3

Basic and Applied Memory Research: Empirical, Theoretical, and Metatheoretical Issues

David G. Payne

Binghamton University

Frederick G. Conrad

Bureau of Labor Statistics

Danny R. Hager

Binghamton University

The first two chapters in this volume offer very interesting and in-depth analyses of the relation between basic and applied memory research. In this chapter we add to this discussion, but before doing so we draw an analogy between the views of Intons-Peterson (chapter 1) and Herrmann and Raybeck (chapter 2) and a classic study in social psychology (Hastorf & Cantril, 1954). Hastorf and Cantril’s study has important implications for our discussion of empirical, theoretical, and metatheoretical issues concerning basic and applied memory research. Hastorf and Cantril were interested in how people perceive complex events, and they took advantage of a real-world event that attracted national interest in 1951. In the fall of 1951, the Dartmouth and Princeton University football teams met in the final game of the year for both teams. During the course of the game there were injuries to key players and numerous plays that resulted in both sides being penalized for infractions. Following the game there were extensive reports of the game in the student newspapers at the two institutions, with each paper reporting very different interpretations of what had transpired. As chance would have it, a film of the game had been made. Hastorf and Cantril (1954) showed the game film to students at the two schools and asked them to count the number of violations committed by each team. Despite the fact that all students saw the same film, there were significant differences between the two samples in terms of the number of violations they perceived during the game; in this case there were clearly effects of the prior publicity about the game, as well as postgame reports, that affected the students’ perception of the same visual events.

The chapters by Intons-Peterson (chap. 1, this volume) and Herrmann and Raybeck (chap. 2, this volume) may represent a situation somewhat analogous to the experiment reported by Hastorf and Cantril. Although Intons-Peterson and Herrmann and Raybeck certainly have not had exactly the same experience with basic and applied memory research, it is reasonable to assume that there is some degree of overlap and that they are commenting on the same general body of research and theorizing. Assume for the moment that this assumption is valid and, with that in mind, consider the very differing “pictures” painted by these authors as they describe the relations between basic and applied memory research.

Striking a rather positive note, Intons-Peterson (this volume) argues that “we strive to explain complex events using theories and models crafted from narrowly contrived situations. These activities depend on mutual feedback from applied and basic research” (pp. 5–6). After reviewing a variety of examples of interactions between basic and applied research, she concludes that “basic and applied research interact, inform, and instruct each other in multiple, interdependent ways. The promise of these mutual interests cannot be overestimated” (p. 18). Intons-Peterson does discuss some of the impediments blocking closer interactions between basic and applied researchers, but in general her message is that there are many areas in which these two groups of workers are benefiting from the knowledge and experience of the other group.

In contrast, Herrmann and Raybeck (chap. 2, this volume) see many problems with the interactions of basic and applied researchers, and they conclude that in general there is little evidence of a symbiotic relationship between researchers conducting basic and applied studies:

Herrmann and Raybeck acknowledge that there have been some cases in which theories from basic research have been applied to real-world problems, but the sense one gets from their chapter is that of frustration on both sides of the fence, with applied researchers finding fault with the theories developed by basic researchers and basic researchers quick to criticize the work of applied investigators on methodological grounds.

It does not appear to us that the differences between the assessments of Intons-Peterson and Herrmann and Raybeck reflect a “glass half full versus glass half empty” situation. Rather, Intons-Peterson has shown that there can be mutual benefits to basic and applied researchers when there is interaction across the basic-applied gulf; Herrmann and Raybeck argue that there is far too little of this going on and there may be fundamental cultural factors that converge to limit interactions between the two groups.¹

We believe that both of these perspectives have merit and that both are reasonably accurate in what they portray—that is, there are some success stories and there are some grumblings among workers in both areas. We are not totally convinced that applied researchers are as frustrated as Herrmann and Raybeck suggest; we do not question their observations, it is just that in our experience (an admittedly small and biased sample), we do not perceive the same level of distress. We are willing, however, to accept their assertion that there are applied memory researchers who feel that the outlook is pretty bleak for basic research being useful in helping them address the problems they are studying.

In the following three sections of this chapter we deal with three fundamental issues pertaining to the relation between basic and applied memory researchers. First, what do basic and applied researchers expect from theories, models, principles, and so on? Although there are many dimensions on which we could focus, we concern ourselves with the issue of the generalizability of predictions across subjects, contexts, tasks, and so forth. The focus here is on the conditions in which basic memory theories are likely to succeed or fail in various applied settings. We also discuss briefly the different uses that theories are put to in the basic and applied domains in order to clarify what may be misconceptions in both areas. Second, we consider those factors that impact communication between basic and applied researchers and ask how that communication might be facilitated. Finally, assuming that there are in fact “cultural differences” between the two groups (an assumption we accept as reasonable), we ask what might be done to minimize differences between the groups and maximize effective collaboration.²

What do Basic and Applied Researchers Expect from Memory Theories?

The authors of the first two chapters in this volume discuss the ability to generalize and extend theoretical accounts to applied settings. Herrmann and Raybeck refer to the cycle of the discovery of basic principles and their application in the real world (with subsequent feedback into the discovery process based on the successes and failures of the basic principles to account for real world phenomena) as the basic/applied cycle. Intons-Peterson reviews several areas in which theories based on laboratory research have worked well in helping structure solutions to real-world problems. Several other chapters in this volume (see especially chapters 15, 16, and 17) provide additional examples. Textbooks in human factors and human performance (e.g., Druckman & Bjork, 1991; Kantowitz & Sorkin, 1983; Proctor & Van Zandt, 1994; Wickens, 1992) provide further examples that include human memory as well as other cognitive processes.

It is also true that there are counterexamples to the success of the basic/applied cycle, where the principles derived from laboratory research have not fared as well when applied to real-world problems (see, e.g., chapter 9, this volume). Herrmann and Raybeck argue (appropriately, we think) that there are many situations in which laboratory-based theories and models fail to account for behavior in real-world situations. Moreover, Herrmann and Raybeck conclude that because of the lack of generalizability of laboratory-based theories, many applied memory researchers become disillusioned with basic memory theories because they feel that they do not capture important aspects of factors that influence memory outside the constraints of the laboratory. If we accept the conclusions that (a) there are cases in which basic memory theories provide useful accounts of real-world memory performance, and (b) there are many situations in which applied researchers find that the basic memory theories are inadequate, then this raises two interesting and important questions. First, what differentiates those situations in which the basic memory theories succeed when applied to a real-world problem from those in which they fail? Second, what do basic and applied researchers expect from memory theories, or, how do basic and applied researchers use theories?

Characteristics of Theories Related to their Success Versus Failure in Applied Settings

We suggest that there is only one situation under which a basic theory can succeed when applied to a real-world memory problem, but that there are multiple conditions that can result in its failure. It is important to note that in both cases—successes and failures—there is a chain of assumptions, inferences, and so forth, between the basic research theory and the application of that theory to solve a real-world problem. As with any chain, this entire chain is only as strong as its weakest link. Successful applications will depend on a sound empirical database, the principle(s) identified in the theory being factors that exert a large influence over memory performance, and a means of translating the concepts in the theory into real world programs, strategies, and so on. If any of these links is weak then it is likely that the endeavor will fail. Let us consider first the specific conditions needed for the theory to succeed. These will in some sense set the stage for the many ways in which a theory can fail.

Relationship Between a Theory’s Principles and Performance. The theory must have identified factors that exert a powerful influence over memory performance, and these factors must have been tested under a wide range of conditions. In the laboratory, researchers endeavor to try to keep conditions as constant and controlled as possible, and this high degree of experimental control (along with sophisticated statistical analyses) can result in demonstrations of effects that are statistically significant, but that are either small in magnitude or restricted to a very constrained set of conditions. When these types of empirical effects—that is, effects that are small in magnitude, only obtained under a very restricted set of conditions, or both—are used as the basis for formulating theories, it is not surprising that these theories fail when applied to real-world problems. There are simply too many other factors that affect behavior in the real world for the factors isolated in the laboratory to exert an appreciable influence. Note that there are two issues being raised here, one concerning the magnitude of the effects and the other concerning the stability of the effects across a range of conditions. Theories that are built upon either very small effects or stable relations that are evidenced only under a constrained set of conditions are likely to be of limited use in real-world settings.

There are, however, a number of memory principles that have been shown to affect memory performance in a wide range of conditions, with effects that are large in magnitude. Table 3.1 presents a listing of some of these principles or factors along with a reference to a review article and an empirical demonstration of the effect. This list is by no means exhaustive, and it is not intended to indicate that the principle or factor always produces the expected results when used in a real-world setting. These principles or factors are those that appear to us to be relatively solid ones that affect memory performance in fairly predictable ways.

Table 3.1
Well-Established Memory Phenomena

Memory Phenomenon	Brief Description	Reference
Encoding Specificity	Retrieval cues are effective to the extent that they reinstate the context present at encoding.	Tulving & Thomson, 1973
Levels of Processing	The amount of meaningful analysis performed at encoding determines how well the event is retained.	Craik & Lockhart, 1972
Distributed Practice	Multiple study opportunities are more effective when they are spaced in time rather than massed.	Dempster, 1988
Chunking	Organizing large amounts of information into smaller, associated “chunks” of information aids memory.	Miller, 1956
Picture Superiority	Memory is better for information presented in a pictorial format than a verbal format.	Paivio, 1971
Concreteness Effects	Memory for concrete information is better than memory for abstract information.	Marschark & Hunt, 1989

Format in Which the Theory Is Stated. To be successful outside the laboratory, a theory must be stated in a format that lends itself to applications in the real world. Quite simply, in order to translate a theory into a program of action for affecting people’s memory performance as they go about their daily lives, there must be an obvious (or direct) means for mapping the constructs of the theory onto real-world factors that can be manipulated, and the theory must be comprehensible to workers outside the specialized area in which the theory was devised and tested. If the theory fails in either of these respects then it is unlikely that the theory will be successfully utilized (or, utilized at all) by applied researchers. Camp and Foss (chap. 16, this volume) and Geiselman and Fisher (chap. 15, this volume) provide several excellent examples of the ways in which factors identified in basic memory theories can be creatively applied when addressing real-world memory problems. Against this backdrop of the necessary conditions for the success of a basic research theory, consider the conditions that may lead a theory to fail. Theories can fail for any or all of the following reasons.

Differences in Context. The conditions under which the data on which the theory is based are sufficiently different from the real world so that it is unreasonable to expect the theory to make correct predictions. This is one of the reasons why many memory researchers have argued that we need to study memory in naturalistic settings (e.g., Conway, 1993). There are certainly many instances of laboratory research that involve conditions, materials, and tasks that are far removed from any conditions one would encounter in the real world. To be convinced of this, one need only consult a recent issue of any of the basic cognitive psychology journals (e.g., Journal of Experimental Psychology: Learning, Memory and Cognition, Memory & Cognition, Perception & Psychophysics) and peruse the methods sections of the articles. These articles present many elaborate and creative tasks, many of which bear little resemblance to anything people do in their everyday lives. Note that this does not mean that the research reported in these articles is without value. Many of the more advanced sciences (especially the physical sciences) have made important discoveries by studying phenomena under conditions that seldom if ever occur in the “real world.”

We believe that there is a place for basic research that adds to our understanding of the functioning of human memory even if the research is conducted using procedures and conditions that do not occur frequently in the real world. A particularly good example of this is the research on implicit memory that began in the mid-1980s (e.g., Graf & Schacter, 1985) and has continued to develop into an established memory field (for a review, see Roediger & McDermott, 1993). Implicit memory tasks are ones in which the subject is required to perform a task that does not require conscious recollection of a prior learning episode, but that may nonetheless show evidence of retention of some aspect of that experience. For example, if subjects study a list of words and are then given the word fragment _I_ _h_ _t and asked to fill in the blanks to form a word, they are more likely to complete it with elephant if the list studied contained the word elephant than they are to complete it with other control items that were not presented in the list. Early research showed that whereas amnesics performed much worse than normal subjects on explicit memory tasks such as recall (which requires the subject to recollect the items having occurred in a specific context), they showed nearly comparable levels of performance on implicit memory tasks such as word fragment completion (Graf, Squire, & Mandler, 1984).

The distinction between implicit and explicit expressions of memory, and the finding of preserved implicit memory in amnesics have been used to develop training programs for amnesics (Glisky & Schacter, 1987). In these studies amnesic patients are taught to perform tasks such as computer programming and simple data entry using a technique known as the method of vanishing cues. This method is founded upon the fact that amnesics have relatively good implicit memory even when their explicit memory performance is very poor. The technique involves presenting commands to the person (e.g., PRINT, SAVE) and then later cueing them for these terms, first with most of the letters of the command (e.g., P _ I N T) and, over time, with fewer and fewer letters presented in the cue. This technique has proven to be very effective and it is unlikely that the technique would have been developed without knowledge of the preserved implicit memory of amnesics. We view this as an excellent example of where a basic research finding aided in solving a real world problem.

A potential danger that arises, however, is when a theory based on the behavior observed under artificial laboratory conditions is used to make generalizations to the real world situations that differ in important ways from the conditions present in the laboratory. We would argue that much of the frustration on the part of applied researchers alluded to by Herrmann and Raybeck (chap. 2, this volume) stems from this problem. We have more to say about this later in the chapter.

Correspondence Between the Components of the Theory and Real- World Factors. A second reason that a basic memory theory may fail in a real-world setting is that the factors identified by the theory do not have a close correspondence with variables in the real world. For example, the study reported by Norman, Brooks, Coblentz, and Babcock (1992) and discussed by Intons-Peterson (chap. 1, this volume) seems to be a case of basic theories of categorization failing to predict performance in a simulated real world task, in this case diagnosing X-ray features when the X-rays are supplemented with patient case histories. The fact that subjects’ performance in these tasks is different from what would be predicted based on theories of categorization indicates that there is some aspect of the tasks, environments, and other such factors that differed between the simple laboratory tasks and the more complex simulated real-world task.

Performance Is Affected by Factors Considered Outside the Scope of the Theory. Basic theories may also falter if the theory ignores important factors that the framers of the theory considered outside the scope of the theory, but yet are potent variables affecting performance in the real world. Many theories in cognitive psychology have nothing to say about the effects of such routine variables as the emotional state of the individual, cultural biases, health factors, and so forth. Applied researchers, as a result of their experience working in various settings, develop an intuitive knowledge about these factors, much the same as clinical psychologists develop their expertise based on clinical experience. Applied researchers frequently come to identify important memorial and nonmemorial factors that exert considerable influence over the behavior of their target audience (e.g., the context in which memory is used, the overall health status of the individuals). If these factors are not accounted for in basic theories then the applied researcher is likely to de-value the theory as an adequate explanatory framework. Alternatively, a young researcher who has not yet had an opportunity to develop a “feel” for these factors may attempt to apply the theories from the lab to the real world and later, when the theories don’t make accurate predictions, the researcher comes to appropriately question the theories’ adequacy in the context of the applied setting.

The Amount of Variance Accounted for by the Theory Is Small. A fourth reason why a basic theory might fail in its application is that the amount of real world variance accounted for by the theory is small, even if the effect is consistent. That is, the theory may have identified factors that influence behavior, but the change in behavior is slight and can be observed only under tightly controlled conditions. This would seem to be the case for many theories that are based on reaction time measures where a difference between the conditions of interest is, say, 20 ms. This difference may be important in terms of assessing the adequacy of the theory as a theoretical account, but it is unlikely that these 20 ms are going to make much of a difference to octogenarians whose main memory problem is remembering to take their medication at the correct time.

Complexity of the Theory. Finally, one additional aspect of the application of basic theories to real-world problems that needs to be considered is the correspondence between the constructs and variables of the theory, and factors and variables in the real world. Quite simply, a majority of the theories and models that have been proposed by cognitive psychologists include constructs and variables which may have no correspondence to the real world. This is especially true of parallel distributed processing, or connectionist, models. This class of models includes constructs such as “hidden units,” “connection weights,” “patterns of connectivity,” and so forth, and as far as we can tell for many of these models there is no way to derive predictions for real-world memory problems. That is, while these theories and models may be useful basic research tools, it is unlikely that the concepts invoked in the theory can be translated into strategies, techniques, and so on, that can be used to solve real-world human memory problems.³

A related concern is that some cognitive psychology models are so complex that it is close to impossible to derive predictions from the models without making a tremendous number of assumptions, and even then you might be forced to resort to computer simulations to derive predictions. As an example, consider the Search of Associative Memory (SAM) model proposed by Raaijmakers and Shiffrin (1980). The original version of this model involved ten parameters (see Table 3.2). Some of these parameters (a, b, c, d) specify the strength of the association between the context, target item, or background and the memory “image” of the target items. Other parameters (e, f, g) represent the increment in associative strength that occurs when an item is successfully recalled. More recent variants of the SAM model have added additional parameters to account for behavior in specific tasks.

Table 3.2
Parameter Names and Functions in the Raaijmakers and Shiffrin Search of Associative Memory (SAM) Theory

Parameter Name	Parameter Function
a	Context Cue to Image Strength
b	Word Cue to Image Strength
c	Word Cue to Self Strength
d	Residual Word Cue to Image Strength
e	Context to Image Increment
f	Word Cue to Image Increment
g	Word Cue to Self Image Increment
Kmax	Total Failure Stopping Criteria
Lmax	Stopping Criterion for a Word Cue
r	Buffer Size

Our point here is not to impugn the SAM framework as a basic memory theory, as one of us has used the model to account for memory performance in a variety of laboratory tasks (Payne, Anastasi, Blackwell, & Wenger, 1994; Wenger & Payne, 1995). Rather, our point is that when one considers how the SAM model might be used to make predictions about real world behavior, two problems arise immediately. First, very few of the parameters listed in Table 3.2 correspond to observable behavior. Therefore, it is difficult to see how the model can be used to make predictions about memory performance in the real world. Second, given the complexity of the model there are often multiple ways in which the model can account for the same observation. For example, Raaijmakers and Shiffrin (1981) noted that the SAM framework can account for the phenomenon of hy-permnesia (i.e., increased recall performance across the retention interval; see Payne, 1987, for a review) in two different ways. For the applied researcher this degree of flexibility (or power, depending on your theoretical predilections) makes the model difficult to apply; how does one know what the model predicts for any given set of circumstances?

Summary. Regardless of its adequacy as a basic memory theory, for any theory developed in the laboratory to be successful when it is applied in the real world, the theory is going to have to meet three criteria. First, it must be based on empirical support from studies employing a wide range of materials, subject populations, tasks, and so forth. Second, in terms of the impact of the factors identified in the theory on observed performance, the factors must be relatively powerful. That is, we know that there are many nonmemorial factors that influence performance in the real world (e.g., health, motivation level). In order for a memory program or strategy to influence performance, the amount of variability in performance levels accounted for by the theory must be substantial, or else the predictions of the theory will not be borne out when tested in the uncontrolled (from the investigator’s point of view) real world. Finally, there needs to be a relatively straightforward manner in which practitioners can translate the concepts of the theory into concrete practice in the real world. Basic researchers spend much of their time debating over whether a specific experimental manipulation provided an adequate test of a theory; applied researchers do not have this luxury. The theory must make concrete predictions that are practical in a real-world setting, or else the theory is in essence useless to the researcher confronting everyday memory problems.

Illustrative Case of a Successful Memory Theory

It is our view that one of the reasons for the success of the levels of processing framework proposed by Craik and Lockhart (1972) is that it met all of the criteria listed previously. The levels of processing framework highlighted the fact that what the learner does when exposed to an event exerts a tremendous influence over the retention of that event. This is not meant to say that there are not flaws with the framework both as a basic memory theory (Baddeley, 1978) and as a tool for applied settings.

The levels of processing framework has been very well received, at least as indicated by its citations in the scientific literature. Roediger (1993) presented data showing the cumulative citations (in the Social Sciences Citation Index [SSCI]) of the original Craik and Lockhart levels of processing article as well as two other influential memory works published at about the same time (Anderson & Bower’s 1973 Human Associative Memory text and Tulving’s 1972 chapter on the episodic-semantic memory distinction). Presented in Fig. 3.1 is an updated cumulative citation count for these three works from the period 1973–1994. All three of these works have been influential in the field, but it is clear that the levels of processing article has been cited far more frequently than the other two works. While it is true that the SSCI citation count may largely reflect citations by basic researchers, it seems certain that the levels of processing framework has also made its impact in the applied domain. For example, a program that is used in many college learning centers involves teaching students to write down questions about textbook materials that they read and lecture notes that they took. The program in large part encourages students to think about the meaning of the materials they read and listen to, thereby encouraging deeper levels of processing than would occur if the students took a more passive role. The program has proven to be quite successful (e.g., Heiman, 1987). The Craik and Lockhart (1972) framework is also frequently cited in popular press memory improvement texts (e.g., Herrmann, 1992).

To summarize, then, the levels of processing framework identified a fundamental fact concerning human memory, namely that the nature and extent of processing an event receives during encoding exerts a tremendous effect on the retention of that event. This has been shown to hold with different materials (e.g., words, pictures, faces), subject populations (e.g., children, young adults, older adults), and types of retention measures (e.g., recall vs. recognition). (For a review, see Cermak & Craik, 1979.) The framework has also been embraced by applied researchers in a variety of contexts.

One final point that we would like to note here is that in our view one of the main shortcomings of the levels of processing view from an applied perspective is that it says nothing about the manner in which memory will be expressed. That is, the levels framework focuses exclusively on encoding processes and is mute on the role of retrieval processes in affecting memory performance. An alternative view, the transfer appropriate processing view proposed by Morris, Bransford, and Franks (Morris, Bransford, & Franks, 1977; see also Bransford, Franks, Morris, & Stein, 1979) remedies this situation by proposing that encoding processes will aid memory to the extent that the processes engaged at encoding match, or are appropriate for, the processes required by the retrieval task. There are two aspects of the transfer appropriate processing view that we think warrant comment. First, the general notion of transfer appropriate processing has proven very useful as a conceptual framework in the basic memory literature (e.g., Roediger & McDermott, 1993) and we believe that in the future the transfer notion will be used more frequently by applied researchers. Second, applying the notion of transfer appropriate processing requires the researcher, regardless of whether they are doing basic or applied work, to do a task analysis on the memory problem facing the individual. Considering the memory task in detail forces the researcher to evaluate how memories are acquired and also how they are expressed. Camp and Foss (chap. 16, this volume) give several excellent examples of situations in which the expression of memory—how memories are used—is just as important as considering how people learn. Researchers who study learning processes are very sensitive to the learning-performance distinction; memory researchers would be well served to keep this distinction in mind also.

What are Theories Used for in Basic and Applied Research?

Herrmann and Raybeck (chap. 2, this volume) discuss several issues related to, but not directly addressing, how theories are used in the two different camps. It is certainly presumptuous of us to profess to know all of the uses of theories in basic and applied research, but we will nonetheless offer the following observations. It is our impression that theories in basic research are used to provide explanatory accounts of the phenomena of interest to the theorist. There is no dictate that these phenomena must necessarily have corresponding expressions in the real world, just as theories in physics designed to account for the behavior of particles in a vacuum or balls rolling down frictionless inclines do not need to generalize directly to leaves falling to the ground or a golf ball rolling across a green. We make this point because it is important to keep in mind that (a) basic theories have utility as explanatory mechanisms in and of themselves, and (b) it is not the intent of all basic researchers to develop theories that can (or even should) be extrapolated to the real world.

On the other hand, applied researchers want theories that can be used to solve, or at least address, problems in the real world. These problems may well occur in contexts far different from the laboratory, and there are myriad differences of both memorial and nonmemorial sorts that can limit the applicability of laboratory-based theories in these situations. In a way, we think that the situation described by Herrmann and Raybeck (chap. 2, this volume) of applied researchers finding that basic theories oftentimes do not generalize to be rather unsurprising; given the differences in the conditions in these two arenas and the different goals for theories in the two arenas it is to some degree surprising that there is very much generalizability at all!

In light of all of the differences between the contexts in which basic and applied research are conducted, it is heartening to note that there have been some efforts to provide a conceptualization of the types of factors that are likely to influence performance in both domains. For example, nearly 20 years ago Jenkins (1979) proposed a tetrahedral model of memory research that identified four classes of variables that are likely to affect performance in any memory task. The four groups of variables Jenkins identified were Subject Variables (e.g., interests, knowledge, goals), the Criteria! Tasks used to assess memory (e.g., recall, recognition), the Orienting Tasks (or encoding conditions, e.g., the instructions given to subjects, the activities subjects engaged in while the target information was presented), and Materials (e.g., the sensory modality in which the items were presented, the psychological organization of the items, pictures vs. words). Jenkins noted that most researchers concerned themselves with one or two favorite classes of variables, and generally held the other variables constant. However, as Jenkins noted then, and as is now well known to students of memory, there are in fact many complex interactions between, as well as within, these classes of variables, and these must be taken into account when attempting to generalize the results of any set of empirical results to a new context, subject population, and so forth. We are certain that there are far more than four main classes of variables that affect performance in real world tasks that depend upon memory for a prior experience. Jenkins’ model is important in that it reminds us that we need to take into account all of the classes of variables that are likely to affect performance in the tasks of interest. Part of the challenge for both basic and applied researchers is to identify which of the many main effects and their higher order interactions exert the most influence on behavior. Because it is infeasible for any researcher to look at all of these variables, we are forced to either look at the ones that seem to be affecting performance in the tasks that we think are important (e.g., remembering procedures for jobs performed at work or home, paying bills on time) or the ones that we believe will reveal important facts concerning human memory more generally. As we see it, there are always tradeoffs in any research endeavor, and one of the differences between basic and applied researchers concerns what “rules” are used in making the decisions needed to launch an investigation.

Communication between Basic and Applied Researchers

Intons-Peterson (chap. 1, this volume) and Herrmann and Raybeck (chap. 2, this volume) discuss a number of factors that effectively serve to limit the amount and nature of communications between basic and applied researchers, and we agree in general with their analyses. They also list several recommendations for steps that might be taken to minimize the consequences of these factors. We will take a different approach here, and consider some of the things that have been done by human factors professionals in an effort to facilitate communication among basic and applied researchers. This is an appropriate comparison to make because (a) in terms of educational backgrounds there is even more variability within the human factors field than there is among memory researchers, most of whom have degrees in related fields (Herrmann & Raybeck, chap. 2, this volume), and (b) human factors workers have been addressing this issue actively over the past decade with what we view as considerable success. We begin with a brief overview of some of the problems of communications within the human factors field. After that we summarize some of the efforts that have been made to facilitate communications and we end with some speculations as to how these might have implications for basic and applied memory researchers.

Overview of Communications With the Human Factors Field. First we discuss the speciality of human factors. The Human Factors and Ergonomics Society (HFES, formerly the Human Factors Society) is one of the largest professional organizations of human factors specialists, with over 5,200 members throughout the world (Human Factors and Ergonomics Society, 1994, p. 2). A committee of the Human Factors Society defined the goals of human factors as follows: “Human factors is that branch of science and technology that includes what is known and theorized about human behavioral and biological characteristics that can be validly applied to the specification, design, evaluation, operation, and maintenance of products and systems to enhance safe, effective, and satisfying use by individuals, groups, and organizations” (Christensen, Topmiller, & Gill, 1988, p. 7). Human factors specialists have a variety of educational backgrounds including (but not limited to) undergraduate and graduate degrees in psychology, industrial engineering, computer science, electrical engineering, systems science, mechanical engineering, mathematics, and physics. There is both basic and applied research conducted by human factors workers in fields as diverse as transportation, architecture, environmental design, education, speech synthesis, farming, sports and recreation, computer systems, consumer products, and large industrial and military systems. Finally, there are subdivisions within organizations such as the American Psychological Association and the Association for Computing Machinery devoted to human factors issues.

With the diversity of backgrounds and target problems being worked on by human factors specialists it is perhaps not surprising that there have been and continue to be concerns raised about the difficulty of translating basic laboratory research into applications to solve real world problems. These in many ways parallel the problems facing memory researchers and outlined by Intons-Peterson (chap. 1, this volume) and Herrmann and Ray beck (chap. 2, this volume). Below we list three of the steps taken by the human factors community to address these problems; note that our intention is not to present an exhaustive listing of these efforts, nor to imply that all of these efforts have been completely successful. Rather, our goal here is to suggest that these complex problems can be addressed and that it is perhaps now time for memory workers to begin this task of improving communications between basic and applied workers.

The human factors community has looked seriously at how basic laboratory research makes its way into the design and development stage in applied areas. For example, if the organization is working on the development of, say, a next generation aviation system, this is a major undertaking, and there are many times and ways in which knowledge from the laboratory might be used in the design of the new system.⁴ To study the process of system design, a workshop was held in the mid 1980s on the Psychology of System Design. Researchers and specialists from a wide range of backgrounds met and discussed their views on topics such as how the design process works, how behavioral science knowledge can aid in the design of machines to be used by humans, what impediments there are to communications within and across organizations, and so forth. The conference resulted in a text (Rouse & Boff, 1987) that provides a wealth of views, if not definitive answers.

Given the complexity of designing large modern systems it is unlikely that any one perspective is likely to become the dominant view on how the design process can or should proceed; and, in fact, having a diversity of approaches is likely to be healthy for society in the long run. Still, a good place to start with any problem is to try to understand the nature of the problem, and Intons-Peterson (chap. 1, this volume) and Herrmann and Raybeck (chap. 2, this volume) have made a very good start at identifying some of the problems facing applied and basic memory researchers interested in using the results of basic research to inform practical applications, and vice versa.

A second issue that the human factors field faces concerns the effective communication of knowledge from the laboratory to workers in applied settings. The essence of the problem is that if human factors workers have a concrete (although typically complex) question for which they need an answer, how do they locate the answer in the basic literature? There are many barriers to be faced here, including not knowing what information is where, problems understanding the procedures and language used in many basic journals, questions about the reliability of the findings, or important boundary conditions, and so forth. Furthermore, oftentimes the persons seeking the answer may not know enough to ask all of the questions that they would if they had more time to delve into the problem at length. So, what was (and still is) needed is a means for providing access to information that will be comprehensible and will provide a quick summary of what is known about certain aspects of human performance.

The goal of developing a comprehensive human factors database on human performance is a lofty one, and there have been several products and sources developed during the past 15 years that have, with increasing levels of depth and sophistication, addressed this need. Before describing these sources, it is important to note that any efforts to develop such a database are going to involve a major and ongoing commitment on the part of many, many professionals and organizations. In the case of the sources described below, funding for this long-term project has come from, among other agencies, the U.S. Army, U.S. Navy, U.S. Air Force, and NASA. One of the first documents published by Integrated Perceptual Information for Designers (IPID) group was the two-volume Handbook of Perception and Human Performance (Boff, Kaufman, & Thomas, 1986). This collection contained comprehensive and in-depth summaries and reviews of the main theoretical views and empirical findings in a wide range of perceptual, cognitive and motor movement areas. The Handbook was designed to be a stand-alone source of information for human factors designers. It was followed by the 12-volume Engineering Data Compendium: Human Perception and Performance (Boff & Lincoln, 1988) that contained information on a wider range of topics than the Handbook. The Compendium also introduced a novel, standardized format for presenting human performance data and theory. The Compendium presents well-established facts in a format that is standardized across topics, and also lists application areas, cross references to other related topics, and so forth. Furthermore, the prose used in the compendium was intended to be as free of jargon as possible; technical jargon is frequently one of the barriers tononspeciahsts’ attempts to read the primary literature. There is also a user’s guide that accompanies the Compendium to help the first time user learn about the overall structure of the Compendium as well as the format used in the Compendium.

Finally, there has been an ongoing effort to expand and update the information in the Compendium and to integrate all of this information into an online database that is accessible by human factors professionals. There are several computer products that have developed out of these efforts, including mainframe and microcomputer systems developed by the Armstrong Laboratory Human Engineering Division, Wright-Patterson Air Force base and the University of Dayton Research Institute. One of the most recent of these systems is the Computer Aided Systems Human Engineering: Performance Visualization System (CASHE: PVS), which lets the user simulate and experience, interactively, a variety of perceptual and performance phenomena on an Apple Macintosh computer. The phenomena included in the CASHE: PVS are wide ranging and include display vibration, speech intelligibility in noise, perceived motion, and visual search and target acquisition. With this system human factors professionals are able to experience (both visually and auditorally) the phenomena of interest. Other aspects of the computer systems developed at Wright-Patterson provide more conventional access to database information.

It is not clear that applied memory researchers would want or need to have anything as elaborate as the systems that have been developed within the human factors community. However, we believe that the human factors efforts in recent years indicate that there can be successful efforts to translate the findings from basic research into principles and guidelines that can then be applied to address real world problems. We also believe that there are some valuable lessons to be learned from the human factors projects reviewed above. Some of these include (a) the need to establish a method of presenting basic research findings in a format that nonspecialists can understand, (b) the utility of consulting both basic and applied researchers when attempting to provide a method of translating basic theory and research into an applied setting, and (c) the critical need to secure funding sources for these efforts. To date, we are unaware of any systematic effort on the part of memory researchers to establish study groups or organizations to aid in the cross-communication between basic and applied researchers. There have been efforts by the National Research Council to develop texts that summarize what is known concerning human cognitive abilities (e.g., Druckman & Bjork, 1991), and we applaud these efforts. It seems to us, however, that there needs to be further effort devoted to facilitating communication between applied and basic researchers in order to identify exactly what the pressing needs of both groups are in terms of their own goals, where they see each other working together, and how these interactions might best be established and maintained.

The final point we wish to mention concerning the efforts of the human factors community is that the HFES has recently established publications (e.g., Ergonomics in Design) that are aimed at providing illustrative cases in which human factors principles have been applied with success. The target audience for these publications is not necessarily human factors specialists; rather, one of the important functions these publications serve is to make these cases accessible (and comprehensible) to nonspecialists. Other disciplines are also making efforts to make basic research available to applied workers. For example, the American Association for the Advancement of Behavioral Therapy (AABT) has recently launched a new publication targeted specifically for practicing clinicians. The journal, Cognitive and Behavior Practice, is intended to disseminate scientific results to mental health practitioners who are not necessarily actively involved in research. An interesting twist to this journal is that although it will report on basic research findings, the editorial board for the journal is made up of practicing clinical psychologists.⁵

There have also been several recent articles in other journals aimed at presenting summaries of the factors that affect memory function in general (e.g., Kihlstrom, 1994) as well as in certain specific contexts (e.g., Lindsay & Read, 1994). These sources are valuable and we are happy to see them appear in print. We believe, however, that memory researchers need to be more effective and proactive in terms of “getting the word out” about findings from the laboratory that may have implications for the applied setting; there is also the need for an effective forum for passing information in the opposite direction, namely applied researchers identifying problems with current theories, possible real world problems that would lend themselves to investigation by basic researchers, and so on.

Improving Collaboration between Applied and Basic Researchers

Herrmann and Raybeck make several proposals for enhancing the collaborations between basic and applied researchers, and we generally support these proposals. We also agree with their conclusion that there will continue to be cultural differences between basic and applied researchers. This is not necessarily a bad thing—cultural diversity may spawn more creative ideas and approaches than would a uniform, homogenous culture. We think that the time is ripe for improving the collaborations between basic and applied researchers, and in fact we think that things are already beginning to improve. Furthermore, society is beginning to force this issue by paying increased concern to the applicability of the research that is funded by taxes, private foundations, and so forth.

There are several facts that support our contention that things are getting better. First, there have been several new journals started that are designed to, in one way or another, bridge the gap between basic and applied researchers. For example, the editors of the journal of Experimental Psychology: Applied have stated that its “mission is to publish original empirical investigations in experimental psychology that bridge practically oriented problems and psychological theory. The journal also publishes research aimed at developing and testing models of cognitive processing or behavior in applied situations, including laboratory and field settings” (Journal of Experimental Psychology: Applied, 1995). We view the addition of the Applied component of the respected Journal of Experimental Psychology series as a wonderful opportunity for promoting the types of collaborative work and exchange of ideas that are needed to bring basic and applied researchers together. Other journals such as Applied Cognitive Psychology and Memory are also devoted to memory research arising from both basic and applied settings. In addition, several journals that typically focus on basic research (e.g., Memory & Cognition, Journal of Memory and Language) have devoted special issues to applied research, or to topics that are of interest to applied researchers (e.g., false memories). Taken together, these and other developments represent a healthy start towards the goal of increased communications between basic and applied researchers.

There have also been a number of major conferences recently that have attracted workers from basic and applied backgrounds. The Practical Aspects of Memory (PAM) conferences (the most recent of which, PAM III, was the starting point for this volume) have served to focus efforts on high quality research designed to address problems of memory in everyday life. There have also been several international conferences (e.g., the First Biennial Conference of the Society for Applied Research in Memory and Cognition, the First and Second International Conferences on Memory) that have brought scholars from across the globe to discuss research on memory in the real world. The fact that there is now a professional society dedicated to applied research on memory topics (Society for Applied Research in Memory and Cognition) is yet another sign that people are working to lessen the gap between the basic and applied worlds.

While these developments indicate that there has been progress in recent years, there is much that remains to be done. Rather than outline specific approaches to deal with the remaining problems, we present here a short list of goals; their ultimate implementation will depend on who takes up the challenge.

First, there are real and important gains to be reaped from communicating with a larger portion of the population. That is, readers of this volume are already interested in memory issues; what memory researchers, both basic and applied, need to do is effectively communicate their successes in trying to understand memory processes and in dealing with real world memory problems. There could be many approaches for doing this (e.g., writing short articles for existing journals and newsletters, becoming involved in applied societies outside one’s primary area of expertise), but what we have been woefully neglectful of is letting the public know that, as a group, memory researchers have a great deal to offer. Rather than focus on an “us versus them” mentality—applied versus basic, or vice versa—we need to come together and publicize our talents. This would serve to bring basic and applied researchers together and it would also increase the public’s awareness of the availability of talented people. It could also serve to help public policy makers when it comes time to make decisions about funding priorities.

Second, graduate training programs need to do more to expose students to applied research. In the modal cognitive psychology graduate program there is no requirement for students to gain experience in applied work. This seems inappropriate given that (a) there are a limited number of academic jobs available (and hence even those students who think that they want a career in academia may not be able to find a position), (b) it only serves to perpetuate the “ivory tower” mystique and myth, and (c) there are many students who plan to pursue a career in an applied setting. Placing students on internships for a term or a year during their graduate career seems to be an excellent opportunity for students to learn more about working in an applied setting; some may decide they like it, others may decide that it is not for them. At the very least this would allow for increased and improved communications between people currently in a basic research environment and those working in the field. Furthermore, this could serve as a source of ideas for research when students return to campus.⁶ Pezdek (personal communication, July 16, 1995) noted that the Claremont Graduate School has a requirement for students to complete an internship program and the students find the experience to be very valuable.

Finally, we think that it is also important to highlight the commonalties between basic and applied researchers. One of the strong themes that links memory researchers regardless of where they conduct their research is a commitment to the use of the scientific method in designing and conducting research. Wright (chap. 4, this volume) presents an excellent summary and review of some of the methodological issues in conducting memory research in naturalistic contexts. While there are many differences between research conducted in the laboratory and that conducted in the real world, in both domains there is a need for conducting quality research; if we are to draw inferences from our research then the same “rules” for drawing these inferences (e.g., avoiding confounding factors, replicating and extending previous research) will hold in both the basic and applied worlds. Fortunately, there are many groups conducting high quality research in both domains, and as a community of memory researchers we can draw both satisfaction and inspiration from the successes of our colleagues.

Acknowledgments

We thank Richard G. Burright, Gep Coletti, Daniel Goldberg, Celia M. Klin, Douglas J. Herrmann, Peggy Intons-Peterson, and Michael Wenger for comments on an earlier version of this manuscript.

References

Anderson, J. R., & Bower, G. H. (1973). Human associative memory. Washington, DC: Winston.

Baddeley, A. D. (1978). The trouble with levels: A reexamination of Craik and Lockhart’s framework for memory research. Psychological Review, 85, 139–152.

Boff, K. R., Kaufman, L., & Thomas, J. P. (1986). Handbook of human perception and performance. New York: John Wiley and Sons.

Boff, K. R., & Lincoln, J. E. (Eds.). (1988). Engineering data compendium: Human perception and performance. Dayton, OH: Armstrong Aerospace Medical Research Center, Wright-Patterson Air Force Base.

Bransford, J. D., Franks, J. J., Morris, C. D., & Stein, B. S. (1979). Some general contraints on learning and memory research. In L. S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human memory (pp. 331–354). Hillsdale, NJ: Lawrence Erlbaum Associates.

Cermak, L. S., & Craik, F. I. M. (Eds.). (1979). Levels of processing in human memory. Hillsdale, NJ: Lawrence Erlbaum Associates.

Christensen, J. M., Topmiller, D. A., & Gill, R. T. (1988). Human factors definitions revisited. Human Factors Society Bulletin, 31, 7–8.

Conway, M. A. (1993). Method and meaning in memory research. In G. M. Davies & R. H. Logie (Eds.), Memory in everyday life (pp. 499–524). New York: North-Holland.

Craik, F. I. M., & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior, 11, 671–684.

Dempster, F. N. (1988). The spacing effect: A case study in the failure to apply the results of psychological research. American Psychologist, 43, 627–634.

Druckman, D., & Bjork, R. A. (1991). In the mind’s eye: Enhancing human performance. Washington, DC: National Academy Press.

Gillman, D., & Appel, M. (1994). Automated coding research at the census bureau (Statistical Report No. RR94/04). Washington, DC: U.S. Government Census Bureau.

Glisky, E. L., & Schacter, D. L. (1987). Acquisition of domain-specific knowledge in organic amnesia: Training for computer-related work. Neuropsychologia, 25, 893–906.

Graf, P., & Schacter, D. L. (1985). Implicit and explicit memory for new associations in normal and amnesic subjects. Journal of Experimental Psychology: Learning, Memory & Cognition, 11, 501–518.

Graf, P., Squire, L. R., & Mandler, G. (1984). The information that amnesics do not forget. Journal of Experimental Psychology: Learning, Memory, and Cognition, 10, 164–178.

Hastorf, A. H., & Cantril, H. (1954). They saw a game: A case study. Journal of Abnormal and Social Psychology, 49, 129–134.

Heiman, M. (1987). Learning to learn: A behavioral approach to improving thinking. In D. N. Perkins, J. Lochhead, & J. Bishop (Eds.), Thinking: The second international conference (pp. 431–452). Hillsdale, NJ: Lawrence Erlbaum Associates.

Herrmann, D. J. (1992). Super memory: A quick action program for memory improvement. Avenel, NJ: Outlet Book Company.

Hoffman, R. R., & Deffenbacher, K. A. (1993). An ecological sortie into the relations of basic and applied science: Recent turf wars in human factors and applied cognitive psychology. Ecological Psychology, 5, 315–352.

Human Factors and Ergonomics Society. (1994). 1994–1995 Directory and Yearbook. Santa Monica, CA: Author.

Jenkins, J. J. (1979). Four points to remember: A tetrahedral model of memory experiments. In L. S. Cermak & F. I. M. Craik (Eds.), Levels of processing in human memory (pp. 429–446). Hillsdale, NJ: Lawrence Erlbaum Associates.

Journal of Experimental Psychology: Applied (R. S. Nickerson, Ed.). (1995). [June issue]. Washington, DC: American Psychological Association.

Kantowitz, B. H., & Sorkin, R. D. (1983). Human factors: Understanding people-system relationships. New York: John Wiley & Sons.

Kihlstrom, J. F. (1994). Hypnosis, delayed recall, and the principles of memory. International Journal of Clinical and Experimental Hypnosis, 42, 337–345.

Lindsay, D. S., & Read, J. D. (1994). Psychotherapies and memories of childhood sexual abuse: A cognitive perspective. Applied Cognitive Psychology, 8, 281–338.

Marschark, M., & Hunt, R. R. (1989). A reexamination of the role of imagery in learning and memory. Journal of Experimental Psychology: Learning, Memory & Cognition, 15, 710–720.

Miller, G. A. (1956). The magical number seven plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63, 81–97.

Morris, C. D., Bransford, J. D., & Franks, J. J. (1977). Levels of processing versus transfer appropriate processing. Journal of Verbal Learning and Verbal Behavior, 16, 519—533.

Norman, G. R., Brooks, L. R., Coblentz, C. L., & Babcock, C. J. (1992). The correlation of feature identification and category judgments in diagnostic radiology. Memory & Cognition, 20, 344–355.

Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart & Winston.

Payne, D. G. (1987). Hypermnesia and reminiscence in recall: A historical and empirical review. Psychological Bulletin, 101, 5–27’.

Payne, D. G., Anastasi, J. S., Blackwell, J. M., & Wenger, M. J. (1994). Selective disruption of hypermnesia for pictures and words. Memory & Cognition, 22, 542–551.

Payne, D. G., & Lang, V. A. (1991). Visual monitoring with spatially versus temporally distributed displays. Human Factors, 33, 443–458.

Payne, D. G., Peters, L. J., Birkmire, D. P., Bonto, M. A., Anastasi, J. S., & Wenger, M. J. (1994). Effects of speech intelligibility on concurrent visual task performance. Human Factors, 36, 441–475.

Proctor, R. W., & Van Zandt, T. (1994). Human factors in simple and complex systems. Needham Heights, MA: Allyn & Bacon.

Raaijmakers, J. G. W., & Shiffrin, R. M. (1980). SAM: A theory of probabilistic search of associative memory. In G. H. Bower (Ed.), The psychology of learning and motivation: Advances in research and theory (pp. 207–262). New York: Academic Press.

Raaijmakers, J. G. W., & Shiffrin, R. M. (1981). Search of associative memory. Psychological Review, 88, 93–134.

Raud, R., & Fallig, M. A. (1993). Automating the coding process with neural networks. Quirk’s Marketing Research Review, pp. 14–47.

Roediger, H. L. (1993). Learning and memory: Progress and challenge. In D. E. Mayer & S. Kornblum (Eds.), Attention and performance (Vol. 14, pp. 509–528). Cambridge, MA: MIT Press.

Roediger, H. L., & McDermott, K. B. (1993). Implicit memory in normal human subjects. In F. Boiler & J. Grafman (Eds.), Handbook of Neuropsychology (pp. 63–131). Amsterdam: Elsevier.

Rouse, W. B., & Boff, K. R. (1987). System design: Behavioral perspectives on designers, tools, and organizations. New York: North Holland.

Schonflug, W. (1993). Applied psychology: Newcomer with a long tradition. Applied Psychology: An International Review, 42, 5–30.

Tulving, E. (1972). Episodic and semantic memory. In E. Tulving & W. Donaldson (Eds.), Organization of memory (pp. 381–403). New York: Academic.

Tulving, E., & Thomson, D. M. (1973). Encoding specificity and retrieval processes in episodic memory. Psychological Review, 80, 352–373.

Wenger, W. J., & Payne, D. G. (1995). On the acquisition of mnemonic skill: Application of skilled memory theory. Journal of Experimental Psychology: Applied, 1, 194–215.

Wickens, C. D. (1992). Engineering psychology and human performance. New York: HarperCollins.