Making People Smarter
Chapter 11 documents the many errors people make in judgment, but the chapter also offers encouragement: We can take certain steps that improve people’s judgments. Some of those steps involve changes in the environment, so that we can, for example, ensure that the evidence people consider has been converted to frequencies (e.g., “4 cases out of 100”) rather than percentages (“4%”) or proportions (“.04”); this simple step, it seems, is enough on its own to make judgments more accurate and to increase the likelihood that people will consider base rates when drawing conclusions.
Other steps, in contrast, involve education. As the chapter mentions, training students in statistics seems to improve their ability to think about evidence—-including evidence that is obviously quantitative (e.g., a baseball player’s batting average or someone’s exam scores) and also evidence that is not, at first appearance, quantitative (e.g., thinking about how you should interpret a dancer’s audition or someone’s job interview). The benefits of statistics training are large, with some studies showing error rates in subsequent reasoning essentially cut in half.
The key element in statistical training, however, is probably not in the mathematics per se. It is valuable, for a number of purposes, to know the derivation of statistical equations or to know the procedures for using a statistics software package. For the improvement of everyday judgment, though, the key involves the new perspective that a statistics course encourages: This perspective helps you realize that certain observations (e.g., an audition or an interview) can be thought of as a sample of evidence, drawn from a larger pool of observations that potentially you could have made. The perspective also alerts you to the fact that a sample may not be representative of a broader population and that larger samples are more likely to be representative. For purposes of the statistics course itself, these are relatively simple points; but being alert to these points can have striking and widespread consequences in your thinking about issues separate from the topics and examples covered in the statistics class.
In fact, once we cast things in this way, it becomes clear that other forms of education can also have the same benefit. Many courses in psychology, for example, include coverage of methodological issues. These courses can also highlight the fact that a single observation is just a sample and that a small sample sometimes cannot be trusted. These courses sometimes cover topics that might reveal (and warn you against) confirmation bias or caution against the dangers of informally collected evidence. On this basis, it seems likely that other courses (and not just statistics classes) can actually improve your everyday thinking—and, in fact, several studies confirm this optimistic conclusion.
Ironically, though, courses in the “hard sciences”—chemistry and physics, for example—may not have these benefits. Obviously, these courses are immensely valuable for their own sake and will provide you with impressive and sophisticated skills. However, these courses may do little to improve your day-to-day reasoning. Why not? These courses plainly do involve a process of testing hypotheses through the collection of evidence, and then the quantitative analysis of the evidence. But, at the same time, let’s bear in mind that the data in, say, a chemistry course involve relatively homogeneous sets of observations: After all, the weight of one carbon atom is the same as the weight of other carbon atoms; the temperature at which water boils (at a particular altitude) is the same on Tuesday as it is on Thursday. As a result, issues of variability in the data are much less prominent in chemistry than they are, say, in psychology. (Compare how much people differ from each other to how much benzene molecules differ from each other.) This is, to be sure, a great strength for chemistry; it is one of the (many) reasons why chemistry has become such a sophisticated science. But this point means that chemists have to worry less than psychologists do about the variability within their sample, or, with that, whether their sample is of adequate size to compensate for the variability. One consequence of this is that chemistry courses often provide little practice in thinking about variability or sample size—issues that are, of course, crucial when confronting the (far messier) data provided by day to day life.
In the same way, cause-and-effect sequences are often much more straightforward in the “hard sciences” than they are in daily life: If a rock falls onto a surface, the impact depends simply on the mass of the rock and its velocity at the moment of collision. We don’t need to ask what mood the rock was in, whether the surface was expecting the rock, or whether the rock was acting peculiarly on this occasion because it knew we were watching its behavior. But these latter factors are the sort of concerns that do crop up in the “messy” sciences—and, of course, also crop up in daily life. So here, too, the hard sciences gain enormous power from the “clean” nature of their data but, by the same token, don’t provide practice in the skills of reasoning about these complications.
Which courses, therefore, should you take? Again, courses in chemistry and physics (and biology and mathematics) are important and teach you sophisticated methods and fascinating content. These courses will provide you with skills that you might not gain in any other setting. But, for purposes of improving your day-to--day reasoning, you probably want to seek out courses that involve a trio of traits: (a) the testing of hypotheses through (b) quantitative evaluation of (c) messy data. These courses will include many of the offerings of your Psychology Department, and probably some of the offerings in sociology, anthropology, political science, and economics. These, it seems, are the courses that may genuinely make you a better, more critical thinker about the conclusions you’re likely to weigh in your daily existence.
Critical Questions
1. fiogf49gjkf0d Why might representing data as a frequency (e.g., "4 cases out of 100") be easier for people to understand compared to a percentage (e.g., "4%") or a proportion (e.g., ".04")? |
|
2. How is the concept of "sampling" from a statistics class relevant to the way that we think about observations made about people and situations in daily life? |
|
3. fiogf49gjkf0d The previous two examples illustrate effects of data format and education. What are some other factors that encourage the use of System-2 reasoning? |
|
Submit to Gradebook:
The Doctrine of Formal Disciplines
Most universities have some sort of college-wide graduation requirements. Students are often required to take a certain number of courses in mathematics, or perhaps a foreign language. They are often required to take courses in the humanities or in laboratory science.
How should we think about these requirements? One widely endorsed view is the doctrine of formal disciplines. The idea, roughly, is that your reasoning and judgment rely on certain formal rules—akin to the rules of logic or mathematics—so you’ll think best (most clearly, more rationally) if you are well practiced in using these rules. Therefore, education should emphasize the academic disciplines that rely on these formal rules: math, logic, and, according to some people, linguistics and the study of languages. (This is presumably because the study of languages sensitizes students to the formal properties of language.)
Chapter 11, however, challenges this doctrine. There are, to be sure, excellent reasons for taking courses in math, logic, and language. But it’s a mistake to argue that these disciplines somehow strengthen the mind in a fashion that improves thinking in general. Why is this? First, the chapter makes it clear that reasoning and judgment do not rely on formal rules (i.e., rules, like those of math or logic, that depend only on the form of the argument being considered). Instead, reasoning and judgment depend heavily on the content that you’re thinking about. It’s therefore wrong to claim that formal disciplines give you training and exercise in the rules you use all the time; you do not, in fact, use these rules all the time.
Second, a number of studies have asked directly whether training in logic, or training in abstract mathematics, improves reasoning in everyday affairs. It does not. We mentioned in the chapter that training in statistics improves judgment, but (as we discussed in the previous essay) the key aspect of this training is likely to be the exercise in “translating” everyday cases into statistical terms, and not knowing the mathematical formulations themselves.
Likewise, some studies do suggest that the study of languages is associated with improved academic performance, but these studies are potentially misleading. For example, it is true that students who take Latin in high school do better on the Scholastic Aptitude Test and often do better in college. But this is probably not because taking Latin helped these students; instead, it is probably because the students who choose to take Latin are likely to be more ambitious, more academically motivated, in the first place.
How, therefore, should we design students’ education? What courses should the university recommend, and what courses should you seek out if you want to improve your critical thinking skills? Part of the answer was offered in the previous essay: Your ability to make judgments does seem to be improved by courses that rely on quantitative but somewhat messy data—disciplines such as psychology, sociology, courses at the quantitative end of anthropology, and so on. And part of the answer is suggested by another issue that arose in the chapter: We mentioned there that people often seem to rely on pragmatic, goal-oriented rules—rules involving permission or obligation, and also rules tied to the pragmatic ways you try to figure out cause-and-effect relationships (“If this broken switch is causing the problem, my car will work once I replace the switch”).
One might think, therefore, that reasoning will be improved by practice in thinking in these pragmatic terms—and evidence suggests that this is right. Short episodes of training, reminding people of the procedures needed for thinking about permission or encouraging people to think through different cause-and-effect relationships, do seem to improve reasoning. Likewise, professional training of the right sort also helps. Lawyers, for example, get lots of practice in thinking about permission and obligation; studies suggest that this training helps lawyers think through a variety of pragmatic, day-to-day problems that have little to do with their legal practice.
Notice, then, that there are two messages here. The broad message is that we can use our studies of how people do think to start generating hypotheses about how we can train people to think more effectively. The more specific message is that the data allow us to start making some recommendations, as we’ve illustrated in this essay.
Critical Questions
1. fiogf49gjkf0d In what circumstances would training in statistics improve reasoning in daily life? In what circumstances would it not benefit reasoning? |
|
2. fiogf49gjkf0d Does training about cause-and-effect relationships improve reasoning in daily life? Why or why not? |
|
3. fiogf49gjkf0d Imagine that you are a university administrator involved in the development of curriculum for students in the biology department. One of the other administrators (Administrator A) wants to include psychology and philosophy requirements, while Administrator B thinks these are unimportant to the understanding of biology. Argue for or against the inclusion of these courses in the curriculum. |
|
Submit to Gradebook: