Throughout most of the history of the study of memory, investigators focused on factors that affect accurate remembering: repetition of information, spacing of repetitions, retention interval, and types of material, retrieval cues, and dozens more. Researchers have usually assessed accurate responding on memory tests or they have measured forgetting, typically defined as omissions of response on tests as a function of delay. Researchers rarely provided systematic investigations of various errors in memory, although a few early studies examined errors (see Roediger, 1996; Schacter, 1995, for the history of memory distortion research).
It has only been relatively recently (since about 2004) that experimental psychologists began the systematic investigation of errors of memory. We may refer to memory illusions as occurring when people remember events quite differently from the way they originally happened or, in the most dramatic case, remembering events that never happened at all (Roediger, 1996). It might seem odd to consider factors that operate prior to the occurrence of an event as affecting its later retention, but in fact such prior factors can be critical.
Even if several people experience “the same” event, they will interpret it differently depending on their prior experiences. Each person perceives an event with different backgrounds and proclivities. Each of us has had different experiences and likewise has different attitudes, knowledge, dispositions, and biases. Bartlett (2002) captured these differences in background knowledge with the term schemata, which are mental structures that organize our past experiences.
Bartlett argued that memory is affected by how well (or poorly) we can encode new experiences in terms of schemata developed from our past experiences. Past experience can also enhance retention. If new information fits well with prior knowledge or schemata, retention of the information is generally better than when information does not fit. This point is aptly demonstrated in the psychological literature; people who are experts in some domain usually remember new information about that domain better than other people with less background knowledge.
In these instances, background knowledge allows us to organize and make sense of incoming information. More prosaically, memory for material that is organized meaningfully is better remembered than when the same material is presented in a scrambled and less meaningful manner. Proactive interference in memory for an event is generally weak immediately after the event is experienced, but increases over time. To return to the case of the police detective who visits many crime scenes, she might have good retention for the most recent crime immediately after she studied it.
But, if queried weeks later, the many prior crime scenes might provoke interference in her memory for the particular scene of interest. In general, the more closely the events in memory resemble one another, the more difficult it is to remember the details of one particular event without interference from the others. Background characteristics of a person can also affect how well he or she retains events. In general, young children usually remember events less well than older children or adults, whereas older adults remember more poorly than young adults.
Relative to young adults, both children and older adults are also more susceptible to interference effects. Memories from early in life, typically before the ages of 3 to 4, are often unreliable. The terms infantile amnesia or childhood amnesia refer to the notion that people generally recall little or nothing from this early period of life once they reach adulthood. However, children are clearly learning during these early years; the difficulty is consciously recollecting the information later (see Howe & Courage, 2003, for a review).
Other people showing signs of profound amnesia are those who have suffered brain injuries (strokes, closed head injuries, damage from neurosurgery to remove tumors, etc. ). Some debilitating neurological diseases (e. g. , Alzheimer’s disease) and other psychiatric illnesses (schizophrenia) can impair memory. Depressed people also generally show declines in memory. Other personality characteristics can also affect retention. For example, some evidence suggests that a measure of how “spacy” people are predicts the occurrence of false memories in several paradigms.
The measure is called the Dissociative Experiences Scale (DES), and it asks people to estimate how often they have experiences such as “listening to someone talk and they suddenly realize that they did not hear part or all of what was said, ” or “not being sure whether things that they remember happening really did happen or whether they just dreamed them” (Bernstein & Putnam, 2001). People who score high on this test tend to be more susceptible to false memories (Hyman & Billings, 1998; Winograd, Peluso, & Glover, 1998).
Encoding refers to the initial registration of information—its perception and the immediate postperceptual processing. Retention or storage refers to the maintenance of information over time, once it has been encoded. Retrieval refers to the utilization of stored information. When performance on a memory test reflects accurate responding—a person successfully remembers some past event—an inference can be made that all three stages were intact. Although this three-stage conception of the learning-memory process is logically sound, in actual practice it is problematic for two reasons (Roediger & Guynn, 1996).
First, it is difficult to separate the processes of encoding and storage. A second problem is that all three stages of the learning-memory process are intertwined and depend on each other. Encoding is an obvious prerequisite to storage, except in unusual cases. In addition, how information is encoded and stored determines what cues will be effective in its later retrieval (Tulving, 2004). It is important to note, however, that encoding and storage alone do not guarantee that information will be remembered.
Retrieval is the critical process that must occur in order to convert latent information into a conscious experience (Roediger, 2000; Tulving, 2004). Despite these caveats, we used this encoding-storage-retrieval framework for simplicity. We discuss a number of variables that can be manipulated during study (encoding), retention interval, and later test or retrieval query. Encoding Factors The distinction between perceiving and remembering is often blurry, as the two processes overlap. When does perceiving and encoding end and retention and memory begin? The answer is arbitrary.
Phenomena such as iconic and echoic memory—lingering sensory representations in the visual and auditory systems, respectively—reveal the fine line dividing perceiving and remembering. If events happen so rapidly that they cannot be accurately perceived and encoded, then they cannot be remembered later. Usually these events would not be considered forgotten because the concept of forgetting presupposes that the events were encoded and potentially could be remembered at some point in the future. However, it is not necessary for an event to be presented to the senses to be remembered.
For instance, when dealing with errors of memory, it is possible for people to remember events that did not actually occur (e. g. , Roediger & McDermott, 1995), a point that we discuss later. These events might have been internally generated by the rememberer—they could have been imagined or constructed—but were never actually experienced. One tactic lawyers sometimes use to discredit eyewitnesses of a crime is to produce a picture of the crime scene so that the judge and jury can see it, but the witness cannot.
Then the lawyer asks the witness about details of the scene, usually resulting in a poor description. This demonstration appears to show poor retention on the witness’ part, but the test is unreasonable. Although the judge and jury can pick out the details of a picture and may wonder why they are not recalled by the witness, virtually no one could pass such a test. Our minds are not capable of making complete records of events; rather, we encode features selectively. Even when we remember events reasonably well, we do not come close to remembering all the fine details that are part of the event.
We are not like video recorders, faithfully taking in and storing all the details in a scene. Bartlett (2002) relied on “the old and familiar illustration of the landscape artist, the naturalist, and the geologist who walk in the country together. The one is said to notice the beauty of scenery, the other details of flora and fauna, and the third the formation of soils and rock. In this case, no doubt, the stimuli, being selected in each instance from what is present, are different for each observer, and obviously the records made in recall are different also” (p. 4).
Many years later, Underwood (2001) captured this idea by distinguishing between the nominal and functional stimuli in learning and memory. A nominal stimulus is the complex event as it exists in the world; a functional stimulus is that part of the nominal stimulus that is coded and may (potentially) be remembered. Because the environment provides a complex array of information, coding is selective: Only some features are encoded for later retention. A false perception typically results in a false memory. But some cases can be tricky to classify. Consider an experiment reported by Roediger and McDermott (1995).
They presented adults with lists of words such as door, glass, pane, shade, ledge, sill, house, open, curtain, frame, view, breeze, sash, screen, and shutter. Participants were asked to recall the words in any order immediately after hearing them and admonished not to guess. After studying and recalling many lists, the adults took a recognition test that included both items from the study lists and items not from the lists (i. e. , lures or distracters). Their task was to differentiate between words that had been presented on the original list (studied, or “old” words) and words that were not on the list (nonstudied, or “new” words).
In addition, for words deemed “old” or studied, participants were further asked to make remember/know judgments (Tulving, 2005). For this decision, they were asked if they recognized the word because they could remember its actual occurrence on the study list or rather because they knew that the word was in the list, but could not actually remember any details about its occurrence. Meaningful material is better remembered than less meaningful information, as previously noted. Similarly, organized material is more easily remembered than disorganized material (e. g. , Marks & Miller, 2004).
Pictures and highly concrete information are better remembered than words or abstract information (e. g. , Paivio, 2001). Information presented slowly is better retained than information presented quickly (e. g. , Glanzer & Cunitz, 2002). Material that receives a person’s full attention is better retained than material that is presented under conditions of distraction (e. g. , Fisk & Schneider, 2000). Repeated information is better retained than information presented only once, and the benefit of a repetition usually increases with the amount of time between the two presentations (Melton, 2004).
Repetitions beyond two will continue to increase retention, although with diminishing returns (Challis & Sidhu, 2003). If conditions are conducive and enable people to reflect on the meaning of information when it is presented, they will remember it better than if attention is directed toward less meaningful aspects of the information (Craik & Tulving, 2005). Similarly, if people actively generate or interact with material, they generally remember better than if it is passively acquired (Jacoby, 1999; Slamecka & Graf, 1999).
For example, people remember the word cold better if they generate it from the clue “opposite of hot” than if they just read the word cold. Another powerful variable aiding retention is distinctiveness: Unusual or distinctive events are better remembered than more mundane and usual events (Hunt & McDaniel, 2003). Similarly (and perhaps for the same reason), emotional events are often easily remembered—the death of a loved one, a national crisis, or other salient and powerful events.
In fact, such events are sometimes said to create “flashbulb memories, ” a phrase referring to the idea that people often believe they remember even fine details of highly emotional events as though they had been caught in a photographic flash (Brown & Kulik, 1998). However, some studies show that the strong impression of accuracy in flashbulb memories is not entirely warranted (Neisser & Harsch, 1992), even though memories for these events are probably better than for any reasonable set of control events.
Of course, many other encoding variables affect retention besides the ones described in the preceding paragraphs. In considering factors that can lead to poor retention, simply take the opposite of these factors. For example, meaningless information is difficult to encode and harder to accurately remember. This is also the case for information that is presented rapidly, under massed presentation conditions, or processed in a superficial manner. In addition, if the event is ordinary, or does not stand out, it will be more poorly remembered than a distinctive event.