HUMAN FACTORS ENGINEERING AND SYSTEMS DESIGN
3.3 Applications of Human Factors to System Design Process
System design can be conceptualized as a problem- solving process that involves the formulation of the problem, the generation of solutions to the problem, analysis of these alternatives, and selection of the most effective alternative (Rouse, 1985). There are various ways to classify the various stages in system design.
Meister (1989), on the basis of a military framework, distinguishes among four phases:
1. System Planning. The need for the system is identified and system objectives are defined.
2. Preliminary Design. Alternative system con- cepts are identified, and prototypes are devel- oped and tested.
3. Detail Design. Full-scale engineering is devel- oped.
4. Production and Testing. The system is built and undergoes testing and evaluation.
To maximize system effectiveness, human factors engineers need to be involved in all phases of the process. In addition to human factors engineers, a representative sample of operators (users) should also be included.
The basic role of human factors in system design is the application of behavioral principles, data, and meth- ods to the design process. Within this role, human fac- tors get involved in a number of activities. These activ- ities include specifying inputs for job, equipment and interface design, human performance criteria, operator selection and training, and inputs regarding testing and evaluation. The nature of these activities is discussed at a general level in the next section. Most of these issues are discussed in detail in subsequent chapters.
The intent of this discussion is to highlight the nature of human factors involvement in the design process.
3.3.1 System Planning
During system planning, the need for the system is established and the goals and objectives and perfor- mance specifications of the system are identified. Per- formance specifications define what a system must do to meet its objectives and the constraints under which the system will operate. These specifications determine the system’s performance requirements. Human factors should be a part of the system planning process. The major role of human factors engineers during this phase is to ensure that human issues are considered in the specification of design requirements and the statement of system goals and objectives. This includes under- standing personnel requirements, general performance requirements, the intended users of the system, user needs, and the relationship of system objectives relative to these needs.
3.3.2 System Design
System design encompasses both preliminary design and detailed design. During this phase of the process, alternative design concepts are identified and tested
52 HUMAN FACTORS FUNCTION and a detailed model of the system is developed. To
ensure adequate consideration of human issues during this phase, the involvement of human factors engineers is critical. The major human factors activities include (1) function allocation, (2) task analysis, (3) job design, (4) interface design, (5) design of support materials, and (6) workplace design. The primary role of the human factors engineer is to ensure joint optimization of the human and technical systems.
Function AllocationFunction allocation is a critical step in work system design. This is especially true in today’s work systems, as machines are becoming more and more capable of performing tasks once restricted to humans. A number of studies have shown (e.g., Morris et al., 1985; Sharit et al., 1987) that proper allocation of functions between humans and machines results in improvements in overall system performance.
Function allocation involves formulating a functional description of a system and subsequent allocation of functions among system components. A frequent approach to function allocation is to base allocation decisions on machine capabilities and to automate wherever possible. Although this approach may appear expedient, there are several drawbacks. In most systems not all tasks can be automated, and thus some tasks must be performed by humans. These tasks are typically
“leftover” tasks. Allocating them to humans generally leads to problems of underload, inattention, and job dissatisfaction. A related problem is that automated systems fail and humans have to take over. This can be problematic if the humans are out of the loop or if their skills have become rusty due to disuse. In essence the machine-based allocation strategy is inadequate.
As discussed previously, there are numerous examples of technocentered design. It has become clear that a better approach is complementary where functions are allocated so that human operators are complemented by technical systems. This approach involves identifying how to couple humans and machines to maximize system performance. In this regard, there is much research aimed at developing methods to guide function allocation decisions. These methods include lists (e.g., Fitts’s list), computer simulation packages, and general guidelines for function allocation (e.g., Price, 1985).
The traditional static approach (humans are better
at. . .) to function allocation has been challenged and
dynamic allocation approaches have been developed.
With dynamic allocation, responsibility for a task at any particular instance is allocated to the component most capable at that point in time. Hou et al. (1993) developed a framework to allocate functions between humans and computers for inspection tasks. Their framework represents a dynamic allocation framework and provides for a quantitative evaluation of the allocation strategy chosen. Morris et al. (1985) investigated the use of a dynamic adaptive allocation approach within an aerial search environment. They found that the adaptive approach resulted in an overall improvement in system performance. Similar to this approach is the adaptive automation approach. This approach involves invoking some form of automation as a function of the person’s momentary needs (e.g., transient increase in workload
or fatigue). The intent of this approach is to optimize the control of human–machine systems in varying environments. To date, few studies have examined the benefit of this approach. However, several important issues have emerged in the design of these types of systems, such as what aspect of the task should be adapted and who should make the decision to implement or remove automation.
Task Analysis Task analysis is also a central activity in system design. Task analysis helps ensure that human performance requirements match operators’
(users’) needs and capabilities and that the system can be operated in a safe and efficient manner. The output of a task analysis is also essential to the design of the interface, workplaces, support materials, training programs, and test and evaluation procedures.
A task analysis is generally performed after function allocation decisions are made; however, sometimes the results of the task analysis alter function allocation decisions. A task analysis usually consists of two phases:
a task description and a task analysis. Atask description involves a detailed decomposition of functions into tasks which are further decomposed into subtasks or steps.
A task analysis specifies the physical and cognitive demands associated with each of these subtasks.
A number of methods are available for conduct- ing task analysis. Commonly used methods include flow process charts, critical task analysis, and hierar- chical task analysis. Techniques for collecting task data include documentation review, surveys and question- naires, interviews, observation, and verbal protocols.
As the demands of tasks have changed and become more cognitive in nature, methods have been developed for performingcognitive task analysis, which attempts to describe the knowledge and cognitive processes involved in human performance in particular task domains. The results of a cognitive task analysis are important to the design of interfaces for intelligent machines. A common approach used to carry out a cognitive task analysis is a goal–means decomposition.
This approach involves an analysis of the work domain to identify the cognitive demands inherent in a particular situation and building a model that relates these cognitive demands to situational demands (Roth et al., 1992). Another approach involves the use of cognitive simulation.
Job DesignThe type of work that a person performs is largely a function of job design. Jobs involve more than tasks and include work content, distribution of work, and work roles. Essentially, a job represents a person’s prescribed role within an organization.
Job design involves determining how tasks will be grouped together, how work will be coordinated among individuals, and how people will be rewarded for their performance (Davis and Wacker, 1987). To design jobs effectively, consideration must be given to workload requirements and to the psychosocial aspects of work (people’s needs and expectations). This consideration is especially important in automated work systems, where the skills and potential contributions of humans are often overlooked.
In terms of workload, the primary concern is that work requirements are commensurate with human abilities and individuals are not placed in situations of underload or overload, as both situations can lead to performance decrements, job dissatisfaction, and stress.
Both the physical and mental demands of a task need to be considered. There are well-established methods for evaluating the physical demands of tasks and for determination of work and rest schedules. The concept of mental workload is more esoteric. This issue has received a great deal of attention in the literature, and a variety of methods have been developed to evaluate the mental demands associated with a task.
Consideration of operator characteristics is also an essential element of job design, as the workforce is becoming more heterogeneous. For example, older workers may need different work/rest schedules than younger workers or may be unsuited to certain types of tasks. Those who are physically challenged may also require different job specifications.
In terms of psychosocial considerations, a number of studies have identified critical job dimensions. Gener- ally, these dimensions include task variety, task identity, feedback, autonomy, task significance, opportunity to use skills, and challenge. As far as possible, these characteristics should be designed into jobs. Davis and Wacker (1987) have developed a quality-of-working- life-criteria checklist which lists job dimensions important to the satisfaction of individual needs. These dimensions relate to the physical environment, institu- tional rights and privileges, job content, internal social relations, external social relations, and career path.
A number of approaches to job design have been identified. These include work simplification, job enrich- ment, job enlargement, job rotation, and teamwork design. The method chosen should depend on the actual design problem, work conditions, and individuals. How- ever, it is generally accepted that the work simplification approach does not lead to optimal job design.
Interface DesignInterface design involves specifi- cation of the nature of the human–machine interaction, that is, the means by which the human is connected to the machine. During this stage of design, the human fac- tors specialist typically works closely with engineers and designers. The role of human factors is to provide the design team with information regarding the human per- formance implications of design alternatives. This gen- erally involves three major activities: (1) gathering and interpreting human performance data, (2) conducting attribute evaluations of suggested designs, and (3) hu- man performance testing (Sanders and McCormick, 1993). Human performance testing typically involves building mock-ups and prototypes and testing them with a sample of users. This type of testing can be expensive and time consuming. Recently, the development of rapid prototyping tools has made it possible to speed up and compress this process. These tools have been used pri- marily in the testing of computer interfaces; however, they can be applied to a variety of situations.
Interface design encompasses the design of both the physical and cognitive components of the interface and includes the design and layout of controls and displays,
information content, and information representation.
Physical components include factors such as type of control or input device, size and shape of controls, control location, and visual and auditory specifications (e.g., character size, character contrast, labeling, signal rate, signal frequency).
Cognitive components refer to the information- processing aspects of the interface (e.g., information content, information layout). As machines have become more intelligent, much of the focus of interface design has been on the cognitive aspects of the interface:
Issues of concern include determination of the optimal level of machine support, identification of the type of information that users need, determination of how this information should be presented, and identification of methodologies to analyze work domains and cognitive activities. The central concern is developing interfaces that best support human task performance. In this regard, a number of approaches have evolved for interface design. Ecological interface design (Rasmussen and Vincente, 1989) is an example of a recent design method.
There are a variety of sources of data on the char- acteristics of human performance that can serve as inputs to the design process. These include handbooks, textbooks, standards [e.g., American National Standards Institute (ANSI)], and technical journals. There are also a variety of models of human performance, including cognitive models (e.g., GOMS; Card et al., 1983), con- trol theory models, and engineering models. These models can be useful in terms of predicting the effects of design parameters on human performance outcomes.
As discussed previously, it is the responsibility of the human factors engineer to make sure that information regarding human performance is in a form that is useful to designers. It is also important when using these data to consider the nature of the task, the task environment, and the user population.
Design of Support Materials This phase of the design process includes identifying and developing materials that facilitate the user’s interaction with the system. These materials include job aids, instructional materials, and training devices and programs. All too often this phase of the design process is neglected or given little attention. A common example is the cum- bersome manuals that accompany software packages or VCRs.
Support materials should not be used as a substitute for good design, however; the design of effective sup- port materials is an important part of the system design process. Users typically need training and support to interact successfully with new technologies and com- plex systems. To maximize their effectiveness, human factors principles need to be applied to the design of instructional materials, job aids, and training programs.
Guidelines are available for the design of instructional materials and job aids. Bailey (1982) provides a thorough discussion of these issues. A great deal has also been written on the design of training programs.
Design of Work Environment The design of the work environment is an important aspect of work system design. Systems exist within a context, and
54 HUMAN FACTORS FUNCTION the characteristics of this context affect overall system
performance. The primary concern of workplace design is to ensure that the work environment supports the operator and activity performance and allows the worker to perform tasks in an efficient, comfortable, and safe manner. Important issues include workplace and equipment layout, furnishings, reach dimensions, clearance dimensions, visual dimensions, and the design of the ambient environment. There are numerous sources of information related to workplace design and evaluation that can be used to guide this process. These issues are also covered in detail in other chapters of this handbook.