June 16, 2016

In the age of big data, survey research will not only survive but thrive

The next frontier of public-opinion research is already visible in the “big data” revolution. Through the digital traces of our everyday activities, we are creating a massive volume of information that can tell us a lot about ourselves. Smart data science can identify patterns in our behaviors and interests. And in some domains, such as predicting consumer spending and who will vote, algorithms may already be surpassing what surveys can do on their own.

But in the age of big data, it’s important to remember what surveys are uniquely suited to do. Asking Americans about their values, beliefs and concerns can tease out meaning from mountains of data and uncover the motivations behind the choices we make – providing a path to understanding not just what we do, but why. If history is a guide, survey research will not only survive but thrive – by taking advantage of what big data provides, and delivering what it cannot.

The survey world is unquestionably facing disruption. Cellphones are replacing the home phones we relied on for decades, and online surveys – of varying quality – are flooding the marketplace with daily numbers, leaving consumers awash in data, and rightfully skeptical. But with every challenge comes a new opportunity. Cellphones have strengthened, not weakened, a pollster’s ability to reach a balanced cross-section of the American public. Online surveys allow researchers to ask new kinds of questions, track individuals’ views over time and reach key populations of interest, all without interrupting people with long phone calls during dinner. The market for information about ourselves will continue to drive innovation and the refinement of best practices in online surveys, just as it did for other survey methodologies in the past.

The future of public opinion is more than just a technical issue. As more data about ourselves becomes available at ever-faster speeds, a broader question arises: What role do we want public opinion to play in a democracy? Critics warn of “hyper-democracy,” where the public’s raw will is so immediate and vivid that it drives the political process and empowers demagogues. A world without survey research is just as threatening; we’d be left with pundits and politicians who tell us what we want rather than listening to what we say.

On Nov. 9, a new president-elect will claim a mandate and use the election outcome as evidence of the public’s will. But elections are imperfect, offering relatively few choices in a limited time frame, and driven, in some cases, as much by personality and circumstance as by substance. Millions of Americans won’t show up to vote, and many of those who do will be unsatisfied with the choices in front of them. Big data may predict this year’s election results, but what we need – now more than ever – is a broader and deeper grasp of the public’s sentiments regarding this country’s direction. And to understand that, we still have to ask.

This article was originally printed in The Wall Street Journal on June 7, 2016.

Topics: Political Attitudes and Values, Polling, Research Methods, Telephone Survey Methods, Web Survey Methods

  1. Photo of Michael Dimock

    is the president of Pew Research Center.

1 Comment

  1. Anonymous6 months ago

    Perhaps it’s my own bias toward qualitative data, but don’t you think there’s a place for text-based responses and the “fuzzier” logic of humanity’s less quantifiable information? For instance, the Presidential candidates have very different rhetorical styles that a quantified study really couldn’t capture, and their supporters similarly voice opinions in language that one could count but might want to examine on a different level.