Anthropometric data are often estimated using predictive formulas or standardized manikins.Most often, these approaches are intended as indicators of the ‘‘average’’ human. As such,its utility can be limited in that there is no individual who is truly average across multipledimensions, and relationships between measures may not be linear or the same betweenpeople. For example, a person with a 50th percentile arm length likely does not have a 50thpercentile leg length (it may be close or quite different). Further, many anthropometric tablesonly present average values (e.g., for center-of-mass location), making estimates of individualdifferences impossible.
A second limitation in the application of anthropometry arises from potential biases. As noted above, most of the larger datasets were derived several decades ago, thus not account2 Physical Ergonomic Analyses 767
ing for general and nontrivial secular trends toward larger body sizes across all populations. Many of these studies were also performed on military populations, and questions arise as to whether the values are representative in general. Additional biases can arise due to ethnic origins, age, and gender. Overall, application of anthropometric data requires careful attention to minimize such sources of bias.
Three traditional approaches have been employed when using anthropometry in design. Each may have value, depending on the circumstances, and differ in their emphasis on a portion of a population. The first, and most straightforward, is design for extremes. In this approach, one ‘‘tail’’ of the distribution in a measure is the focus. In the example above for door height, the tall males were of interest, since if those individuals are accommodated, then all shorter males and nearly all females will as well. Alternatively, the smaller individual
may be of interest, as when specifying locations where reaching is required: If the smallest individual can reach it, so will the larger ones.
The second approach, design for average, focuses on the middle of the distribution. This has also been termed the ‘‘min–max’’ strategy, as it addresses the minimal dimension needed for small individuals and the maximal dimensions for large individuals. A typical nonadjustable seat or workstation is an example of designing for the average. In this case, both the smallest and largest users may not be accommodated (e.g., unable to find a comfortable posture).
Design for adjustability is the third approach, and this seeks to accommodate the largest possible proportion of individuals. For example, an office chair may be adjustable in height and/or several other dimensions. While this approach is generally considered the best among the three, with increasing levels or dimensions of adjustability comes increasing costs. In practice, designers must balance these costs with those resulting from failure to accommodate some users.
In all cases, the design strategy usually involves a goal or criterion for accommodation. Where the large individual is of concern (e.g., for clearance), it is common practice to design for the 95th percentile males. Similarly, the 5th percentile female is used when the smaller individual is of concern (e.g., for reaching). When the costs of failure to accommodate individuals is high, the tails are typically extended. From the earlier example, it might be desirable to ensure that 99.99% (or more) of the population can fit through a doorway.
Application of anthropometry in the design process usually involves a number of steps.
Application of anthropometry in the design process usually involves a number of steps.
Key anthropometric attributes need to first be identified, then appropriate sources of population data (or collect this if unavailable). Targets for accommodation are usually defined early (e.g., 99%) but may change as costs dictate. Mock-ups and/or prototypes are often built, which allow for estimating whether allowances are needed (e.g., for shoe height or gait in the doorway example). Testing may then be conducted, specifically with extremes of the population, to determine whether accommodations meet the targets.
Maury A. Nussbaum
Industrial and Systems Engineering
Virginia Polytechnic Institute and State University
Blacksburg, Virginia
Jaap H. van Diee¨n
Faculty of Human Movement Sciences
Vrije Universiteit
Amsterdam, The Netherlands
Mechanical Engineers’ Handbook: Materials and Mechanical Design, Volume 1, Third Edition.
Edited by Myer Kutz
Copyright 2006 by John Wiley & Sons, Inc.
