7.5: Retrieval

There are two forms of retrieval. In recall, we actively remember a fact, name, etc. The example of the phone buttons requires you to recall which letters go with which numbers. By contrast, in recognition, we need only recognize something when we perceive it. The Lincoln-head penny example doesn’t require you to describe or recall the face of a penny; it simply asks you to recognize the correct picture when you see it.

Although retrieval is a natural word for the elicitation of information from memory, reconstruction would often be more accurate. Retrieval is the joint effect of what is stored in the brain and of our present inferences about it. You can begin to see this if you try to remember the things you did yesterday and the order in which you did them. Yesterday’s events do not pop up in memory, one by one, in the right order. You have to do some reasoning to see what makes sense. It might go something like this:

Well, let’s see, at noon I drove to Wendy’s, but since I stopped by Homeland on the way I must have gone there before Wendy’s. Then I went to the bank. Hmmm . . . . No that can’t be right. That doesn’t make sense, since I was broke and I had to go get money from the bank to pay for my moon pie and fries. So, I guess I went to the bank between going to Homeland and Wendy’s…

The way we reconstruct items in memory is influenced—sometimes dramatically— by the context in which we remember. One way context affects memory is by providing retrieval cues. Retrieval cues are features of the situation that help us retrieve information from memory. For example, if you are trying to recall someone’s name, picturing them or recalling other information about them often helps you to remember.

Memory of an event occurs (by definition) after that event, and many things going on at the later time affect what we remember, how we remember it, and the way that we organize it into a meaningful pattern. Not only do we fill in gaps to help make sense of the earlier event; our memory of an earlier event is also colored by our attempts to make sense of the present.

Many features of a context can influence our reconstruction of the past. These include our current beliefs and attitudes, emotions and moods, expectations and set, motivations and goals (including the goals to look good and maintain self-esteem), the way questions are worded, and other peoples’ suggestions. We will now examine the ways such factors can influence our memories.

Current Attitudes and Beliefs

We tend to remember our earlier beliefs, opinions, attitudes, and even our behavior as being more like our current beliefs and attitudes than they in fact were. Greg Markus conducted a ten-year study of changes in people’s political attitudes over time. In 1973, he surveyed a group of graduating high school students, along with many of their parents. He asked them about their attitudes toward the legalization of marijuana, women’s rights, affirmative action programs, equality for women, and several other social issues. Ten years later, he asked the same people (1) what their current attitudes on these issues were, and (2), what their earlier attitudes, in 1973, had been. Both the students’ and the parents’ memories of their earlier attitudes were much closer to their current attitudes than to the attitudes they had really expressed back in 1972.

In another study, people were asked to report on their political views in 1972. Four years later, they were asked what their current views were and what their earlier views had been. Many people’s views hadn’t changed, and 96% of the people in this group (correctly) reported that their views had remained constant. But some people’s views had changed, and 91% of them (incorrectly) reported that their views had not changed.

People sometimes also remember their earlier behavior as being more in line with their current views and behavior than it really is. Linda Collins and her coworkers asked high school students about their use of tobacco and alcohol. Two and a half years later they asked them (1) what their current patterns of use were, and (2) what their earlier pattern of use, two and a half years earlier, had been. Their memories of their earlier pattern of use were closer to their current pattern than to the pattern they had reported earlier.

These results may explain why each generation of parents and teachers wonder why the current generation seems to be going to hell in a handbasket: “Why can’t today’s teenagers be more like we were when we were young?” Parents and teachers may be comparing their remembered version of their past (which is much more like their current views than their own past really was) with today’s generation, rather than comparing how things really were in the past with today’s generation. The effects discussed here are relatively modest, and people often do accurately recall their earlier views. But there is a definite tendency to see our earlier beliefs and attitudes as more like our current beliefs and attitudes than they really were. This fosters the view that our beliefs and attitudes are more stable and consistent over time than they are. This can lead us to suppose that our future beliefs and attitudes will be more like our present ones than they will turn out to be. To the extent that this happens, we have an inaccurate picture of ourselves.

Current Moods and Emotions

Cognition and emotion—thought and feeling—are more intertwined than we sometimes suppose, and our moods and emotions can affect memory. Although the evidence is cloudy, there is some evidence that people who learn material in one mood recall it more easily when they are in that mood. And studies of actual patients over a several year period showed that when people are sad or depressed they tend to remember more negative things. For example, they are more likely to remember their parents as unsupportive, rejecting, even unloving, then people who aren’t depressed. This raises the question whether people are depressed because they had a bad childhood or whether they tend to remember having a bad childhood because they are depressed (it could be a bit of both).

Context and Retrieval Cues

It is often easier to remember something if we are in the context where we experienced it. This is called context-dependent retrieval. Being back in the original context jogs the memory by providing more retrieval cues. For example, you would probably find it easier to remember names of last year’s acquaintances if you walked back through your old dormitory; it’s full of cues that would help you remember the people who lived there. Or suppose that you are in your kitchen and think of something that you need do on the way to campus. You walk into the hall and can’t remember what it was. Often going back to the kitchen helps you recall; it contains cues that help you remember what you forgot.

The importance of context shows up over and over. For example, students do better when they are tested in the room in which they learned the material. And smells are particularly powerful at evoking memories that are associated with them; they provide a cue that can awaken memories that are hard to access in other ways. It is also easier to remember something if we are in the same physiological state that we were in when we learned it. Here the context is physiological, inside our skins, and our own internal states provide a retrieval cue. This is called state dependent retrieval. For example, if you learned something after several drinks or cups of coffee, it will probably be easier to remember under those conditions.

Framing Effects: The Collision

When someone asks us to remember something, the way they word or frame their request can influence what we remember. Half the people in a group were asked, “How frequently do you have headaches?” and the other half were asked, “If you occasionally have headaches, how often?” The average response of the first group was 2.2 headaches a week, while that of the second group was 0.7 headaches a week. Similarly, it has been found that if you survey the people coming out of a movie and ask half of them, “How long was the movie?” and the other half, “How short was the movie?” those asked the first question will think the movie was longer.

In a study having more obvious real-life implications, Elizabeth Loftus and her coworkers asked subjects to watch a film of a traffic accident. Later they were asked:

• How fast were the cars going when they _____________ each other?

The blanks were filled in with different verbs for different groups of subjects. When people were asked how fast they had been going when they “smashed” each other, subjects remembered them going faster than when they were asked how fast the cars were going when they “contacted” each other. The results were:

1. smashed: 40.8/mph
2. hit: 34.0/mph
3. contacted: 30.8/mph

They were also more likely to remember seeing broken glass at the scene, even though none was present, when the collision was described in the more violent terms. Here, the experimenter’s wording affected what people remembered. If such small changes of wording can produce such dramatic effects, we must wonder what effects leading questions from a skillful lawyer, hypnotist, or therapist might have.

Schemas

We can tie some of these examples together with the notion of a schema. Our beliefs about the world in general also play a role in our construction of memories.

Consider the sentence:

• Wilbur was annoyed when he discovered he had left the mustard out of the basket.

What is the setting? Why should the mustard have been in the basket? Where is Wilbur likely to be when he discovers the mustard isn’t there? Someone from another culture might have trouble answering these questions, but you saw straightway that Wilbur has gone on a picnic and that he left the mustard out of the picnic basket. That was easy—but how did you know this?

There is now considerable evidence that we have well-organized packets of generic knowledge about many things, including picnics, graduate student offices, classrooms, visits to restaurants, and so on. These packages of information are called schemas. We won’t worry about the exact nature of schemas, which isn’t well understood in any case, but the basic idea will be useful.

Most of us have a packet of information about the typical picnic, a picnic schema. In the typical picnic, we pack food in a picnic basket, take along ketchup and mustard, eat outside, and so on. We can have picnics without any of these features, but such things are part of our picture of a typical picnic.

Schemas are very useful because they help us organize our knowledge and automatically fill in many details. A little information may activate the schema, and then we use the generic knowledge in it to quickly draw further inferences about the situation. For example, mention of a basket and mustard activate our picnic schema, and we can then use it to draw inferences about what Wilbur is up to. Similarly, your schema for a graduate student office probably includes having books in it, so it is natural to infer that it does.

Schemas enable us to form accurate expectations about a situation based on just a little information about it. These expectations may be wrong, but we will often be surprised if they are. For example, our schema of a classroom includes having a roof, and if you walked into a classroom and found no roof, you would be surprised.

Schemas figure in memory in the following way. If you remember a few fragments of experience that activate a schema, you then tend remember other things that are included in that schema. The knowledge in the schema helps you to fill in the gaps. Often this filling in is accurate. Most graduate student offices do contain books. Again, it is part of many people’s schemas of classrooms that they have fluorescent lighting. It turns out that many people think they remember that a given classroom had fluorescent lighting, even if they didn’t notice the lighting. In most cases classrooms do have such lighting, so this gap in memory would usually be filled in accurately. But we will be mistaken when we are asked about a classroom with some other sort of lighting.

Stereotypes

Not all schemas are accurate. Stereotypes are schemas, mental pictures we have of clusters of traits and characteristics that we think go together. Most of us have various racial, ethnic, and gender stereotypes. Many of these are inaccurate, and they can lead us to perceive and remember and infer things in a distorted way. For example, you may have a stereotype about the typical New Yorker that includes being rude and pushy. If so, you are more likely to predict that a given New Yorker will be pushy, more likely to interpret a New Yorker’s behavior as pushy, and more likely to remember the behavior as pushy.

In later chapters, we will explore in more detail how the stereotypes and biases involved in racism and sexism work, along with how they impede judgment and safeguards to protect ourselves from these schemas.