Ergonomics on the Brink of the 21St Century

991219

ERGONOMICS ON THE BRINK OF THE 21ST CENTURY

David Meister

San Diego, California

USA

Introduction
Fundamental assumtions
Measurement
The adequacy of ergonomics theory
What needs to be done?
References

INTRODUCTION

The instigation of this paper was a suggestion by the Editor, M. Helander, that the author discuss something called Activity Theory (Bedny & Meister, 1997) and its relationship to what may be termed information processing theory.

This is a perfectly reasonable topic for a research study, but it evoked questions in the author’s mind: what is the relevance of Activity Theory or any other behavioral theory to the discipline we call Human Factors Ergonomics (HFE)? If one asks such a question it becomes necessary first to define what is meant by HFE, its purposes, the relationship between its research and practice, etc. This is necessary unless one is prepared to deal with the topic only superficially.

To discuss the HFE discipline as a whole, it is necessary to make a number of assumptions which are described in the following section. The author has a very pronounced point of view about HFE which may clash with that of other HFE professionals. For this reason it becomes necessary to, as one might say, "put one’s cards on the table."

This paper is therefore not the usual experimental paper. It looks at the totality of HFE and it applies certain criteria which the more scholarly reader may find disquieting.

It is particularly appropriate at the start of a new millenium that HFE professionals review the status of their profession. In this paper HFE is viewed not only as an intellectual phenomenon, but also as an historical and cultural one (for further detail on this, see Meister, 1999). Everyone recognizes that there has been an extraordinary momentum in the expansion of technology in this century. There has been technology as far back as Ancient Egypt – see James & Thorpe (1994) – but nothing like what we have seen since World War II. It is possible that this technology would have inevitably required a discipline like HFE, even if it had not been called by that name or had had other characteristics. For example, can one think of a behavioral theory stemming from an engineering base? One of the brighter lights in the theoretical arena has originated in this way (see Vincente, in press).

It may be only speculation to say that this extraordinary technological momentum forced consideration of the human (the user), because, as the momentum accelerated, it became more consumer-oriented, it incorporated the user more and more into that technology. Technology has always paid attention to the user, but primarily in Taylor’s (1919) concept as a "productive unit," rather than for his or her own safety, comfort, or pleasure.

This expanding technology enhanced disposable income and therefore contributed to technology’s further expansion. Although it is not usual for HFE professionals to think about economics in connection with what they do (except when research funding dries up), when technology became a more consumer process in the 19th century it forced consideration of the human in the human-technology relationship.

If then we look at HFE in its larger context, it is possible, even highly desirable, to ask questions such as: where are we as a discipline (which means, how effective are we) and where is HFE moving as it enters the 21st century? To ask such questions may be disturbing to professionals who were trained in a somewhat limited laboratory context, but HFE is too important to be confined in our thinking to an experimental laboratory.

FUNDAMENTAL ASSUMPTIONS

The Human-Technology Relationship

From the standpoint of the preceding, the author makes certain assumptions which the reader must recognize if he or she is to understand the paper. The first is that HFE is an autonomous discipline, distinct from its major predecessors, psychology and engineering. That is because it is concerned with a relationship, the human-technology relationship. Psychology is primarily, if not exclusively, concerned with the human; engineering is concerned, primarily, if not exclusively, with technology. It must be emphasized that only HFE is concerned with the relationship between the two. This may seem to the reader as if it were merely a cliché, but it is the foundation of a line of thought without which little of the following will make much sense.

HFE research and practice must therefore stem from an attempt to relate the two aspects of the human and technology. Research seeks to develop a quantitative equation between the two; practice seeks to transform one into the other by deriving behavioral design principles that can be incorporated physically in that design. The research equation between the human and technology seeks to explain in research how variations in technological configurations, i.e., different ways in which hardware and software can be configured, will produce differences in human performance resulting from these variations. The transformation requires the translation of behavioral principles into physical form, thus crossing the physical/behavioral domain barrier.

Types of Knowledge

Research produces both explanatory and instrumental knowledge.

Explanatory knowledge, as produced by traditional experimentation, merely suggests why a certain type of human performance has occurred. Instrumental knowledge takes the explanatory conclusions and suggests how the transformation process previously referred to can be accomplished. An example is how research findings can be changed into design guidelines of the Smith & Mosier (1986) type. Instrumental knowledge may or may not be based on experimental data; it may also reflect the consensus experience of skilled HFE practitioners, who are often both researchers and practitioners.

Traditionally, knowledge has been represented as theory tested by experimentation. It is therefore necessary to look at HFE in terms of the theories it produces. Because HFE is distinctive only in its technological aspects, that theory must be examined not only in terms of human responses to an already developed technology (which is what most HFE research actually consists of) but also in terms of how effectively it describes and implements the design process that produces this technology.

HFE theory, such as it is, and there is not much of it, is largely psychological in orientation. In other words, following the stimulus-organism-response (S-O-R) paradigm, it explains human performance as a response to technological stimuli, but has little or nothing to say about how design produces these stimuli. The author estimates that only about 10% -- 15% of papers given at the Human Factors and Ergonomics Society annual meeting and published in its Proceedings has any relation to design (Meister, 1998). This may be acceptable to many HFE professionals, because they think of themselves as primarily psychologists (Hendrick, 1996), but it does not satisfy those professionals who work more directly with technology and who must deal with the transformation problem.

MEASUREMENT

The essence of science is measurement. If one examines measurement in HFE, it consists solely of standard psychological methods, e.g., observation, interviews, the recording of human actions in terms of error and time, the scaling of attitudes, etc. This is inevitable, given the limited number of ways in which human performance can be manifested, short of an ability to examine human thought processes directly, through telepathy, for example. Our measurement goal is, as explained previously, to equate the human’s performance quantitatively with technological variations. This can be done only by deriving a probabilistic (i.e., a predictive) value for major technological dimensions. A few primitive attempts have been made in this direction under the rubric of human reliability, (Meister, 1999), but much more must be done.

We conceptualize the design process as one that seeks a solution to a problem. Although most design is incremental, in the sense that it builds upon what has been created in earlier system versions, there is always something "new" about design. In major design projects of a highly innovative nature, such as, for example, the development of the first missile, the Apollo space module, the Stealth aircraft, the design was very new and the element of uncertainty which is inherent in all new design was very great.

Design as a Behavioral Process

From the HFE standpoint, therefore, design is problem solving, since the purpose of the designer is to find the best (most effective) means of implementing a function; and that means is often obscure. From that standpoint, that of problem solution, akin to the processes involved in playing chess, design is a behavioral, cognitive process of immense interest to HFE, because of the centrality of design to HFE practice.

HFE theory must therefore seek to explain the problem-solution process, and to harness that explanation to the human-technology equation. One criterion of theoretical adequacy is then the extent to which engineering psychology as described in HFE texts actually relates to design problem solving behavior. Unfortunately, very little HFE theory bears any relation to the design process and is in fact only a slightly altered version of experimental psychology theory. It is probably heretical to say so, but it is possible that this is because HFE, outside of the design/technological process, is essentially only experimental psychology in a special dress; it is only when HFE is considered in terms of design (i.e., in terms of the human-technology transformation relationship) that it becomes a distinctive discipline.

There is nothing wrong with being an experimental psychologist (the Founding Fathers, Fitts, Small, Chapanis, Singleton, Broadbent, were all experimental psychologists); but design (and roughly half of HFE professionals work in some relationship with design) is what gives HFE its uniqueness, a quality the author, and one hopes others besides, prizes.

Information Processing in Design

If design is the solution of a problem (which is a fundamental assumption of our thinking), then it is part of an information gathering and interpreting process, because that is primarily what problem solution is. The author conceives of design as a continuing effort to answer a number of questions which are listed as Table 1. The effort to find answers to these questions generates efforts which culminate in decision making and testing.

Empirical Studies of Design

In the same vein, if design as a behavioral process is so integral to HFE, there is a great need to explore empirically what designers (and this term includes HFE specialists working with engineers) actually do in real design situations, not in a laboratory. Unfortunately, there is a great lack of data about how designers design, because again engineering has focused on technology and HFE has focused on the human response to that technology. Both ignore the relationship which is at the heart of the discipline.

The logic described by the assumptions of this paper requires HFE to allocate some of its intellectual resources to the behavioral study of the design process. Among the few efforts to study design empirically are the work of McCracken (1990) and those categorized as Design Rationale (Moran & Carroll, 1996). But much more must be done.

Table 1. Questions Posed by the Design Problem

Initial Design Analysis
1. What is the scope of the design problem (is it ‘new’ design, update, or revision?) How difficult is the problem?
2. What type of SEP (system, equipment, product) is involved (e.g., aircraft, automobile, word-processing computer, general purpose commercial product); and what parameters are involved as a consequence (e.g., visual factors, strength limits, cognitive decision making?
3. What questions about the details of the design problem must be answered?
4. Are all SEP requirements fully described in the design specification?
5. What behavioral implications (e.g., human thresholds) can be drawn from these requirements; and what further information is needed that stems from these implications?
6. What technology and information can be transferred from predecessor SEPs?
7. What characteristics does the human in the SEP have (e.g., specific constraints, like physical handicaps, or needs and desires to be satisfied by design)? How important is the human role in the SEP? because this determines how much emphasis the human will receive.
8. What is the human in the SEP required to do; what is it anticipated that s/he will do; what personnel requirements, if any, have been specified in the SEP requirements; how will the human’s performance affect SEP performance?
9. What factors, technological and non-technological (e.g., money, time) will affect personnel and SEP functions?
10. What design tools, analytical or evaluational, should be used to assist in answering the preceding questions?
Detailed Design
(All the preceding questions may/will be asked as design progresses.)
11. What additional information must be collected and analyzed as a result of increasingly detailed design; and where can this information be found?
12. What are the various ways in which the SEP components (e.g., the human-machine interface HMI) can be configured? How adequate is the HMI design?
13. What criteria should be used to decide among the various design alternatives? Which alternative is the best, considering all factors? What compromises must be made among competing performance criteria and design characteristics?
Design Verification
(All prior questions may still be asked, in addition to the following):
14. How should the evaluation of the final, total SEP be performed, including behavioral criteria and test methodology?
15. What last minutes fixes, if any, must be made to the SEP production model?

Approaches to HFE

In addition to theory, or perhaps as quasi-theory, we have what one can call "approaches." These have some of the characteristics of theory, but without the formal logic and hence the presumed precision of formal theories. The most recent HFE approach is what has been termed " user-centered design" (Norman & Draper, 1986). Other variations that also stem from the socio-technical approach of Emery & Trist (1969) are called participatory ergonomics (Wilson & Haines, 1997), and socially centered design (Stanney, Maxey & Salvendy, 1997). These orientations represent an effort to achieve a more "humane" approach to design, reflecting perhaps a belated recognition that technology is supposed to serve humans rather than the reverse. Although HFE has always been concerned about the ultimate user of technology (this, after all, is the HFE rationale, to make technology maximally useful, safe, comfortable, etc.), the user-centered approach represents a more direct recognition of the user. It is possible to link this approach to certain types of behavioral theory, such as Activity Theory, which will be described later, or to industrial psychology or industrial ergonomics, both of which have been historically more interested in conditions affecting the worker such as the work environment, work-rest schedules satisfaction, etc. In the history of HFE there have been two (not necessarily opposed, but certainly uncoordinated) approaches, the one concerned with a more cognition or externally – centered focus on design, the other concerned more with the worker’s internal concerns. The former has been dominant in American/Western forms of HFE, the latter more influential in Europe and Australia. The user-centered approach as it is exemplified in Norman and Draper (1986) may be a tendency to bring the two together.

Rapid Prototyping as User-Centeredness

One sees user-centeredness most vividly in what has been termed "rapid prototyping," which became feasible with the introduction of computer technology. In design this has permitted very rapid simulations of interim design efforts, which in earlier days were called "mockup tests." The essence of prototyping is that it introduces the user into the design process as a source of information and as a reflection of user preferences. The user as an arbiter of design goes back to World War I, when aces were asked to fly new fighter aircraft and judge their combat worthiness (a tradition which is still respected in the aviation industry). In the past user opinion was provided by subject matter experts (SMEs) who presumably had special knowledge that was valued in design. With the emphasis on commercial products we see the user being utilized in prototyping not so much as a source of expert knowledge, but as a source of preference or as a consumer, as he or she has always been used in marketing research. If we wish to sell a product to potential buyers, it makes sense to ask them what they like and why they like it, and to incorporate the user in this way into the design process. Others may wish to go further by making them participants in the actual design process, but many designers might have reservations about this.

Reliance on the user in design may be considered as one means to solve the design problem other than through the designer’s personal cognition. However, user contributions to design, especially desires and preferences, must always be adapted to functionality considerations of technological capability (can a particular user desire be designed and manufactured), cost, and reliability.

Moreover, the user in design works best for products designed for mass purchase; it is not quite so effective for systems with a single purchaser, such as the US government, involving a new tank or a new paper pulping mill.

The Limited Nature of Human Response

Everyone will agree that measurement is the heart of science. If, as the author believes, HFE is a distinctive science oriented around the human-technology relationship, one would expect to find a distinctive measurement methodology.