"If God exists," a pre-recorded interviewer recently asked 928 American voters, "do you approve or disapprove of its performance?" According to Public Policy Polling (.pdf), the North Carolina firm that conducted the survey, 52 percent of respondents gave God a vote of confidence, 9 percent "disapproved" of "its" performance, and 40 percent were "not sure" how they felt about rendering judgment on the Almighty by pushing a button on their telephone keypad.
For Public Policy Polling, this quirky query, planted in the midst of more conventional questions about mere mortals John Boehner and Rupert Murdoch, amongst others, functioned as an excellent publicity magnet, earning it coverage from CNN, NPR, and many other media outlets.
For the polling industry in general, however, the question played more like a scathing if subtle critique of its shortcomings. While Public Policy Polling may have crafted it as a joke, the question ended up functioning no differently than "real" survey questions do. It was asked and answered. The answers were tabulated and then presented to the public as a scientifically precise fact about an important, even profound subject. Fifty-two percent of the survey's respondents "approve" of God's performance. It feels like something substantive is being said, insight is coalescing out of nebulous, infinite data — and yet what does it actually mean?
Seventy-five years ago, when George Gallup was pioneering the polling industry, his company, the American Institute of Public Opinion, employed hundreds of "field reporters" who conducted in-person interviews with individuals who'd been carefully chosen to approximate the demography of the country as a whole. Sampling the population in this manner has been the cornerstone of the polling industry ever since — but things have changed since Gallup's heyday. America's population grew bigger and more diverse. In-person interviews shifted to phone interviews. New ways to impose our brief but emphatic opinions on strangers began to supersede the public opinion survey — first there were bumper stickers, then T-shirts with slogans, then Internet message-boards.
These days, when, say, 52 percent of a survey's 928 respondents approve of God's performance, should we really equate that with 52 percent of the American public? Or just those members of the American public who are still willing to be interrupted by robots at dinnertime to discuss their theological beliefs in a highly directed, artificially simplified fashion because they haven't yet figured out Twitter?
And what happens when contact is actually made? Certainly polls can help political donors who want to get the most bang for their money figure out which candidates have the best shot at winning. When the goal is to collect more complicated iterations of public opinion, things grow more problematic, just as they have since the polling industry's earliest days.
For example, how many people frame their thoughts about God in terms of job performance? For most people, it's likely a new idea to judge their Creator as if He were a candidate for Employee of the Month or up for re-election. Some would no doubt find the very premise nonsensical or blasphemous. Others would rightfully demand time for extended deliberation before arriving at anything they'd consider a serious answer.
In this case, Public Policy Polling is not so much measuring public opinion as it is creating it. That is, it's asking people to think about God in a way they probably haven't thought of before and then forcing them to fit their reflexive, hastily arrived-at "opinions" into three categories (approval, disapproval, and ambivalence) that are so broadly reductive they don't end up telling us much.
Indeed, while the survey recognizes that people may define God in many different ways — that's why it asks respondents if they approve or disapprove of "its" performance rather than "His" performance — it fails to address the criteria for "approval" are numerous too. One respondent may give God a thumbs-up because he believes He's doing a good job of punishing sinners and non-believers through earthquakes, floods, and famine. Another may be grateful to God for New York's recent passage of a same-sex marriage law. Every time a respondent indicated his approval of God's performance, it contributed to that scientifically precise approval rating of 52 percent. But all such arbitrary consensus really signifies is the efficiency of Public Policy Polling's auto-dialers.
Even as "opinion" polls devolve into little more than a series of keystrokes obtained by a team of patient and indefatigable robots, polls appear to be proliferating. In an age of ubiquitous spin and data overload, we find anything that seems emphatically objective and decisive, anything that convincingly simulates actionable information, dangerously seductive. And unfortunately for us, those robots can always find at least 1000 Americans who are unbusy enough to answer whatever multiple-choice question the polling industry decides to throw at them. Press "1" if you're pretty okay with the fact that public policy is being shaped in this manner. Press "2" if you think this foreshadows the eventual destruction of mankind in, say, 2027. Press "3" if you have no opinion.