It's a scene we statisticians know all too well. The passenger next to us asks, "So, what do you do?" "I'm a statistician," we reply, bracing for impact. "That was the worst course I ever took," our seatmate exclaims. "I had to take it for my major, but didn't understand any of it." It's a recurrent narrative we've experienced for decades.
But times have changed. From Google chief economist Hal Varian's well-circulated 2009 quote in The New York Times that "the sexy job in the next 10 years will be statisticians" to the assertions of numerous blog posts and articles, statistics is "hot."
And the dead-on predictions of the 2012 presidential election by not only statistician Nate Silver, but also Huffington Post's Mark Blumenthal and Simon Jackson, political scientist Drew Linzer, and others, have captured the attention of the public and media about the power of statistics like never before. Now, the reactions of our seatmates have transformed from groans into eager curiosity, due in large part to this watershed event.
What is statistics? To many, statistics is the class they took in college or figures on the sports pages. But statistics is so much more. Statistics is the science of learning from data and measuring, controlling, and communicating uncertainty. Statisticians do this by developing models to describe data. These models help us design methods to collect data, draw conclusions from data, and characterize the uncertainty in the findings.
For example, polls used by Silver and others to make their election predictions were designed by statisticians to learn about the true proportion of voters in a state favoring one candidate over another. In a poll, a sample of voters -- selected to be similar in composition to the entire population of voters -- yields a sample proportion that should reflect that for all voters. But it is likely a bit different, being based on only the sample, leading to uncertainty about the true proportion. Intuitively, the larger the sample, the less the uncertainty. Statistics provides tools for designing the sampling plan, quantifying the uncertainty, and determining how large a sample is needed to control the uncertainty below an acceptable threshold. Silver and his colleagues developed statistical models to combine the results of many polls to make their predictions and characterize the uncertainty in these.
Polls are a recognizable example of statistics in action. Most people don't know that statistics also benefits science and society in numerous ways. For instance, advances in medicine depend critically on statistics. For decades, statisticians have designed and analyzed clinical trials, the gold standard studies for comparing treatments. In the pharmaceutical industry, statisticians are involved in every step of drug development -- from discovery through all phases of testing and clinical trials required for Food and Drug Administration approval.
And statisticians collaborate with medical and public health researchers on analyses of vast databases of patient information to uncover risk factors for disease, guide clinical practice, and generate new hypotheses.
Statisticians are working with medical and genomic scientists in the quest for personalized medicine, developing models of high-dimensional patient data to determine how to tailor treatment based on patient characteristics. The Environmental Protection Agency uses statistical modeling to uncover relationships between pollutants and illness and mortality and to establish regulatory standards.
Statistical models drive those purchase suggestions that pop up on Amazon and the ads that accompany Google searches. Businesses use statistics to develop marketing strategies based on Internet search and transaction data. Across the 14 federal statistical agencies, complex surveys -- including the U.S. Census -- are designed and conducted by statisticians. The results are used to determine congressional districts and allocate billions of dollars in resources for schools, health care, and transportation.
Scientific breakthroughs, such as the existence of the Higgs Boson and discovery of the accelerating expansion of the universe -- awarded the 2011 Nobel Prize in Physics -- used statistical models to establish that the findings were not simply artifacts of imprecise measurement.
Although of fundamental impact, these and myriad other contributions of statistics are widely unknown to the public, which is why more than 1,600 organizations in 112 countries are joining together to promote 2013 as the International Year of Statistics. This global campaign features informational and educational resources and videos at the Statistics2013 website, and events are taking place worldwide throughout 2013.
The timing is ideal. The world is increasingly data driven and data dependent, and the avalanche of information being generated through online searches, electronic medical records, advances in genomic science, social networks, retail transactions, and remote sensing of the environment hold great potential for improving human health and business productivity and for guiding scientific discovery and public policy decision-making. Statisticians are essential to realizing the promise of this age of Big Data, ensuring that sound decisions are made and mitigating the threat of false discoveries.
It is this data deluge that has made statistics "sexy." A 2011 McKinsey Global Institute study projects demand for 140,000 to 190,000 additional individuals with data analytic skills by 2018 in the U.S. alone, and 4 million worldwide. The opportunities for statisticians abound.
Modern applications of statistics have transformed the way it is taught. Meanwhile, interest in statistics is growing substantially among students. The number of high-school students taking the Advanced Placement Statistics exam has increased three-fold over the past decade, 7 percent in the last year alone. And the number of undergraduates majoring in statistics has risen dramatically. However, the number of graduate programs and Ph.Ds, essential partners in cutting-edge scientific and business innovation, has grown only modestly. The demand for their skills in health sciences research alone will be significant, says a National Institutes of Health advisory group.
So if your seatmate says, "I'm a statistician," ask about her work. Statistics will be more important than ever to scientists, policymakers, and you as we navigate the emerging data-rich era. As H.G. Wells purportedly said, "Statistical thinking will one day be as necessary for efficient citizenship as the ability to read and write."
The truth is stranger than fiction. Step into the world of weird news. Learn more