Should We Follow the Science Instead of the Votes? A thought experiment in replacing elections with surveys. By Harvey C. Mansfield
The Democrats, ever solicitous of the people’s welfare and comfort, want to make voting easier. The Republicans, guardians of public morality, want to make sure votes are genuine. So why not abandon elections and replace them with surveys?
Surveys turn citizens into “respondents” answering from home by phone or computer. Respondents are scientifically selected to represent a slice of the population. Answering is easy, to please Democrats, and since your qualities and attributes are selected without regard to your name, there’s no risk of fraud, which should please Republicans. Now that we have surveys made reliable by the science of polling, why do we need elections with their hoopla, ceremony, and expense—not to mention their chanciness, rowdiness and unreason?
An objection to this question comes to mind at once. Polls often go wrong, failing to predict accurately the result of a following election. It would seem that we need elections to check on the surveys. But no—the objection takes for granted that an election is superior to a survey for reckoning the people’s will. That should be taken as a question, not an assumption. We must not underestimate the power of science. We must entertain the possibility that the survey is correct and the election—because it fails to follow the method of science—is incorrect.
This was done in 1995 by one of the founders of survey science, the late Sidney Verba, a friend and colleague of mine at Harvard. In a speech on “the citizen as respondent,” he made the claim that voter surveys are both more democratic and more accurate than elections because they reach those who don’t vote. Nonvoters are different from voters; they are less well-informed and less active on their own account, hence more vulnerable. The political scientist can reach out to them to capture their unvoiced opinions or even generously to articulate their feelings for them.
On their own, nonvoters might be content to let their fellow citizens do the voting for them, since voters seem to think voting so useful and important. But science, particularly political science, can make them active. To activate someone who on his own is passive is the work of what we call an “activist.” This is the service performed for democracy by political science and kindred intellectuals—to be more popular than the people, more democratic than untutored, uninterested democracy.
Another objection looms. How can someone else, even if supplied with the equipment and method of a scientist, know my will better than I do? The sovereignty of the people seems to rest on the sovereignty of the electorate as expressed in elections. This is what Verba believed. Despite the superior claim of science that he had the insight to glimpse, he deferred to our constitutional system of elections, however defective and incomplete. But let us look into the abyss where he stopped.
An election is a form of what survey scientists call “self-reporting.” The voter votes what he believes to be his will, giving by his vote his own report of his will. He knows better than any outside source what his own will is. That is his very human self-confidence; he knows himself. But such self-reporting is the very contrary of science. Galileo didn’t take a survey of whether the earth moved, for obviously common opinion would say it doesn’t. Science says otherwise.
This is true not only of the laws of physics but also those of medicine and psychology. A doctor wants to know your symptoms, not your diagnosis—you are too ignorant for that. He will listen to a patient for clues rather than for his views. The same with a survey scientist; he knows better than you what your will is, at least in the form of a vote. Knowing the determinants of your will, he can tell you how to vote better than you know on your own. Science challenges the popular notion that people know what they are doing.
Moreover, if one wants to object that science can’t predict individual results, so that we still need to count the votes, science would reply that the laws of statistical probability, being universal, are stronger than actual results by counting, which are variable and chancy. Note that the U.S. census uses models to get a more accurate count of population than can be had from canvassing by knocking on doors. A vote can be affected by bad weather and distracting events. One of those events might be a partisan political campaign before an election, something not done before a survey. It might seem an advantage that surveys address respondents generally in a mood of calm, unexcited by contentious accusations and warnings full of menace. They might occasionally be mildly interested by having heard the sort of advertising that argues for the qualities of Campbell’s Soup, but nothing frenetic in the way of partisan exaggeration.
Here we have found a point that gives us pause. Surveys treat voters as consumers; they don’t see anything special in voting that would require elections instead of surveys. They look for voters’ “preferences,” a usage survey science borrowed from economics. But voters don’t have preferences in the manner of preferring vanilla over chocolate. Lovers of vanilla don’t care whether chocolate is sold or not, but voters want their opinions to prevail over rival and contradicting ones. Concerning abortion, for example, a voter wants a society that thinks as he does, abhorring it or promoting it, or somewhere in between. Voters want to rule.
To rule is to join together to take responsibility for governing society. Rulers stand up for what they believe; in a democracy, the people rule by participating in elections. They don’t merely register their personal likes and dislikes; they say what they think is the common good for the community. Is it the liberal view of Democrats, focusing on including everyone equally, particularly the most vulnerable? Or the conservative view Republicans espouse, that some members of the community deserve more than others if their achievement and contribution are greater?
These differences of principle, though often not explicit or elaborated, lift elections to a level above surveys. Elections, but not surveys, are about self-government; they are actions, not responses. They decide who rules in our country. They have a greater dignity that a well-run free country, not to mention a well-considered political science, should recognize. To express that dignity, an election should be treated as a ceremony. It should be, as it used to be, an occasion when the community votes together on a single day rather than registering a preference, when convenient, that increasingly resembles online shopping.
Election Day should be the day we all vote, not a deadline for voting as it has become. It should require getting off your duff and making your way to the polls, unless you are unable or away from home. Those not voting can be urged but should not be pestered to vote. Their nonvote should be respected as if it were their vote. Quietly satisfied or modest citizens should have their say too, for they make a point the rest of us should consider. Are we really so wise as to be able to govern ourselves? And yet what better assumption can we make, perilous though it may be?
Surveys assume that we know or leave it that we “don’t know.” Elections put our wisdom to the test; they engage our assertions and our doubts together. No wonder that our parties seek advantage in how they are managed. But there is a greater truth about elections: We should be proud that we hold them. When you think about it, accusations of voter suppression and fraud show that the integrity of elections reflects our dignity as citizens. So let’s keep elections distinct from surveys. V.O. Key, a famous political scientist of the mid-20th century, said “voters are not fools.” They sort of know what they are doing. Let them rule our country in the normal, not quite scientific way of constitutional elections.
Neither party wants to abolish elections, but there’s a tendency, particularly among Democrats, to think of them as surveys and of citizens as respondents. We need to remind ourselves that the central act of our common freedom is not to “follow the science” but to run the country.
Mr. Mansfield is a professor of government at Harvard.
Comments are closed.
