English 101 & 102

Evidence

  1. Definitions
    1. Evidence vs. Citation
  2. Assessing Evidence
    1. Analyzing Evidence
    2. Evaluating Evidence
      1. Sufficient
      2. Relevant
      3. Representative
        1. Representative of what?
        2. “There's no way to know”
      4. Relevant vs. Representative
  3. To Sum Up

Defintions

Argument—that is, the attempt to prove a point—is all about supporting claims with evidence. (The other part of argument is reasoning, which I cover in separate handouts for 101 and 102.)

A claim is a statement about something, which could, in theory, be supported with evidence. It is an assertion about the way things are, or were, or will be, or should be. Claims are, almost by definition, controversial, in the sense that not everyone agrees with them. That is why they require evidence.

Evidence is the concrete facts used to support a claim. Ideally, evidence is something everyone agrees on, or something that anyone could, with sufficient training and equipment, verify for themselves.

Evidence comes in many types, which I discuss in detail here.

At its most basic, evidence is something that can be perceived with the senses. A great deal of scientific evidence is observable phenomena, like the change in color when two chemicals are mixed in an experiment, or the light emitted by a distant star, or the structure of bones in a fossil. In some cases, though, what you’re observing is very removed from the claim. For example, if I say “Einstein said that space is curved” as a fact to support my claim, the observable fact is the written record of what Einstein said, not a physical phenomenon that relates directly to my claim.

There is no clear-cut line between a claim and a “fact.” Things that are facts in one context might be claims in another. For example, “the earth orbits the sun” was once a claim (i.e. an unproven or controversial assertion), but now nearly everyone accepts it as a fact. Many arguments begin by treating something as a claim, try to prove it, and then, once they have proved it to the author’s satisfaction, treat it as a fact to be used as evidence to support some new claim. However, in well-constructed arguments you can usually distinguish pretty easily between the claim and the facts used to support it. Generally speaking, successful arguments will rely on well-established facts for their evidence, at least in the beginning. If it’s hard to tell the difference between the claims and the facts, this might be one sign of a badly constructed argument, one that relies on faulty evidence or reasoning to support its claims.

Evidence vs. Citation

Don’t confuse evidence with citation. Evidence is the facts used to support the claim. Citation tells the reader where the writer got the facts. Just because a writer does not cite her or his sources, does not mean she or he has no evidence.

Assessing Evidence

Since arguments are so basic to the way we acquire knowledge, develop understanding, make decisions and shape the world, it is crucial to be able to tell good ones from bad ones—and understand what makes them that way. A huge part of what researchers do is debating each others’ arguments, by evaluating their evidence and the reasoning. This is true of chemists, English professors, historians, psychologists, oceanographers, geneticists, philosophers, anthropologists… you name it. It also applies to politicians, government agencies, and the people who want to have an influence on what they do. And, of course, it applies to lawyers, when someone is accused of a crime or a breach of contract (i.e. in trials).

Evidence is one of the two basic ingredients of argument (the other is reasoning), and so a great deal of effort goes into ensuring its quality, and challenging the quality of evidence in arguments we disagree with.

Assessing evidence can be divided into two main steps:

  1. analyzing it, and
  2. evaluating it.

I discuss each of these below.

Analyzing Evidence

Analyze is a broad term that can mean many things, but here I mean describing it neutrally in a way that shows its parts.

Analyzing evidence can itself be broken into three steps:

  1. identify the point the author is trying to prove (the claim),
  2. identify the specific facts the author gives to support the claim, and
  3. explain how the evidence is supposed to relate to the claim.

(I say “supposed to” because it doesn’t always work. If it doesn’t, we say the evidence is irrelevant, which I discuss under “Evaluating Evidence,” below. But when analyzing the evidence you should be able to tell how the author intended the evidence to support the claim, even if they were not successful. Before you can evaluate you must understand what was meant.)

Sometimes it can be tricky to tell the difference between the claim and the evidence. There is no clear-cut dividing line between them, and sometimes a claim in one context is evidence in another. But one simple rule of thumb is this:

Once you know what the claim is, you are in a better position to see what the facts are that the author has provided to support that claim.

For example, in “Time to clean house on torture” by Letta Tayler (the Guardian, 12/3/10), at the end of the second paragraph the author writes that “both the Obama and Bush administrations sought to quash criminal investigations in Europe into illegal counterterrorism activities such as kidnapping and torture by Bush-era officials.” The next three paragraphs present examples of specific acts by Bush administration officials, the investigations into those acts by Spanish and German governments, and efforts by Bush or Obama to stop those investigations. To relate this to our three steps (identify the claim, identify the evidence, explain how the evidence supports the claim):

  1. The claim is the statement at the end of the second paragraph, that the two US Presidents did this.
  2. The evidence is the list of specific, concrete facts: who did what.
  3. The evidence consists of examples of what she claims happened. (Other ways that evidence could support the claim might include numerical evidence showing how many times something had happened, statistical evidence showing its frequency, physical evidence demonstrating that an event or action had taken place, documentary evidence showing that the claim is supported by official records, and so on. See Types of Evidence for a list of different forms that evidence typically takes.)

Evaluating Evidence

Once you have identified the claim, the evidence, and their relationship, you’re on much stronger ground for evaluating the evidence.

When we evaluate evidence we go beyond simply describing what it is and how it relates to the claim. We also say whether it is good or bad. Obviously this is an essential step in evaluating the overall quality of an argument. If the evidence fails for any reason, the argument fails and the claim is not proven.

There are a lot of things we could look at in deciding if a piece of evidence is good or bad. The obvious one, and the most basic one, is truth or accuracy. Is it really true? If it’s not, of course it can’t support any claim.

A lot of academic debate is focused on proving that a piece of evidence is or is not true. For example, in the debate over global warming one key piece of evidence was the shrinking of glaciers around the world. Only when scientists had made enough observations of the glaciers over a long period of time were they able to present evidence that they were confident was true or accurate.

There are practical problems with evaluating the truth or accuracy of a piece of evidence, however. Often we don’t have the time, the resources, or the expertise to determine if a so-called “fact” is really accurate. This is even more of a problem in a three-week assignment where you’re reading about a topic you may know nothing about.

In the real world we are often forced to evaluate evidence whose accuracy we can only guess at. Fortunately, there are other criteria we can use, even when we don’t know about the accuracy.

As a practical matter, in this assignment we will not be concerned with the accuracy of the evidence. Instead, we will give the author the benefit of the doubt, and assume all of their facts are correct.

Even if the evidence is completely accurate, it may still be faulty. We will look at three other questions you can ask about the evidence to help evaluate its overall quality.

  1. Is the evidence sufficient? Is there enough of it, or do you need more to feel convinced?
  2. Is the evidence relevant? Is it really about the claim the author wants to prove, or did they go off on a tangent, providing facts that don’t have anything to do with the claim?
  3. Is the evidence representative? This is the hardest one to understand, but also the most important. Representative evidence is evidence that accurately portrays the object of study, not distorted and not just a select piece. Another word for representative is typical.

I discuss each of these questions in detail below.

Sufficient Evidence

Every fact an author provides might be accurate, and yet they might leave out crucial information needed to prove the claim: They might have insufficient evidence. One key step in evaluating evidence, therefore, is to decide if it is sufficient.

How do you know there’s enough? What makes evidence sufficient to prove a claim?

There is no simple answer to this question. To a large extent it’s a matter of individual judgement, based on what you know about a topic, the assumptions you make, your ability to draw conclusions from the facts available.

Let’s say I want to prove it’s going to rain soon. This is my claim. For evidence, I point out that it’s cloudy. This might be true, but is it enough? Probably not, since it’s often cloudy without actually raining. Suppose I add that the barometer is dropping, the temperature is dropping, wind is picking up and the air smells “damp.” Is that enough to prove it’s going to rain? For many people, that would be enough to convince them that rain is coming, or at least likely.

Generally speaking, of course, more evidence is better, and more types of evidence are better. In the first example above, there was just one piece of evidence (clouds) and one type (physical detail). In the second, there were four additional pieces of evidence (barometer, temperature, wind and smell), and two types (barometer and temperature are numerical evidence, while wind and smell are physical detail, although from different senses—touch and smell).

You can also consider the total evidence that could be included, if one were able, and ask whether the author has provided a significant portion of that. For example, say I wanted to prove that Community and Technical Colleges in Washington are suffering from budget cuts, and I give examples from 15 of the 34 colleges in the state. I’ve cited nearly half the entire group under discussion. For many, this would be enough to show that the colleges in general are suffering.

It’s usually easier to spot claims that you think do not have enough evidence than to explain why you think something does. But even when discussing insufficient evidence, try to think about what it needs: what would be “enough,” and why?

Each case is different and will require its own evaluation. Don’t expect to have the “right” answer, because the definition of “enough” is too subjective. Instead, think about whether the evidence feels like enough to you, and why. Think about how many separate facts or examples have been given, the different types of evidence, and the total amount of evidence that could be given if there was room. Aim for a convincing explanation of your reasons, not the “right” answer.

Relevant Evidence

Evidence is relevant when it has a definite relationship to the claim.

Notice that I said definite. The relationship does not have to be direct or clear, but it has to be there. Of course, a direct and clear relationship is preferable, but it’s not required.

Here, too, evidence may be 100% accurate and yet worthless, because it does not relate to the claim.

Irrelevant evidence is one of the most common problems in arguments, and is used at times by unscrupulous writers and speakers in a deliberate attempt to confuse or mislead. However it’s more commonly used by people who are confused themselves and think they see a relationship where there is none.

In testing whether evidence is relevant, don’t jump to conclusions. Just because it looks irrelevant at first glance does not mean it has no connection to the claim. Sometimes the connection is there but not direct or obvious. A good technique is to start by assuming the evidence is relevant and then try to figure out how. This trains your mind to spot connections that may be hidden. If after giving the author the benefit of the doubt in this way you still cannot see a connection, you can be more confident that the evidence really is irrelevant.

Relevance is not a binary (yes-no, either-or). Rather, it is a matter of degree. Suppose I’m shopping for a used car. I’ve found one I’m interested in and I want to know if it’s a good car. The person selling it points out that it has a really nice paint job. Is this relevant? Somewhat—after all, most people would prefer a car that looks good. But it’s not as relevant as how well the engine runs, the condition of the transmission, or the bent frame that the paint job is trying to disguise.

Relevance is probably the easiest of our three criteria to evaluate. The simplest way to do it is, if you think the evidence is relevant, to explain how it relates to the claim. If you think it is not, explain why it does not connect, or give an example that would be more relevant.

Representative Evidence

Representative comes from the word represent. Evidence represents, or gives us a picture of, the topic, and representative evidence gives us a complete and undistorted picture.

Another word for representative is typical. Representative examples are those that are typical, or most like the majority of other items in the same group.

This is different from saying the evidence is true or accurate. I can give you a true statement that nevertheless completely distorts reality. For instance, in trying to prove the age of my students, suppose I point to the oldest or the youngest person present. The facts are true—I’ve given their age correctly—but I’ve given a distorted picture, because my evidence is not representative. (This is sometimes called selective evidence or cherry picking, because you select or pick only that evidence that supports your position.) On the other hand, if I chose a more typical student, I’d be giving a better picture of the age of most of the students. (Notice that it’s not likely to tell you about all of the students, but most. Representative evidence is rarely 100% representative.)

Representative evidence is absolutely essential. One reason is simple and practical: we rarely have room to include every single example of whatever we’re talking about. Therefore, the examples we do use should be the ones most like all the others—most typical, most representative.

Let’s say I want to prove that Americans are generous, and to prove it I point to Bill Gates. He gave away billions last year! Does this prove my claim? Is Gates a typical American? No—he has tons of money to give away, and he has made that a special goal for himself. Therefore, he does not represent Americans in general. He is not the best example. A much better example would be an ordinary working person who makes an average income, because such a person is more likely to represent a typical American.

Researchers go to a lot of trouble to make sure their evidence is representative. For example, in surveys and polls, they work hard to get a random sample of people to talk to. Why? Because picking people at random means you get a typical or representative example. You don’t accidentally limit yourself to members of a certain income group, or ethnicity, or gender, or occupation, or some other category that might distort your results.

Sometimes a single example is all we need, because that one example is completely representative. For instance, if I want to demonstrate the physics of bicycles, I don’t need twenty-five bikes to illustrate my point. One typical bike will do the trick. More often, however, no one example is perfectly representative. We need at least a few to cover the ground. Even then, however, we usually have room for just a small subset of the total, so the ones we choose should be as representative as possible.

This is a major reason why statistics are such an important form of evidence. Many of the subjects we are interested in are way too big to cover with just a few examples. Go back to the question of whether Americans are generous. There are over 300 million Americans, and they are incredibly diverse in age, income, religion, ethnicity, gender, sexuality, political beliefs, region, language, and more. There is no way you could pick enough examples to represent the entire group.

What you can do, however, is give a statistical breakdown. We might say, for example, that 37% of Americans between 31 and 40 years old gave between $100 and $500 to charity in the past 12 months (Chronicle of Philanthropy, March 21, 2010). No one example, or small group of examples, could accurately represent this complex picture, or tell us how many Americans fit this category.

By definition, then, statistics—assuming, of course, that they are accurate—are designed to give a more representative picture of the world than individual examples. Not only that, but they can give you a very precise breakdown. In general you can say that if an example is representative then “most” of the subject is like it. Statistics, on the other hand, let you say much more exactly how much of the group you’re studying fits the description, as a percentage of the whole.

This is why the term “anecdotal evidence” is a form of criticism in many fields. An anecdote is a little story or example, and as such it is much less likely to be representative. One or two or even a dozen examples are no substitute for an accurate statistical measure.

Representative of what?

When deciding if a piece or a collection of evidence is representative, it is crucial to ask what it is trying to represent. “Representative of what?” should be the first question you ask.

To look at the Bill Gates example again, we might ask, is he supposed to represent Americans’ income or their attitude toward giving? He is clearly not representative when it comes to income, but he might be representative of Americans’ attitude about giving. Maybe most Americans would do the same if they had his money. My point here is not that Bill Gates is typical in this regard, but to point out that an example can be either representative or unrepresentative, depending on what you think it’s supposed to represent. That is why it is essential to ask, “representative of what?”

“There’s no way to know…”

Sometimes people say, “This is just one example, so we cannot know if it is representative of not.”

This is a mistake. Think about the bicycle example: even with one example, we can know if it’s representative. (Remember our question: Representative of what? An old-fashioned velocipede, with a front wheel diameter of four feet, would not be representative of most bicycles in use today, nor would a bicycle built for two—a tandem.) The fact that it’s just a single example does not tell us if that example is representative.

Furthermore, it is almost always a mistake to say “There is no way to know.” It might be difficult to find out, and it might be beyond our knowledge, but with sufficient research you can almost always discover if a given example is representative.

So, what do you do? How do you decide if a piece of evidence is representative? This is where you must use your judgement and your knowledge of the world. Most often, you have to make an educated guess. It’s not a total shot in the dark—it’s educated, based on what you know about the subject, or about how things work in general, or your common sense and logic—but it is a bit of a guess because you’re not an expert in the subject. (Some people like to do a little research on the side to help them decide, but that is certainly not required for the assignment.)

Deciding if evidence is representative requires that you use your knowledge and judgement about the topic and about things in general—what some call your “opinion.” You cannot figure this out based solely on what’s in the essay itself.

The Difference between “Relevant” and “Representative”

Relevant and representative are different terms and should never be used to mean the same thing:

Another way to say this is,

Go back to the example of the car and the paint job. If I ask what kind of shape the car is in, and the sales person points to the paint job, we might feel that evidence is not very relevant to the claim that it’s a good car.

On the other hand, what if the sales person points to the paint on the left front fender, which is in really good shape, and ignores the scratches and peels on the rest of the car? Then we could say that the evidence—the paint on the fender—is not representative of the entire paint job.

In the first case, the question is whether the evidence (paint job) relates to the claim (good car). In the second case, the question is whether the evidence (paint on the fender) gives an accurate picture of the thing being represented (paint on the entire car).

To Sum Up

When evaluating an argument’s evidence, your first task is to analyze it:

  1. identify the point the author is trying to prove (the claim),
  2. identify the specific facts the author gives to support the claim, and
  3. explain how the evidence is supposed to relate to the claim.

Then, you must evaluate it. In this class, that means three things:

  1. Is it sufficient?
  2. Is it relevant?
  3. Is it representative?

In each case you must give your reasons.