Human Factors Methods

A Practical Guide for Engineering and Design

Neville A. Stanton, Paul M. Salmon, Laura A. Rafferty, Guy H. Walker, Chris Baber, Daniel P. Jenkins

• Publisher: CRC Press
• Year: 2017
• ISBN:
  9781317120155
• Edition: 2
• Topic: Tecnologia e ingegneria
• Subtopic: Salute e sicurezza sul lavoro

This second edition of Human Factors Methods: A Practical Guide for Engineering and Design now presents 107 design and evaluation methods as well as numerous refinements to those that featured in the original. The book has been carefully designed to act as an ergonomics methods manual, aiding both students and practitioners. The eleven sections represent the different categories of ergonomics methods and techniques that can be used in the evaluation and design process. Offering a 'how-to' text on a substantial range of ergonomics methods that can be used in the design and evaluation of products and systems, it is a comprehensive point of reference for all these methods. An overview of the methods is presented in chapter one, with a methods matrix showing which can be used in conjunction. The following chapters detail the methods showing how to apply them in practice. Flowcharts, procedures and examples cover the requirements of a diverse audience and varied applications of the methods. The final chapter, a new addition, illustrates the EAST method, which integrates several well-known methods into a teamwork analysis approach.

1 Introduction to Human Factors Methods

What is a Human Factors Method?

Most readers will be able to identify an example of a Human Factors (HF) problem from their own experience of work or study. An HF problem will more than likely possess some, or all, of the following attributes. It will be a problem which impacts negatively on overall system performance. It will involve humans in systems not behaving as they were expected to. It will be a problem which existing methods of design, evaluation and procurement have somehow not captured, despite indepth analysis. Above all, it will usually be frustratingly resistant to a whole range of purely technical interventions. HF problems therefore have the very real capability to reduce overall system performance to a level that is substantially less effective than was originally intended.

This creates an interesting situation, again one that many readers may find resonates with their own experience. Faced with a ‘Moore’s Law’ of accelerating technological progress, some authors have noted how much more difficult it is becoming to compete on functionality, reliability or manufacturing costs (Green and Jordan, 1999). What they mean is that technology is not the only precursor to the success of a system and that functionality, reliability and manufacturing costs relate more to what a system ‘is’ (in terms of its technical/engineering content) rather than to what a system ‘does’ (in terms of harnessing that technology to enable people to accomplish meaningful real-world tasks). HF problems are troublesome because they do not reside exclusively within the purview of engineering; nor are they the exclusive domain of human scientists. HF problems reside at the interface of both, and HF methods are the means by which they can be tackled.

HF and Design

Much has been made about the timeliness of HF input into projects and it is true to say that considerable time, effort and expense can often be saved by early intervention rather than being faced with a completed system which requires considerable re-design. Unfortunately, this is a common problem. That being said, the appropriateness of a particular HF analysis will of course depend on a number of factors, including which stage of design the project is at, how much time and resources are available, the skills of the analyst, access to the end-user population and what kind of data are required (Stanton and Young, 1999). Fortunately, many HF methods are flexible with regard to the design stage they could be applied to, even if the system itself is no longer as flexible in terms of subsequent changes. Figure 1.1 provides an illustration of this, showing that, in terms of overall effort, HF input can add most value early on in the design process when it is still relatively inexpensive to carry out modifications to mock-ups and prototypes.

There are many methods explained in this book which lend themselves well to being applied at the very early stages of design. In addition to this, there are many methods explained which may be used in a predictive as well as an evaluative manner. Called ‘analytical prototyping’, this is the process of applying HF insights to systems which do not yet exist in physical form. Figure 1.2 presents a generic design process in which different HF methods become applicable and useful at different stages. At the start, we begin with methods that are suited to ‘analytical prototyping’ and to modelling the constraints of a particular problem domain to reveal opportunities for unexpected behaviours. The analysis would then proceed forward with analyses of human error, usability and interface evaluation, amongst others. Each method would be chosen to suit the particular stage of the design life-cycle. For example, in the early stages, methods would be chosen to enable designers and engineers to diagnose important HF dimensions of their proposed systems. In later stages, methods would be chosen which reflect the fact that a physical manifestation of the system now exists and that users themselves can start interacting with it.

Figure 1.1 Illustration to show that HF effort is better placed in the early stages of the design process

Figure 1.2 Application of HF methods by phase of the design process

Given that most HF problems emerge from unexpected interactions at the boundary between people and systems, the need to engage in an evolutionary, iterative, ‘design-test-design’ process emerges as a consistent theme in projects within which the authors have worked. This book is not about the design process per se, but suffice to say that HF problems can be avoided with a systems approach to which HF methods lend themselves very well.

Human Factors Integration

From Figures 1.1 and 1.2, it is clear that the flexibility of application to the various design stages bodes well for HF methods. Something else that bodes well is ISO-13407, ‘Human-Centered Design of Systems’. This international standard codifies the requirement to ensure adequate focus on the potential users of systems at all stages in the design and development process. In support of this, a practice called Human Factors Integration (HFI) represents the ‘business process’ which seeks to meet this aim. HFI is about:

providing a balanced development of both the technical and human aspects of equipment procurement. It provides a process that ensures the application of scientific knowledge about human characteristics through the specification, design and evaluation of systems. (Ministry of Defence, 2000: 6)

HFI exists in various forms within certain sectors of industry. In the UK Ministry of Defence, for example, the HFI process covers six domains: Manpower, Personnel, Training, HF Engineering, System Safety and Health Hazards. The HFI process is intended to be seen as an activity which supports attention to all six domains during the entire system design life-cycle, from design to manufacture, to use and to maintenance and disposal. This book contains methods which can be used to support all these domains. We cover methods that are essential to System Safety and to Manpower, and that can support Training and Personnel. Issues relating to Health Hazards relate in turn to risk analysis, which requires additional knowledge and techniques beyond the scope of this book, yet the methods presented interface well with this partnering domain.

Scientist or Practitioner?

HF methods are built on a foundation of robust human science, with actionable methods forming a major part of the HF discipline. The International Encyclopaedia of Human Factors and Ergonomics (Karwowski, 2001) has an entire section devoted to methods and techniques. Many of the other sections of the Encyclopaedia also make reference to, if not provide actual examples of, HF methods. HF consultants provide services to industry in which HF methods are routinely deployed. In short, the importance of HF methods cannot be overstated. These methods offer the engineer, the designer and the specialised HF practitioner a structured approach to the analysis and evaluation of practical problems. The overall approach can be described using the scientist-practitioner model (Stanton, 2005). As a scientist, the process of applying these methods is as follows:

• extending the work of others;
• testing theories of human-system performance;
• developing hypotheses;
• questioning everything;
• using rigorous data collection and analysis techniques;
• ensuring the repeatability of results;
• disseminating the findings of studies.

As a practitioner, the application of these methods is as follows:

• addressing real-world problems;
• seeking the best compromise under difficult circumstances;
• looking to offer the most cost-effective solution;
• developing demonstrators and prototype solutions;
• analysing and evaluating the effects of change;
• developing benchmarks for best practice;
• communicating findings to interested parties.

In applying the methods contained in this book, you will work somewhere between the poles of scientist and practitioner, varying the emphasis of your approach depending upon the problems that you face. HF methods are useful in the scientist-practitioner model because of the structure and potential for repeatability that they offer. There is an implicit guarantee in the use of methods that, provided they are used properly, they will produce certain types of useful products. It has been suggested that HF methods are a route to making the discipline more accessible to all (Diaper, 1989; Wilson, 1995; Stanton and Young, 2003). This is entirely appropriate given the multi-disciplinary nature of the problems in which HF professionals and engineers/designers will encounter each other. However, despite the rigour offered by methods, there is still plenty of scope for the role of experience. The most frequently asked questions raised by users of ergonomics methods are the following:

• How deep should the analysis be?
• Which methods of data collection should be used?
• How should the analysis be presented?
• Where is the use of the method appropriate?
• How much time/effort does each method require?
• How much, and what type, of expertise is needed to use the method(s)?
• What tools are there to support the use of the method(s)?
• How reliable and valid is/are the method(s) (Stanton and Annett, 2000)?

This book will help to answer some of these questions.

Reliability and Validity

To the engineer or designer, the human sciences in general (and possibly HF in particular) may fall victim to the popular, albeit wholly inaccurate, perception of being a rather ‘woolly’ field. This is not so. HF methods certainly deal with problems which, to engineers or designers, may seem alarmingly loose in definition and which do not conform to any readily identifiable chain of cause and effect. However, in facing such problems, HF methods will provide a welcome source of structure and rigour.

Researchers have identified a dichotomy of Ergonomics methods: analytical methods and evaluative methods (Annett, 2002). They argue that analytical methods (i.e. those methods that help the analyst gain an understanding of the mechanisms underlying the interaction between human and machines) require construct validity, whereas evaluative methods (i.e. those methods that estimate parameters of selected interactions between human and machines) require predictive validity. Construct and criterion-referenced validity play a role in the development of HF theory itself. There is a difference between construct validity (how acceptable the underlying theory is), predictive validity (the usefulness and efficiency of the approach in predicting the behaviour of an existing or future system) and reliability (the repeatability of the results). This distinction is made in Table 1.1.

Table 1.1 Annett’s dichotomy of ergonomics methods (adapted from Annett, 2002)

This presents an interesting question for HF. Are the methods really mutually exclusive? Some methods appear to have dual roles (i.e., both analytical and evaluative, such as Task Analysis For Error Identification), which implies that they must satisfy both criteria. It is plausible, however, as Baber (2005a) argues in terms of evaluation, that the approach taken will influence which of the purposes one might wish to emphasise. The implication is that the way in which one approaches a problem – or, in other words, where on the scientist-practitioner continuum one places oneself – could well have a bearing on how a method is employed. At first glance (particularly from a ‘scientist’ perspective), such a ‘pragmatic’ approach appears highly dubious: if we are selecting methods piecemeal in order to satisfy contextual requirements, how can be certain that we are producing useful, valid, reliable output? While it may be possible for a method to satisfy three types of validity: construct (i.e. theoretical validity), content (i.e. face validity) and predictive (i.e. criterion-referenced empirical validity), it is not always clear whether this arises from the method itself or from the manner in which it is applied. The solution, simply stated, is that care needs to be taken before embarking on any application of methods to make sure that one is...