head. So in today’s world it’s as much about internal perceptions.
To measure these, we can’t simply ask you. Yes, it’s true, we could ask who you see as particularly influential in your life. And why. But if we did that with everybody – through an online survey, for example – we would discover only whose fans were best organised to take part in a survey. Self-proclaimed experts from all walks of life – whether environmentalists, linguists, authors or even the pope himself – can count on loyal supporters in such cases, and there are plenty of gurus out there to do their bidding too. And then there are the trolls and hackers, who, for example, managed to select Christopher Poole, founder of hacker website 4chan and then just 21-years-old, as the world’s most influential person via the online voting poll on TIME magazine’s website.
Network beats voting
Some media have opted for a more elegant variation on the survey – they don’t ask everybody, just an expert jury. This prevents embarrassments such as the 4chan hack and generally leads to results that are plausible for the editors – after all, they are the ones who have selected the jury. The jury result, as enlightening as it may be, is thus intrinsically subjective.
At the GDI, we have chosen another way to identify the world’s most influential people – by measuring their weight in global discourse.
Anyone who affects people’s thoughts must leave traces in their communications. And for those communications that are publicly available, it’s possible to measure how central certain people are. To do this, we use the Coolhunting software supplied by Galaxyadvisors. This software not only measures whoever ‘shouts the loudest’ (i.e. has the most likes or followers), but can also calculate the strength of the relationships between the relevant people. The so called betweenness centrality measures the control a node has over the flows of a network – how often is this node on the path between other nodes? Closeness measures how easily a node can access what is available via the network – how quickly can this node reach all others in the network? Our study involved more than 600 candidates from around the globe, whose relationships we measured in three areas of the English-speaking infosphere: Wikipedia, Twitter and the internet.
Not everyone who has an influence on us can be a candidate in the search for the world’s most influential people. The GDI defines a thought leader as a living person who works primarily as a thinker, and is known and influential beyond the borders of their discipline. This definition allows for a clear delineation in most cases: Aristotle doesn’t make the cut as he’s dead, nor Lady Gaga as she’s not primarily a thinker. Active politicians are not included in the pool of candidates as their influence stems not so much from their ideas but from their actions, and the same is true of managers and entrepreneurs. Nevertheless, there are always borderline cases: what about former politicians such as Al Gore or Yanis Varoufakis? Or former businessmen such as Bill Gates? And should we include visionary entrepreneurs such as Elon Musk, whose influence stems far more from their visions than their products? Physicist Stephen Hawking is active and well known outside his own field, is the same true of economist Paul Krugman? Do architects act more as thinkers or doers? In borderline cases, every decision changes the pool of candidates and thus also the overall result.
Language barriers curb influence
Starting in 2014, we carried out influence surveys for other linguistic and cultural regions alongside the global thought leader rankings. In many parts of the world, English is perceived and used as an international language; however, for many other major languages, regional peculiarities can be identified and measured. We first added the German-speaking infosphere in 2014, the Spanish and Chinese last year, and most recently the Arabic-speaking world and Switzerland in 2016. The methodology used in the analyses is fundamentally the same, although certain details are localised; for example, in the Arabic thought leader analysis, social media receives a higher weighting than Wikipedia, which is used less frequently in this part of the world. Due to China’s strict policy of internet censorship, the Chinese analysis differentiates between influence inside and outside mainland China.
A person’s position in these rankings is known as their ‘influence rank’. Just like Google’s PageRank, this term aims to make something measurable that was previously thought to be unmeasurable: a quality – in this case, an influential quality. The question as to the quality of this measurement is asked ever more frequently of Google (a million times a day through the actions of users and search engine optimisers). But it plays a role for us too in the choice of candidates, the analysis and the feedback on the results. We have grappled with some of the criticisms that have arisen in this area for some time now.
If a jury is subjective, the same must also be true of the candidate selection in the thought leader analysis.
In principle, yes. When we first tested these methods, we realised that a subjective or even one-sided selection leads to correspondingly skewed results. Consequently, we have since relied heavily on suggestions from experts over a wide range of regions and areas of the global discourse. This greater diversity leads to more objective results, although it still can not guarantee 100% objectivity.
If it’s not online, it’s not measured.
True. Discussions are not taken into account if they occur on online platforms that are not externally measurable, such as Facebook or WhatsApp. But our subset of measurable data is the most relevant, at least in terms of intellectual dialogue. Relevant discourse that provokes no response in online media, blogs or Wikipedia is almost inconceivable in today’s world.
The selection criteria and weighting could be completely revised.
True. We try to do this every time – and then we discuss the results. For example, if under the new criteria, left-wing linguist Noam Chomsky emerged as the world’s second most influential person, we would deem the result implausible and refrain from using this weighting method again. So, just like Google’s PageRank, GDI’s influence rank is not immune from manipulation or influence – but only to a small degree, of course.