Jennifer M Bonds-raacke & John Raacke. 21st Century Psychology: A Reference Handbook. Editor: Stephen F Davis & William Buskist. Volume 1. Thousand Oaks, CA: Sage Publications, 2008.
Over two millennia have passed since Aristotle first began to reason about the properties and structure of memory. In addition to Aristotle, many other historical figures have spent time trying to decipher the complexities of memory. Plato was known to describe memory as a wax tablet to be written on. Additionally, philosophers such as Descartes and Kant portrayed simple and complex ideas as being the building blocks of thought and memory. Nevertheless, none of these great thinkers were actually capable of experimentally testing their beliefs. Only within the last century or so has the idea of memory come to be tested using scientific procedures. However, even with new advances in methodologies, at the heart of all memory research is the question that Aristotle attempted to answer many years ago: What is memory?
On the surface, this appears to be an easily answered question. Most people intuitively can provide an answer to this question. The most common response to this question when posed to introduction to psychology students is that memory is a storage space housed in the brain for remembering information. This answer is simple and noncomplex, and most people on the surface would agree that it does a pretty decent job of describing memory on some level. This becomes especially true when considering the universal definition that memory is the mental process of acquiring and retaining information for later recall. However, when one begins to ask simple questions about memory to help further define the concept, inadequacies in the aforementioned definitions become striking. For example, how does memory form? How do people store their memories? Where exactly is memory information stored? How much memory information can one store? How long can memory information be stored? How does one recall memories? Thus, although memory appears simple enough on the surface, it is far more complex than most realize.
Theory and Methods
Researchers have proposed many theories (and accompanying methodologies) to explain how information is encoded, stored, and retrieved from memory. First, individuals who can be considered pioneers in memory research will be presented. Their contributions and current status in the field will be examined. Next, sensory memory will be covered, explaining how environmental stimuli are entered and retained in memory. Atkinson and Shiffrin’s theory of short-term memory and Baddeley’s working memory will be outlined and current empirical data regarding each presented. Finally, long-term memory will be examined including (a) the taxonomy of long-term memory, (b) flashbulb memories, and (c) autobiographical memories.
Pioneers in Memory Research
Herman von Ebbinghaus in the late 1800s was the first to take a scientific, systematic approach to the study of human memory. What was even more impressive than his being the first to undertake researching this topic was the ingenuity he demonstrated in developing his methodology and statistical procedures. Unlike many of today’s researchers studying human memory, Ebbinghaus did not have a group of research participants. He also did not have a laboratory or technical research equipment. Rather, Ebbinghaus served as the sole participant on his studies of memory. One of his many inventions was the use of the meaningless consonant-vowel-consonant (CVC) triads. Ebbinghaus would learn a list of CVCs to a set criterion (e.g., one or two recitations without error), put aside the list for a period of time, and later relearn the list to the same criterion. He would then compute a savings score, which depicted the number of trials necessary for relearning. Ebbinghaus is famous for the graphical depiction of this information in the forgetting curve. Although Ebbinghaus has been criticized for his use of nonsense syllables, his pioneering work greatly shaped and continues to influence the field of memory research today.
Unlike Ebbinghaus, Sir Frederic Bartlett (1932) was interested in the study of meaningful information and memory. Bartlett’s methodology involved having participants examine the original material (e.g., stories, folktales, pictures) and recall the original material shortly after viewing and again after varying time intervals. Bartlett’s research provided early evidence for the reconstructive nature of memory. Bartlett asserted that people used their knowledge and experiences combined with elements from the original material. Although Bartlett’s work was first published during the period of behaviorism and therefore received little recognition, his contributions were recognized with the rise of cognitive psychology and have influenced many memory researchers.
Other pioneers of memory research included Binet, Semon, and Ribot. Alfred Binet is well known in psychology for his work on intelligence testing. Not surprisingly, Binet researched memory as it related to the school system. Semon described the three stages of encoding, storage, and retrieval and laid the groundwork for later researchers on the encoding specificity hypothesis. Ribot investigated problems with memory. Freud’s view of memory had a lasting (and controversial) impact on the field. And researchers in the stimulus-response framework, such as Pavlov, Thorndike, and Lashley, dominated the field until this framework was replaced by the information-processing approach.
Information from the environment is constantly available via the senses. For example, sights and sounds are all around. How information from the environment is registered and retained for later use is the topic of sensory memory. Sperling’s (1960) classic research was on the capacity and duration of visual sensory information, referred to by Neisser (1967) as iconic memory. In Sperling’s studies, he presented participants with a 3 × 4 matrix of letters (i.e., three rows with four letters in each row) for 50ms. In the first of these studies, participants were to report as many (or any) of the letters that they could remember seeing. This is called the whole report procedure or condition. Results indicated that participants could only remember about four or five letters—roughly a 37 percent accuracy rating.
Sperling’s next step was ingenious. Because participants reported seeing more letters than what they could verbally report, Sperling developed a new condition. In this condition, participants were still exposed to the matrix of letters presented for the same length of time. However, immediately following the presentation of the matrix, participants were exposed to one of three tones (i.e., high tone, medium tone, and low tone). If participants heard the high tone, they were to report the letters on the top row. If participants heard the medium tone, they were to report the letters in the middle row. And finally, if participants heard the low tone, they were to report the letters in the bottom row. Participants had no way of knowing which of the three tones they would be given and therefore had no way of knowing which line of the matrix to focus their attention on. This is called the partial report procedure or condition. Results indicated that performance increased to roughly a 76 percent accuracy rating. Sperling also varied the amount of time between the matrix presentation and the tone, which signaled to participants when to report. As expected, he found that performance decreased in the partial report condition as the amount of time delay before reporting increased.
Sperling’s original research has since been extended to investigate other issues in iconic memory. For example, Averbach and Coriell (1961) investigated the role of interference and found that the visual cue used (i.e., a circle marker) can interfere with the original perception. In addition, Sperling investigated the influence of the postexposure field on recall. He found that recall was significantly better when the visual field following the presentation of the matrix was dark rather than bright. Although Sperling’s research on iconic memory has stimulated later research and produced a wealth of literature, it is not without critics. Specifically, Haber (1983) has stated that Sperling’s studies are not ecologically valid and therefore do not reflect normal visual perception.
Sperling’s research inspired others to compare the whole report condition to the partial report condition in other areas of memory. Darwin, Turvey, and Crowder (1972) did just that to investigate the capacity and duration of auditory sensory information, referred to by Neisser (1967) as echoic memory. In their experiments, referred to as the three-eared man, participants heard a matrix of letters and digits in the space of one second. One message (comprising three letters and digits) was played to the right ear, one to the left ear, and one to both ears (serving as the third ear). Similar to Sperling’s research, where participants were visually presented with the entire matrix at one time, Darwin et al. played all three messages at the same time. In the whole report procedure or condition, participants were instructed to report as many (or any) of the letters/digits that they could remember hearing. In the partial report condition, participants were given a visual indicator to know which message to report. Results indicated a higher accuracy rating for the partial report condition. Results from both Sperling and Darwin et al. indicate that almost complete records of visual and auditory information are available for a brief period of time after exposure. This information must be encoded (e.g., paid attention to) or it will be lost quickly.
Short-Term Memory and Working Memory
Atkinson and Shiffrin (1968) proposed a stage model of memory. In their model of memory there are three stages: (a) sensory memory, (b) short-term memory, and (c) long-term memory. According to the model, sensory memory is where environmental information is registered. This information fades quickly unless it is attended to for further processing (see section above on sensory memory for more information).
The second stage is short-term memory. Short-term memory is where new information from sensory memory is transferred and where old information from long-term memory is retrieved. This model argues that short-term memory has a limited capacity and duration. Evidence for two separate memory storages for short-term memory and long-term memory includes George Miller’s (1956) classic work on the limited capacity of short-term memory (i.e., seven plus or minus two) as well as J. Brown’s (1958) and L. R. Peterson and M. J. Peterson’s (1959) reports where participants quickly forgot simple three-letter stimuli. Miller showed that short-term memory has a limited capacity by using a digit span task. Specifically, Miller repeated random number lists with participants immediately recalling the lists until they could no longer recall them. Results showed that regardless of the type of material used (i.e., digits, letters, etc.), people were not able to recall more than seven plus or minus two items. He did note that these items could consist of chunks of information in which multiple items are combined to form a single item. For example, instead of the letters I, R, and S taking up three item spaces, they would take up only one space as IRS. This process of combining information would later be called chunking.
Additionally, J. Brown and L. R. Peterson and M. J. Peterson (1959) had participants view nonsense trigrams (e.g., DCL) and then instructed participants to count backward by threes from a very large number before reporting the trigram 18 seconds later. Originally, these results were interpreted to support Atkinson and Shiffrin’s short-term memory storage by demonstrating the brief nature of short-term memory (the accuracy rate was roughly 5 percent).
The third and final stage of Atkinson and Shiffrin’s model is long-term memory. Long-term memory is considered to have an unlimited capacity and potentially permanent duration.
Atkinson and Shiffrin’s model of memory is no longer the dominant model of memory, in part due to the separate stages for short-term and long-term memory. Anderson outlines the significant issues surrounding the idea of a separate short-term memory store as including (a) effects of rehearsal, (b) coding differences, and (c) the retention function. Atkinson and Shiffrin’s model proposed that information was entered into long-term memory by being rehearsed in short-term memory. However, simply rehearsing information does not always improve long-term memory. For example, Craik and Lockhart’s (1972) depth of processing theory asserts that rehearsal improves memory only if the rehearsal is meaningful. To illustrate this point, Craik and Lockhart showed participants a word and then had participants make three types of judgments. The shallow-level judgment asked if the word was in capital letters. The intermediate-level judgment asked if the word rhymed with another word. Finally, the deep-level judgment asked if the word fit into a sentence. Results indicated that the more deeply processed words were remembered best.
Next, a separate store was argued because of evidence suggesting that short-term memory was sensory in nature and long-term memory was semantic in nature. However, later findings indicated that sensory and semantic information can be used at short and long delays. Finally, a separate store for short-term memory was argued because of the retention function. The retention function refers to results of J. Brown (1958) and L. K. Peterson and M. J. Peterson (1959). Yet when examining these results, it must be noted that the typical pattern (with rapid initial forgetting) occurred only after participants had seen many trials. In other words, the participants were experiencing interference.
Even if a separate short-term memory storage is not required to explain research findings, information must still be rehearsed to be retained in long-term memory. Baddeley (1986, 1992) asserted that a more accurate term for the place where information is rehearsed (and worked on) is working memory (a term first used by Miller, Galanter, and Pribram, 1960). The term working memory implies a place where mental effort is used and conscious effort is exerted. In Baddeley’s tripartite theory of working memory, there are three basic systems. The central executive is the main system, supporting the two slave systems: the phonological loop and the visuospatial sketchpad. Each of these components will be discussed below, including the more recently added fourth component, the episodic buffer.
The phonological loop processes phonological information and was proposed to account for research findings, in particular the classic digit span procedure. The phonological loop has two systems: a phonological store and an articulatory rehearsal system. The phonological store holds auditory information for about two seconds, and the rehearsal system cycles the information through a phonological store (much like subvocalization; Baddeley, 2001). Despite some critics (e.g., Neath & Nairne, 1995), the phonological loop is a simple account of data, and evidence in support of the loop includes research from patients with neuropsychological deficits and speech and language deficits (Baddeley, 2001). The visuospatial sketchpad processes visual and spatial information. There is neurological and imaging research to support the idea of the sketchpad containing both components. However, researching the two components separately has posed a challenge (Baddeley, 2001). The central executive system is involved in tasks such as decision making, reasoning, and comprehension. It is also involved in attention (e.g., focusing, dividing, and switching attention). The fourth component, the episodic buffer, serves as a third slave system. The episodic buffer processes information from long-term memory and aids in integrating that with current information in working memory.
One line of research in working memory that has been extensively investigated is how people search through items in memory when looking for a specific target—a working memory search. Although this seems like a bizarre task, it is one that is done by everyone on a daily basis. For example, imagine that a person is looking for a new release DVD at a movie rental store. He or she has seen what the DVD looks like on television and now has a memory for the DVD box. Upon arrival at the store, he or she searches the rows of DVDs, scanning each until the one that matches the record in memory is found, or not. This would be an example of a search in working memory. Generally, when one does a search, a mental representation of the target item is formed in working memory and then a search is made through a set of items until a match is found. This type of search task was developed by Sternberg (1966).
Sternberg (1966) was interested in determining how individuals search information in working memory. In Sternberg’s experiment, he presented subjects with digit spans up to 6 digits long. Following the 6-digit span, Sternberg would then present participants with a single digit. This single digit would represent the target digit, and participants would have to decide if the target digit existed in the previous span. For instance, assume the digit span was 3, 4, 7, 0 and the target digit was 7. In this case, the correct answer for the subject would be yes, the target was in the digit span. If the target digit had been 2, then the correct response from the subject would have been no, the target was not in the digit span.
During this task, Sternberg recorded subjects’ reaction time when it came to judging if the target was in the digit span or not. Results showed that as the digit span increased from one to six numbers, the reaction time increased as well. Sternberg proposed that participants in this task were encoding the digit span into memory, followed by encoding the target digit, and then searching through their memory to find a match. He was led to theorize that searching in working memory takes place as either a parallel or serial search. In a parallel search, the target digit is compared to all digits in the span simultaneously. This type of search predicts a flat reaction time as digit span increases. However, in a serial search, the target digit is compared to each digit in the span, one at a time in order of digit span presentation. This type of search predicts that each comparison takes a small amount of time and thereby as the digit span increases, so does reaction time. From his work, Sternberg was able to conclude that since participants’ reaction time did increase with the number of digits in the span, they must have been using a serial search.
Once Sternberg (1966) had concluded that participants were engaging in a serial search, he questioned if the serial search was exhaustive or self-terminating. In an exhaustive search, participants are searching the digit span in memory for the target digit through the whole digit span. Sternberg indicated that this would occur when the target digit was not in the digit span. For example, if the digit span consisted of 4, 2, 6, 7, 0 and the target digit was 9, then the participant would have to search the whole span in order to know that the target was not in the span. In a self-terminating search, once the target digit is found in memory, the search stops. Therefore, if the digit span was 4, 2, 6, 7, 0 and the target digit was 6, the search would terminate after reaching the third digit in the span. This would result in a far shorter search as compared to the exhaustive search. Sternberg believed that a self-terminating search would occur only when the target digit was present in the digit span. However, when comparing reaction time for a target digit that was in the digit span and for one that was not, the reaction time was the same. This indicated not only that participants were using a serial search in working memory, but also that the search was always exhaustive.
The final stage to Atkinson and Shiffrin’s (1968) model is the concept of long-term memory. The purpose of long-term memory is to organize and store information in memory. Long-term memory involves the processes of acquisition, retention, and retrieval of information. Most researchers believe that long-term memory is unlimited in its capacity to store and organize information. Additionally, most researchers believe that once that information is stored and organized in long-term memory, it will be kept permanently. Squire distinguished between two main divisions in long-term memory. Specifically, long-term memory can be divided into declarative (i.e., explicit) and nondeclarative (i.e., implicit) memory.
The first subdivision of long-term memory is declarative memory. Declarative memory is memory that requires a person to consciously recall information. Declarative memory can be divided into either semantic or episodic memories. Semantic memory is one’s memory for general knowledge and facts. For example, knowing that a car has four tires or that the first president of the United States of America was George Washington are examples of using semantic memory. On the other hand, episodic memory is a memory for episodes in one’s life—for example, knowing the date of one’s own birthday or knowing what one ate for lunch yesterday. Often, the term episodic memory is used interchangeably with the term autobiographical memory.
Autobiographical memory, the memory for events in a person’s life, has been used as a tool to explore many areas of interest in psychology (Conway & Pleydell-Pearce, 2000). For example, autobiographical memory has been used to study distressing personal events (Chapman & Underwood, 2000) and the lives of famous people (Neisser, 1981). Researchers have also investigated changes in autobiographical memory across one’s lifespan, from infantile amnesia to changes in the function of memory as people age. Studies have identified some general characteristics of autobiographical memory including the following: They are experienced as having occurred at a unique time, being experienced by the self, containing visual imagery, being highly affect laden, and being remembered as highly accurate (Leahey & Harris, 2001). Other important issues of concern, such as the stability and accuracy of the autobiographical memory, are also examined in these studies.
To address these issues of concern and to develop this area of research in human memory, researchers have developed many techniques for studying autobiographical memory. For example, one commonly used technique involves having participants keep a diary to record their daily events and later testing participants over the contents of the diary. Another technique is to have participants record an event that is occurring when they are contacted by beeper at random time intervals (Brewer, 1988).
One particularly interesting type of autobiographical memory is a flashbulb memory (Brown & Kulick, 1977). A flashbulb memory occurs when a vivid, rare, or significant personal event takes place, usually within the context of hearing about a major (typically tragic) public event. Unlike much memory research, which can be performed within a laboratory setting, research on flashbulb memory cannot. Typically, when a significant public event occurs, researchers must act quickly to take advantage of this opportunity to examine flashbulb memories. For example, Talarico and Rubin (2003) studied people’s flashbulb memory following the events of the 9/11 terrorist attacks. Specifically, the researchers tested undergraduates’ memory the day after the events and then again 1, 6, or 32 weeks later. At each session, Talarico and Rubin asked the participants to record their memories of when they first heard of the attacks as well as a recent “everyday” memory. Results from this study showed that as time passed, the consistency in their flashbulb and “everyday” memories declined. However, the researchers also found that as time passed, the participants believed that their flashbulb memories were highly accurate. Thus, the researchers concluded that although flashbulb memories are not remembered any better than “everyday” memories, a person’s confidence in them is generally much higher.
However, not all research on flashbulb memories has arrived at the same result. A study conducted in England in the early part of the 1990s produced a different result. Researchers in England studied the impact of Margaret Thatcher’s sudden resignation in 1990. Using a similar methodology as in the aforementioned studies, Conway (1995) found that participants who were more invested in the event were more consistent in their recall of their flashbulb memories over time. Therefore, the level of importance attached to the event for the individual in the study appeared to be the determining factor when evaluating consistency. In other words, people are more likely to form flashbulb memories if they view the event as especially important to them and it has an emotional effect.
Recently, researchers used autobiographical memory to investigate memories of media experiences. Hoekstra, Harris, and Helmick (1999) asked young adults to recall from their childhood a movie that had seriously frightened them. Almost all participants (98 percent) had such a memory and could describe the experience and its effects vividly, reporting sleep disturbances, specific and nonspecific fears, and preoccupation with stimuli from the film. Later research used autobiographical memory to look at memories for frightening movies seen on dates (Harris et al., 2000). Sex differences existed in the behaviors participants engaged in while viewing the movie (e.g., women reported being more jumpy than men). Another study (Harris, Hoekstra, Scott, Sanborn, & Dodds, 2004) had participants report their memories for watching romantic movies on dates. Sex differences were found in who originally chose to watch the movie, behaviors while watching the movie (e.g., men reported laughing less than women), and cognitions engaged in while watching the movie (e.g., men reported thinking about their date more than women did). Autobiographical memory has also been used as a tool to study other areas in media psychology, including televised sporting events (Bonds-Raacke & Harris, 2006), portrayals of gay and lesbian characters (Bonds-Raacke, Cady, Schlegel, Harris, & Firebaugh, in press), and memories for music. See Harris, Bonds-Raacke, and Cady (2005) for a complete review of using autobiographical memory to study media.
Over the years, memory has been researched extensively and along many different lines. Many of these lines have provided psychologists with answers to the structure of memory, how retrieval is affected, and ways to enhance one’s memory. Thus, this section will discuss models of memory, forgetting, factors affecting recall, and ways to improve memory.
Models of Memory
With the advent of the cognitive revolution, a wealth of research on memory has developed. Much of this research has evaluated how information is organized, stored, and recalled from memory. Specifically, researchers have developed models of memory to account for data generated in the lab. These models provide a systematic framework for formulating predictions, thereby supplying researchers with a way to connect research in the laboratory with events in the real world. The most popular models of memory include the SAM model (Raaijmakers & Shiffrin, 1981), the Neural Networks model, and Anderson’s (1983) ACT model.
The SAM model is a broad-based mathematical model of memory that seeks to account for results obtained using list learning paradigms. The idea here is that words have a certain memory strength or familiarity value. This strength is based on three cues: (a) the context in which the item is encoded, (b) other items to be remembered, and (c) the item itself. These three cues will determine the strength of the word in one’s memory. The latter two cues are dependent on the number of rehearsals an item receives during the encoding process. The greater the rehearsal number, the stronger the cues and the greater the memory strength will be. As with any memory task, the context in which the item is remembered is also important to recall. Therefore, the memory strength is mediated by the context in which the encoding took place. Additionally, in the SAM model, it is these three types of cues that will either facilitate or inhibit later retrieval of information. However, the SAM model has been criticized in the past for having too many assumptions, not all of which can be empirically tested (Roediger, 1993).
A Neural Network model of memory is different from the mathematical model of SAM. A Neural Network model is a computer simulation of neurons in the brain. Each neuron in the network receives input from the environment or nearby neurons and then produces an output as represented by an activation function. In a person, input information occurs through physical energy (i.e., light), whereas output information is expressed through mechanical energy (i.e., movement). In a neural network, these types of energy are represented by activation of either an excitatory or an inhibitory nature. Activation levels are generally expressed as a number from 1.0 to −1.0.
Each neuron is connected with other neurons by links that are excitatory or inhibitory. These networks of words and concepts vary in complexity from the simple to the extreme. The strength of the links between neurons in this model is determined by weights. Weights are a direct function of the neurons’ learning history. The more an item (neuron) is encoded and used, the stronger the weight. Therefore, the output of a neuron in a Neural Network model is determined by the activation of an input neuron and the weights attached to the underlying links for that given neuron. If the weights are strong, then the links will be activated and additional neurons excited or inhibited. If the weights are weak, then no information will be outputted. This theory is very popular among applied fields due to its mathematical nature and its ability to learn new information and make predictions (Massaro, 1988).
The final model of memory is Anderson’s (1983) ACT model. Unlike many models that focus solely on declarative memory, the ACT model evaluates declarative, procedural, and working memory. This model also explicitly evaluates the impact of the environment, whereas other models (i.e., SAM model) do not. The emphasis with ACT has been to model the acquisition of problem-solving skills.
Each of the three aforementioned types of memory is represented by ACT. Declarative memory provides the model with information about facts and general knowledge; this is where permanent memory resides. In ACT, the procedural memory holds production rules, which are either weakened or strengthened by their fit to environmental demands. Finally, as with previous research on working memory, this type of memory is the workhorse of ACT. Information from the environment requires information to be recalled from declarative and procedural memory. The structure in the ACT model that allows for the integration of this information (i.e., knowledge and production rules) and the changing of this integration as a result of environmental influence is working memory. Due to this flexibility, ACT has been very popular and has accounted for results in skill acquisition (Anderson, 1983), processing in working memory, and problem solving.
Typically, people think about their memory not when trying to learn information but rather when memory has failed them. When one’s memory fails, the most common reason cited by people is that they have forgotten some piece of information. Forgetting is defined as a loss of memory. The interesting question is, why does this event occur? Once information is learned, the rate of forgetting is rapid. Ebbinghaus showed this in his work with the forgetting curve. This idea that information simply fades with time is known as the decay theory.
However, time is not the only factor in forgetting. Retention of information is also affected by interfering material. McGeoch (1942) distinguished between two different types of interference. The first type of interference is called retroactive interference. Retroactive interference occurs when new information disrupts previously acquired information. For example, after moving to a new home, most people will get a new phone number. After a few years, many people find it hard to recall their previously held phone number. This is an example of retroactive interference in that the new phone number disrupts the previously held phone number. The second type of interference is proactive interference. Proactive interference occurs when previously learned information disrupts the acquisition of new information. For example, when learning information for a new psychology class, some people struggle with information from other psychology classes interfering with learning the new information. This would be an example of previously learned information disrupting the retention of newly learned information.
Additional research on interference has looked at the fan effect (Anderson, 1974). The fan effect describes a function in which reaction time on a recall test increases in proportion to the number of facts associated with a concept. This can be explained by some of the models of memory in that spreading activation can cause interference. Specifically, when a particular piece of information about a concept is to be recalled, spreading activation with that concept will determine the speed of recall. If other pieces of information have higher strength, that piece of information will cause interference and slower recall. Thus, the more facts associated with a concept, the slower the retrieval of any one of the facts.
Therefore, there appear to be two separate mechanisms that produce forgetting: decay and interference. However, many psychologists suggest that decay of information really does not occur. Rather, psychologists have suggested that all forgetting is a function of interference. Specifically, although memories appear to decay over time, it is more likely that memories are interfered with by new memories that participants have encoded.
Factors Affecting Recall
In addition to research on forgetting, there is a plethora of information evaluating other factors that affect recall from memory. One of the best-known factors that affect recall is serial position. When conducting a recall task, there are two types of recall a researcher can request. The first is known as free recall. Here, participants can recall items in a list in any order they choose. The second is known as serial recall. In this task, the participants must recall items in a list in the correct order of presentation. In this task, the participants must rehearse not only the item itself but also the position of the item in the list. This added level of complexity to the task makes it more difficult than the free recall task. As the researcher shows more and more items, the participants tend to perform more and more poorly on the task. Researchers consistently see the same curve during this task. Specifically, participants have better recall for those items at the beginning and the end of the task list and extremely poor recall for items in the middle; this pattern is known as the serial position effect or serial position curve.
Two separate events, known as serial position effects, are happening to create the serial position curve. First, participants have accurate recall for items presented at the beginning of a list. This accuracy for early information is called the primacy effect. If participants have a strong primacy effect, then they are engaging in good rehearsal over the earlier items. This good rehearsal is increasing the likelihood that information is moved to long-term memory—for example, remembering the first few items on a grocery list but not the rest. However, if participants have a weak primacy effect, then they are engaging in insufficient rehearsal of early list items and decreasing the likelihood of items moving to long-term memory.
Second, participants have extremely accurate recall for items near the end of the list. This accuracy on the final items on the list is known as the recency effect. Research has shown that most people have a strong recency effect due to the capacity of short-term memory (i.e., 7 +/−2 items). Not too surprisingly, in order to eliminate the recency effect researchers provide a distracter (similar to L. R. Peterson & M. S. Peterson, 1959) task following the last item in the list to be remembered. Glanzer and Cunitz (1966) applied a distracter task, a counting task, following a serial position task. Results indicated that participants had a weak recency effect as recall was delayed, indicating the impact of short-term memory on late item recall. However, the primacy effect was unaffected by this task, supporting the moving of earlier items into long-term memory via rehearsal.
The physical context of learning information, such as where a person was during encoding, can impact a person’s recall ability at a later point by providing cues. However, not all research has demonstrated the validity of this effect, and researchers have argued that the likelihood of this effect’s taking place depends on the degree to which participants integrate the context with their memories. There are several other types of context effects, including emotional context, mood congruence, and state-dependent memory.
Like physical context, emotional context has been shown to either improve or hinder recall. However, emotional context effects appear to occur only under special circumstances.
A stronger effect is witnessed with mood congruence. Unlike the physical context and emotional context effects, mood congruence does not occur because of the context of learning. Rather, participants are presented with positive, neutral, and negative words to learn. At the time of recall, researchers induce either a positive or a negative state. Results show participants recall more of the words that match their moods.
Related to mood congruence is the concept of state-dependent memory. State-dependent memory extends the concept of context to the internal state of the individual. Specifically, research has shown that people are better at recall when their internal state at test is the same as their internal state during encoding. Interestingly, much research on state-dependent memory has looked at the effects of alcohol and drugs on recall of information.
Although it has been discussed indirectly, much research has focused on improving people’s ability to retain and retrieve information from their memories, specifically by focusing on ways to enhance the encoding of information. Examples of this include (a) elaborative rehearsal, (b) depth of processing, and (c) the use of mnemonic devices.
Earlier in the chapter, we mentioned the most basic type of rehearsal: maintenance rehearsal. Maintenance rehearsal is a low-level repetitive recycling of information. This is the type of rehearsal one would use when remembering a phone number. However, in terms of aiding encoding and memory, once that recycling stops, there is no permanent memory. Yet, Craik and Tulving (1975) identified another type of rehearsal. This type, elaborative rehearsal, is more complex in that it uses the meaning of the information for storing and remembering. When one uses elaborative rehearsal, information is encoded more deeply and thus has a greater probability of being retained.
These two types of rehearsal were used to explain Craik and Lockhart’s (1972) depth-of-processing model. As mentioned earlier, when information is presented to be encoded, information is encoded at different levels. Some information will receive little attention and little rehearsal and therefore will not be encoded very deep. Other information will be subjected to more intentional and elaborative rehearsal and deeper, more meaningful processing. It is these types of processing that will determine the likelihood that information can be recalled at a later date.
Interestingly, another way to improve memory through improving encoding also has to do with the level of processing. Many people use mnemonics to encode information at a deeper level so that recall will improve. The term mnemonic means “helping memory.” A mnemonic device is one where strategic rehearsal is employed. The strength of mnemonics is based on the fact that information is repeatedly rehearsed, the information is built into existing frameworks, and the mnemonic device produces an enhanced ability to recall information.
One of the earliest developed and known mnemonic devices was the method of loci. This technique uses a combination of mental imagery and physical locations when encoding information. When using this method, a person will come up with a list of 10 to 15 easily remembered locations. For example, one could develop a list based on locations passed on the drive home from school or work. A person would take the first item to be remembered and mentally place that item in the location. Then, the second item to be remembered would be placed it in the second location, and so on. When the time comes for the information to be recalled, the person can just mentally take a drive and go “looking” for the items of information placed in each location. Research has shown that the distinctiveness of the mental image will aid in the recall of information.
The second most popular mnemonic device is that of the peg-word method (Miller, Galanter, & Pribram, 1960). In the peg-word method, a prememorized set of words serves as a set of mental “pegs” onto which the information to be remembered can be “hung.” The peg-word method relies on rhymes with the numbers one through ten. For example, in the peg-word method, one is a bun, two is a shoe, three is a tree, and so on. The information that is to be remembered is then “hung” on the “pegs” with the person forming a mental image of the rhyming word and the information to be remembered. For example, if the information to be remembered is a grocery list and the first item is ketchup, a person would make a mental association between “one is a bun” and ketchup. Therefore, a person could picture a bun being smeared with ketchup. Again, the distinctiveness of the mental image will aid in the recall of information.
The topic of memory has been discussed and studied for centuries. Although the topic itself is relatively old, much of the empirical research findings have come about only in the last 125 years or so. Space limitations preclude us from discussing all memory topics. However, other chapters in this Handbook present such topics as false memories and eyewitness testimony. Finally, even though this chapter and psychologists have just scratched the surface of human memory, the information gathered thus far has led to a wealth of knowledge explaining how people encode, store, and retrieve information.