Credit: MsSaraKelly
Credit: MsSaraKelly

On Privilege and Luck, or Why Success Breeds Success

Ask successful people about the secrets of their success, and you’ll probably answers like passion, hard work, skill, focus, and having great ideas. Very few people, if any, would reply with “privilege and luck”. We’re often blind to these factors and they make for less inspiring stories. But time and again, we see that the advantages that give us a head-start and the accidents that ease our path can make or break a career.

In 1968, sociologist Robert Merton noted that in several areas of science, advantage accumulates. Well-known scientists, for example, are more likely to get further recognition than equally productive peers of lesser renown. Merton called this the Matthew effect after a biblical verse that says “For unto every one that hath shall be given, and he shall have abundance: but from him that hath not shall be taken away even that which he hath.”

Merton focused on science but the Matthew effect pervades every area of our society, from bestseller lists to sports leagues. Several experiments have shown that small, random, initial advantages can spiral into huge ones. Success can breed success, and inequality breeds more inequality. Haves have more. Have-nots continue having not.

The latest example comes from Arnout van de Rijt at the State University of New York. His team went to four well-known websites and randomly distributed small bursts of success.

On Kickstarter, where people raise money for specific projects, they picked 100 out of 200 projects and donated a small proportion of their funding goal. On Epinions, where users review products and are paid based on the quality of their appraisals, they gave a “very helpful” rating to some new, unrated reviews. On Wikipedia, where dedicated editors can get status awards to honour their commitment, the team gave awards to a random subset of the most productive editors. And on Change.org, where people call for signatures to support their campaigns, the team gave a dozen signatures to 100 out of 200 early-stage campaigns.

In each case, success came in a different form: money; endorsement; social status; and expressions of support. But these small initial gains always snowballed into significant later ones.

In all four experiments, the early beneficiaries were all significantly more likely to be even more successful. For example, in the Kickstarter study, 70 percent of the projects that got a kickstart went on to attract more funding, while just 39 percent of the unchosen projects did. The lucky projects also attracted more than twice the number of later donations.

View Images
Credit: van de Rijt et al, 2014.

These effects lasted for a long time. Two weeks after the experiment, the endorsed reviews on Epinions still had more positive ratings than the others. Three months later, and the lucky Wikipedia editors still had more awards than their equally productive peers.

Van de Rijt’s study also showed that initial bits of success suffer from diminishing returns. They repeated the Kickstarter study but this time, they made either one donation or four, all at the same amounts. They found that 32 percent of the projects with no donors attracted more funding, compared to 74 percent of those with one donor, and 87 percent of those with four. So, a bit of initial success leads to more success, but lots of initial success doesn’t necessarily lead to much more success.

They found the same thing in Epinions. They could boost a reviewer’s eventual standing by giving them one very helpful rating, but giving them four such ratings didn’t do a lot more. A little snowball will careen down the slope into a big one, but a huge snowball won’t create a gigantic one. What matters is that someone kicked off a snowball at all.

View Images
Note: the experiments were followed for different times, so the x-axis is normalised. Credit: van de Rijt et al, 2014.

If the team is right, this means that it would be very hard to exploit the spiralling nature of success through brute force. As they say, “the susceptibility of reward systems to deliberate manipulation may be restricted mostly to interventions favouring those individuals who cannot muster any initial success otherwise”. That might be friends telling each other about an unknown band, or a charity offering jumpstarting loans to underappreciated projects—something to get the ball rolling.

But Duncan Watts, who studies social networks at Microsoft Research, says that these results may not generalise to bigger issues, like careers, societal trends or financial bubbles. “Clearly it’s impractical and unethical to randomly assign people to receive early career advantages, or randomly publish negative news about the housing market, so it’s going to be tricky to get experimental evidence in these systems,” he says. “That’s why the authors have chosen to study the systems they have.  But it’s important to remember that these are all rather simple and special compared to the systems that we really care about.”

Still, the results from van der Rijt’s study are clear: despite their equal merit, some projects or people came to stand above the others, simply because of a small and arbitrarily assigned advantage that they were totally blind to. In other words: privilege and luck.

A skeptic might argue that this effect is a good thing. In this study, advantages were bestowed randomly but in the real world, perhaps they are offered on merit, so that small differences in quality are gradually amplified. But we rarely get the chance to assess merit in a systematic way. No one goes through all of Kickstarter to evaluate every project they see. No one looks at every book in the store before deciding which one to buy.

That is not to say that skill, passion and hard work don’t matter. They clearly do, but studies like this tell us that we can’t assume that success is down to the qualities that emblazon motivational posters, or that people without success somehow lack these qualities.

Reference: Van der Rijt, Kang, Restivo & Patil. 2014. Field experiments of success-breeds-success dynamics. PNAS. http://dx.doi.org/10.1073/pnas.131683611