There are times when errors in logic seem minor and picayune – and there are times when they undermine the whole point being made. In this piece by Joe Keohane at the Boston Globe’s website, we find a case of the latter.
Keohane’s thesis is that people’s beliefs about public issues are impervious to facts, and that this is partly because people hate to admit being wrong. That’s not actually news, of course; people on both the left and the right would say that about each other within the first 45 seconds after being awakened out of a dead sleep. Keohane backs it up with information from a study, however, and that’s where the poleax is applied to the logic.
Here’s the relevant paragraph:
On its own, this might not be a problem: People ignorant of the facts could simply choose not to vote. But instead, it appears that misinformed people often have some of the strongest political opinions. A striking recent example was a study done in the year 2000, led by James Kuklinski of the University of Illinois at Urbana-Champaign. He led an influential experiment in which more than 1,000 Illinois residents were asked questions about welfare — the percentage of the federal budget spent on welfare, the number of people enrolled in the program, the percentage of enrollees who are black, and the average payout. More than half indicated that they were confident that their answers were correct — but in fact only 3 percent of the people got more than half of the questions right. Perhaps more disturbingly, the ones who were the most confident they were right were by and large the ones who knew the least about the topic. (Most of these participants expressed views that suggested a strong antiwelfare bias.)
What’s the logic problem here? It’s the implication that the pieces of information respondents were quizzed on are of primary relevance to forming a valid political opinion about welfare.
To make a case about people’s use and misuse of facts, an academic study should focus on facts that matter to the opinions in question. If a survey asked people what percentage of the federal budget is taken up by the Defense Department, respondents might know or not know, but that has very little meaning for the quality of their opinions about the topic of “national defense.” The case is similar with federal welfare. The facts relevant to people’s opinions would include such quantities as the percentage of welfare recipients who remain on welfare longer than, say, 2 years, the percentage who go on welfare and then have additional children while on welfare, and the percentage of the black population living on welfare.
My guess is that those with a “strong antiwelfare bias” would overestimate these percentages, while those with a strong pro-welfare bias would underestimate them. Readers can decide for themselves who would probably err by the greatest margin.
I wouldn’t claim that people on either side of the issue would get the answers to relevant questions “correct” with pinpoint accuracy. What I do assert is that you don’t know anything about how informed people’s opinions are, if you ask them questions that are unrelated to the criteria they use to form their opinions.
The criteria people use are, on the whole, very reasonable. Antiwelfare voters dislike welfare because they believe it creates a subculture of irresponsible dependency. They don’t measure this effect by the percentage of the federal budget that goes to welfare programs. They measure it by what they see and read and hear reported, about the lives of welfare recipients and the fates of their communities. They look at statistics on welfare recidivism, single motherhood, crime, gangs, low educational achievement, and male mortality among the urban poor. They probably couldn’t repeat precise numbers even one day after reading them in a newspaper or hearing them on the TV news – but their recollection that the statistics indicate a problem will be a well-founded part of a logical thought process. This thought process, in all its humble simplicity, has a rational integrity that contrasts starkly with the illogic of assuming that people’s opinions are “wrong” because they don’t know pieces of information irrelevant to those opinions.
It’s worth pointing out that pro-welfare voters also decide their opinions based on criteria far other than the percentage of the federal budget represented by welfare expenditures, or the percentage of recipients who are black, or the average subsidy amount. And it’s worth making the further point, obvious though it should be, that in considering these voters – both pro-welfare and antiwelfare – the rest of us would not judge the “rightness” of their opinions about welfare by how many of the irrelevant questions they got right.
Keohane’s argument highlights a deeper truth, however, if perhaps inadvertently. That truth is that what he calls “facts” would be more accurately called “elements of a narrative” – and the narrative is what it’s all about. His piece has the potential to generate a narrative that “antiwelfare voters don’t know the facts about welfare.” That theme isn’t untrue, in the strictest sense, but it also isn’t comprehensive enough to be an accurate assertion; and it is, in the context of the example he cites, decidedly irrelevant. Antiwelfare voters know enough to form their opinion based on observations – about social patterns and lifestyle – that hold up quite well in empirical study. But the Illinois Urbana-Champagne survey didn’t ask them about the criteria they actually use to make their judgments.
Unpacking narratives like this is keeping a lot of people employed these days, so we probably ought to thank the MSM for its incorrigible tendencies in this regard. I agree with Joe Keohane on what he says toward the end of his piece: that those with a media voice need to ensure a price is paid for the retailing of bad information. But a person who knows that welfare correlates strongly with out-of-wedlock birth, multigenerational dependency, low educational achievement, and crime isn’t retailing bad information – or forming his opinions without good information – if he isn’t able to recite the percentage of the federal budget that goes to welfare programs. Let’s keep these things straight.
Cross-posted at Hot Air.