Is Science Rigged for the Rich?

A recent paper published by the Centre for Economic Policy Research, titled "Access to Opportunity in the Sciences: Evidence From the Nobel Laureates," found that 67 percent of science Nobel Prize winners have "fathers from above the 90th income percentile in their birth country." The authors, affiliated with Imperial College London, Dartmouth College, Princeton University, and the University of Pennsylvania, claim that their paper reveals extreme inequality in the science world and suggests that undiscovered geniuses from poor backgrounds never had the chance to show what they could do for humanity.

The study received considerable press attention, including a piece in The Guardian claiming that it showed "a lot of talent wasted…and breakthroughs lost."

"The Nobel prizes highlight that we have a biased system in science and little is being done to even out the playing field," wrote scientist Kate Shaw in Physics World. "We should not accept that such a tiny demographic are born 'better' at science than anyone else." 

This study contains several statistical and conceptual errors, making its findings meaningless. It provides no evidence that unequal opportunity in science limits human progress. 

For starters, how did the authors determine who was "born better" and thus had a better chance of winning a Nobel Prize? The study examined what their fathers did for a living. It found that since 1901, people with scientists for fathers had 150 times the chance of winning a science Nobel than the average person. 

Scientists earn more on average, which allegedly shows that coming from a wealthier family gave them a boost. But it's common sense that the children of scientists will have an advantage in winning Nobel Prizes. Children of successful people often excel in the same fields as their parents. The size of the advantage may seem surprising, but this is typical when you look at the extremes of the bell curve. Even small initial advantages can result in extreme differences in outcome.

Suppose instead of Nobel Prizes in science we were talking about an Olympic gold medal for the 100-meter dash. Suppose everyone in the world got to participate. There would be thousands of people a step or two behind the winner.

Now, suppose that 10 percent of the population—say, anyone with a left-handed mother—had started the race with a two-step head start. The average runner with a left-handed mother would only be two steps ahead of the pack, but we can almost guarantee that the winner would be one of them.

But the authors don't treat winning a Nobel Prize like a race, they suggest it's like winning a coin-flipping contest in which innate talent, culture, and hard work don't matter.

"If talent and opportunity were equally distributed," they write, "the average winner's parent would be at the 50th percentile."

Let's say everyone in the world participates in a coin-flipping contest to get 24 heads in a row, which is similar to winning a Nobel Prize. The one percent with scientist fathers gets two free heads, giving them a modest 8 percent advantage and 300 times the chance of winning the contest. 

The same mathematics applies to children of scientist fathers, who have 150 times the chance of winning a Nobel Prize. That could result from a modest eight percent advantage in scientific talent and opportunity. The bell curve strikes again.

So why would having a scientist father put someone 8 percentile points ahead of the pack? The study authors say it's their families' higher income or education, but those are not the first factors I would point to.

One key factor is genetics. Though we haven't identified a Nobel Prize gene, some helpful qualities for scientific accomplishment—like IQ, lack of major congenital disabilities, conscientiousness, and curiosity—are partly influenced by DNA. Another factor is culture, and having a scientist father makes it more likely you were born in an atmosphere that values science.

Of course, children of scientists are likely to have more opportunities. According to the study's authors, that's the problem we need to fix. When writing about the paper's findings on X, co-author Paul Novosad quoted Stephan Jay Gould: "I am, somehow, less interested in the weight and convolutions of Einstein's brain than in the near certainty that people of equal talent have lived and died in cotton fields and sweatshops."

The paper's authors write, "Our evidence suggests that there is a large number of 'missing scientists'—individuals who could have produced important scientific discoveries, but did not receive the complementary inputs required over their lives to do so." 

Of course, improving education and opportunities for workers in cotton fields and sweatshops is a worthwhile goal. However, the paper misunderstands how scientific discovery works. Just because the children of nonscientists aren't getting their share of Nobel Prizes doesn't mean they aren't making valuable contributions to science or other fields.

Scientific progress is based on the contributions and discoveries of thousands of people whose names we never hear. Geniuses are important, but innovation doesn't depend on one individual. We'd have Newton's laws without Isaac Newton, we'd understand radioactivity without Marie Curie, and we'd have found the Higgs boson without Peter Higgs. Literature is different: We wouldn't have Shakespeare's plays and sonnets without Shakespeare.

Redirecting all children into science to help equalize Nobel Prizes won't mean more Nobel Prizes, only perhaps different winners. It would likely mean more scientists, but perhaps more than we can fund. It could deprive the world of top contributors in other fields like literature, politics, arts, and entertainment—fields where, unlike science, top contributions cannot be duplicated by others. And it won't necessarily equalize outcomes, because children of high socioeconomic status will still have advantages over children of low socioeconomic status, whatever fields people choose.

Nobel Prize winners also aren't always the most productive scientists. Some recipients win for a single insight or a fortuitous observation. Often the winners seem to be nearly random selections from several people who published similar findings around the same time. Some Nobel Prizes were awarded for work that turned out to be wrong.

The paper did show that children of engineers, doctors, business owners, lawyers, and judges were also more likely to win Noble Prizes, although they had a smaller advantage than the children of scientists. 

Again that advantage probably had more to do with genetics, interests, and culture than family wealth. 

The paper also has another significant problem: The authors use the father's occupation to guess childhood income and education, which in turn are used to guess socioeconomic status. However, these are not perfect correlations.

The authors are applying group characteristics to individuals, which is a classic statistical error known as the ecological fallacy.

There are plenty of Nobel winners whose childhood socioeconomic status was typical of their fathers' professions. But there are also plenty who don't fit the mold.

Ada Yonath, who won the 2009 Nobel Prize in chemistry, had a father who was a business owner and rabbi, which the authors coded as the 98th education percentile. However, Yonath's father was actually an impoverished grocer who died when she was young, meaning she had to take on multiple jobs to support her family.

Harold Urey, who won the 1934 chemistry Nobel, was the son of a minister, placing him in the 98th education percentile. However, his father died when he was six, and he spent his childhood in poverty.

Linus Pauling won the 1954 Nobel Prize in chemistry. His father owned a business, and Linus was coded at the 97th wealth percentile. However, the business was a drug store, and Pauling's father got sick when he was five and died when he was nine. Pauling's practical-minded mother thought going to college was a waste of time.

The authors acknowledged this issue but claimed that the Nobel Prize winners in their study were, if anything, more likely to be born to fathers at the high socioeconomic status ranks of their fields, and therefore, the imperfect correlations strengthened their results. 

This is circular reasoning. The authors want you to start by assuming their finding is true—that socioeconomic status is a causal factor in winning science Nobel Prizes.

Good scientific inquiry doesn't start by assuming what the author is trying to prove. This bias leads researchers to make false assumptions about evidence.

It's like a detective who assumes someone is guilty and considers having an alibi as additional evidence against her. Innocent people don't need alibis.

If you don't assume family socioeconomic status is the main factor in winning science Nobel Prizes, there's no reason to think the winners' fathers had higher-than-average socioeconomic status for their fields. And therefore the errors in guessing wealth and education from profession weaken the authors' case rather than strengthen it.

The authors of this study fail to realize that their data actually show that science Nobel Prizes seem to be more meritocratic than anyone would have guessed. There is certainly more advantage to having the right parents for winning Oscars, top political offices, and sports awards. But good news doesn't make for sensational headlines or viral social media posts.

• Motion Graphics: Adani Samat
• Graphic Design: Nathalie Walker
• Audio: Ian Keyser
• Producer: Cody Huff