Anthropometry, the study of people in terms of the their physical dimensions, has classically been performed by the physical anthropologist. More recently, in an attempt to create more efficient people-machine interfaces, engineers and designers have become increasingly aware of anthropometric terminology and dimensions.
This increased awareness on the part of practicing professionals has brought to light the need for a standard glossary of terms related to anthropometry as it is used in the fields of engineering and design. Unfortunately, most designers and engineers are not versed in medical terminology, nor are most anthropologists familiar with engineering. As a result, the majority of available references fail to communicate to both audiences. Therefore, the Z.94.15 Committee was formed for the purpose of identifying those dimensions required in industry, translating their medical definitions, and compiling the resulting terms into a working glossary.
The main purpose of the standard is intended to be that of a technical reference. As such, the practicing professional should be able to consult the standard when encountering unfamiliar terms in texts or when devising a proper reference term for further definition of written or presented material.
A historical overview of the development of anthropometry is discussed in order to provide the reader with an awareness of the varying backgrounds from which anthropometry grew and also, as a result, an indication of the problems of application to the disciplines of industrial engineering and design.
A HISTORICAL OVERVIEW
The following work is intended to give the reader a reference source on the measurements most commonly taken in anthropometric research, as well as their relevance for equipment design. Since all human beings change, it is important to have a means of quantifying the variation of these physical changes. Anthropometry is the formal name for the techniques used to express quantitatively the form of the human body. Thus, anthropometry is the measurement of humans, whether living or dead, and consists primarily in the detailed, accurate, and consistent measurement of the dimensions of the human body (Montagu, 1960:3).
In 1920, the physical anthropologist, Ales Hrdlicka, summarized the purposes for taking anthropometric measurements. He stated that these measurements were taken for the purposes of industrial design, artistic expression of the human body, military purposes, medical, surgical and dental research and procedures, detection of bodily defects and their correction and forensic identification (Hrdlicka, 1920:8). His approach was narrowed considerably since he did admit later on in his treatise (ibid.. 1920:8) that the "industrial and artistic systems are of little interest to us."
To practitioners of anthropometry and scholars in anatomy and physical anthropology, it soon became evident that standardization of measurements and measuring techniques was sorely needed and that this methodological disarray was very prejudicial to progress in the field. From 1870 to 1920, many international meetings were set up to discuss, refine, and standardize anthropometric techniques. Among the most significant of the conferences were the first and second craniometric conferences held in Munich and in Berlin in the years 1877 and 1880 respectively and the thirteenth General Congress of the German Anthropological Societies, held in 1882 at Frankfort-on-Main.
It must be noted that the field of physical anthropology was concerned basically with anthropometric measurements until the early 1950s when the theoretical orientation of the field changed radically and interest in anthropometry was replaced by an ever-growing interest in human evolutionary processes.
Studies of movements of the body were undertaken in industry in this country during the second decade of the present century in order to improve per capita work output. During the 1920's, many management personnel realized the value of placing the work within easy reach of the workers and this led to a large amount of research geared to studying maximum and normal work areas dimensions. During the 1930's and the 1940's, a large amount of military anthropological research was done establishing effects of bodily dimensions and physiques on the design and use of such military equipment as apparel, aircraft, battleships, tanks, etc. During the same period O'Brien addressed the issue of fitting the masses through her work in adult and children garment sizing. These studies helped in synthesizing data from psychology, physiology, anthropology and medicine with engineering to form a field generally known as human engineering.
After World War II, research emphasizing the fitting of the machine to the human became more fully developed as both commercial and military outfits continued to perform research on body dimensions and work space requirements. The design of motor vehicles and mass transit facilities was reevaluated in light of this new orientation of human-machine design specifications.
Recently, it became necessary, as the technological revolution continued, to change the approaches, the types of data gathered, and the measurement instruments in anthropometric research. Dynamic anthropometry was established as it was necessary to develop a scheme of measurement, taking into account three-dimensional coordinate systems. The science of biomechanics (essentially the interdisciplinary study of the mechanical nature and behavior of biological materials) is now closely related to the fields of classical anthropometry and engineering anthropometry.
With the advent of the computer it is now possible to consider the effect of more than two measurements at a time through the use of multivariate analytic routines in which one does not have to determine dependent and independent variables, and can determine the interaction between and within a large amount of variables at the same time
TYPES OF DIMENSIONS
As mentioned in the historical overview, different types of dimensions have been defined for different applications: These include static and dynamic anthropometry. Static dimensions may be subdivided into circumferences, lengths, skinfolds, and volumetric measurements. Dynamic dimensions include link measurements, center of gravity measurements, and body landmark locations. Static and dynamic anthropometry arc also referred to by the names structural and functional anthropometry, terms which more explicitly express the body and its action. Static dimensions are taken with body parts held in fixed, standardized positions. These dimensions are easily obtainable but not so easily applied since design applications often involve the body in functional attitudes. Dynamic dimensions are taken with the body at work, in motion or in workspace attitudes. These measurements are in more difficult to obtain with application limited to a particular workshop or movement studied. Functional dimensions account for interactions among body members. For example, the limit of functional arm reach is not due to arm length alone but is affected by shoulder movement, some trunk movement and the function to be performed by the hand.
Functional anthropometric models used in workspace design are statistical in that they describe the probabilistic location of body landmarks of a population of users in a given workspace.
Human body measurements, when taken from a sufficiently large sample, follow a normal distribution. Static and dynamic dimensions are generally given as percentiles (e.g. one can discuss a 10th percentile female popliteal height as well as a 10th percentile female eye location in a given workspace attitude): a measure which describes the proportion of a subject population less than or equal to that value in a given dimension. Percentiles provide a univariate basis for estimating the proportion of a population accommodated or inconvenienced by a specific design. Multivariate analysis techniques are becoming increasingly employed in anthropometric design as explained in subsequent sections.
Anthropometric Design Principles
Accommodation of anthropometric characteristics is important to the design of anything which must be operated or maintained by a human. This section will examine some of the design principles associated with anthropometric accommodation. The fundamental principle of design must be that all intended users of an item will be able to use it effectively and efficiently, and not be limited by their size.
There are five approaches to achieving anthropometric accommodation:
Select workers who fit the existing design
Custom fit each individual
Have (several) fixed sizes
Make it adjustable
Design for the extreme individual
Selecting workers who fit the existing design is appropriate when it is not practical to modify the design. This approach tends to be self-propagating because once a population of users becomes restricted, the restrictions tend to become institutionalized. It is usually less expensive in the future to continue to design to a limited population than to relax the restriction and expand the design range to be accommodated. Examples of situations where this is approach is used is in commercial and military aircraft. These cockpits accommodate a certain range of sitting heights to afford outside vision and a certain range of limb lengths to afford reach to controls. After several generations of aircraft were built for this population, it would be economically disadvantageous to change the characteristics of the pilot population.
Custom fitting the design to each individual user is a good approach when only a very small number of items are being produced, For example, if a one-of-a-kind race car is being built for a specific driver, that driver's size is the only one that requires accommodation. The small population of astronauts have always had custom-fitted space suits.
Having several fixed sizes is the approach used where (1) there is a large population of users, (2) the unit price is relatively small, and (3) adjustment is impractical. An example is shoes and clothing, which come in a variety of fixed sizes, and the user selects the one best matching his/her personal fit criterion. When using this approach, inventory schemes must be developed. A inventory scheme must define the number produced and stocked at each location for each size to accommodate the demand. Because size is approximately normally distributed, the demand for sizes near the mean will be greater than for sizes near the extremes. One common mistake is not empirically validating the inventory scheme, since the fit criteria applied by individuals may not match the theoretical criteria assumed by the designer.
Making the design adjustable is the approach used where (1) there is a medium-sized population of users or (2) adjustability is practical, or (3) the unit price is relatively large. In automobiles, for example, the high unit cost requires adjustability to fit any potential user. In an office chair, the unit cost is not great, but neither is the cost making chairs adjustable.
Designing for the extreme individual or dimension is appropriate where some limiting factor can define either a minimum or maximum value which will accommodate the population. One example is locating an emergency stop button within the smallest reach. Designing for the extreme individual or dimension is a special case of approach 3 (several fixed sizes) or even of approach 4 (make it adjustable). Designing for the extremes of an unrestricted population would increase the standard 80 inch (2 meter) high doorway by 20 inches (0.5 meters) to accommodate the tallest people, who may be over 8 feet tall.
Usefulness of anthropometric measures for design. Designers must take care when using anthropometric data and selecting which measures to use. For the most part, anthropometric measures were not defined to be useful to designers, but rather for the ease and repeatability of making the measures. Because the human body is pliable and comes in a great range of sizes and proportions, measures were defined to make the best use of surface features and bony landmarks. Anthropometric measures are usually two dimensional (unrelated lengths) and do not describe functional abilities, such as reach envelopes. For this reason, the best use designers can make of anthropometric data is for scaling functional capabilities, and relating the characteristics and performance of specific individuals to populations.
Defining the population. The first and most critical step in anthropometric accommodation is defining the population to be accommodated. If the user population is small, or if only a small number of measures are needed, the designer can consider measuring the actual user population. If this is impractical, the next best practice is to
find a matching population which has already been measured.
Selecting a correlated survey
To find a matching population, one must consider many factors, such as gender, age range, and race. The table below has a sampling of the richness and variety of existing surveys from which designers can choose.
In selecting a matching population, pay particular attention as to how the population was restricted. Military surveys generally have the greatest number of measures, some approaching 200 measures per individual. However, most military populations have restrictions on size and weight, and a usually have more limited age range than is found among civilian workers.
While this section contains a number of definitions of anthropometric measures, this is only a small subset of the hundreds available. The designer should also be aware that measures with the same name may have definitions which differ from one survey to another. So even if time measure has the same name, it is important to compare the actual definition to verify that it is the same.
General Design Requirements
Design and sizing should ensure accommodation. compatibility operability, and maintainability by the user population. Generally, design limits should be based upon a range from the 5th percentile (small) female to the 95th percentile (large) male values for critical body dimensions. For any body dimension, the 5th percentile values means that five percent of the population will be equal to or smaller than that value, and 95 percent will be larger; conversely, the 95th percentile value indicates that 95 percent of the population will be equal to or smaller than that value, and 5 percent will be larger Therefore, use of a design range from the 5th percentile female to 95th percentile male values will theoretically provide coverage for 95 percent of the user population for that dimension. (See FOOTNOTE).
[FOOTNOTE: If the population being accommodated were all the same gender, the design range of 5th percentile male to 95th percentile male values will theoretically provide coverage for 90 percent of the user population. However, in a mixed population, the design range from the 5th percentile female to 95th percentile male values will theoretically provide coverage for 95 percent not 90 percent of the user population. The correct value is 95 percent because of the Distributive Law of Algebra: mA + mB = m(A + B). In a mixed male/female population, let M be the number of men and F the number of women, then the total number of people is (M + F). Since no men are excluded by the 5th percentile female cutoff and no women are excluded by the 95th percentile male cutoff, then the following is true: Excluded are 5% of M plus 5% of F equals 5% of (M + F), which means 95 percent remain after the exclusion, not 90 percent]
Table of Miscellaneous Civilian Anthropometric Surveys
Law Enforcement Officers
Health and Nutrition, Education Survey
Australian Female Pilots
German Office Workers
Swedish Industrial Workers
Swedish Civilian Women
French Young Men
Air Traffic Controllers
National Health Exam Survey
English Civilian Women
Dutch Civilian Women
Women in Dept. of Agriculture
Safety considerations: Where the design is safety critical (that is, failure to perform could result in serious injury to personnel or equipment), it is recommended that design range from the 1st percentile female to 99th percentile male values be accommodated.
Univariate Standard scores, or Z scores. While some may be more familiar with Z scores than percentiles, they are, in fact, two ways of representing the same phenomenon, that is locating individual measures within a distribution. Anthropometrists prefer percentiles because they are easier to explain, and since anthropometric measures are approximately normally distributed, the two can be related using a table of areas under the normal curve. Percentiles have the additional advantage of being computed from cumulative frequencies, and do not require the distributions to be normal, so they also apply to strength measures, which are often not normally distributed. Percentiles are more straight forward, since they have the same sign, whereas Z scores have different signs above and below the mean of the distribution. To convert a percentile to a Z score, simply use the percentile as an area, and look up the Z score in the table. For example, a 95th percentile has an area of the mean (50th percentile) plus an additional 45 percent, which equates to a Z = +1.645. Similarly, a 5th percentile value is 45 percent below the mean and equates to a Z = -1.645.
Clothing allowances: Because the anthropometric data represents nude body dimensions taken in standard anthropometric postures, suitable allowances should be made for postural variation, light or heavy clothing, flying suits, helmets, boots, body armor, load-carrying equipment, protective equipment, and other worn or carried items.
Strength considerations: In most instances where strength is the parameter being accommodated, the design should not link strength and body size. The correlation between body size and strength is low, so a designer should not assume that small people are weak and big people ate strong. Good design practice requires accommodating size and strength separately where both are relevant. Strength is task specific, so one should not apply a strength measured in one location and direction of force to a situation where the location or direction of force is different. Strength accommodation should also consider the duration of the exertion, that is, endurance.
Multivariate Analysis. Univariate percentiles are not appropriate when two or more dimensions are simultaneously used as criteria for design (Moroney and Smith, 1972). When the operator or maintainer must simultaneously perform two or more different actions in such a way that they interact, multivariate techniques should be used to verify accommodation. For example, a vehicle operator must simultaneously have sufficient sitting height/seat adjustment to see over the forward instrument panel, sufficient arm length to reach all hand operated controls, and sufficient leg length to operate all foot pedals. To verify accommodation, individuals with extreme combinations of both proportions are used: tall torso with short limbs, short torso with long limbs. etc. (Hendy. 1990).
Non-additive Percentiles: There is no such thing as a 5th percentile person. However. there are people with one body dimension equal to a 5th percentile population value. Because of lack of proportionality, however, it is unlikely that any one individual will have more than a few dimensions close to 5th percentile. In practice, to exclude only the smaller 5 percent of a population when considering two or more simultaneous variables, the design cutoff value of each variable must be lowered to less than the 5th percentile value.
The percentile values reported in anthropometric surveys are suitable only for univariate accommodation and should not be used for designs where two or more dimensions are used simultaneously as design parameters. While the means of different measures of body parts can be added, the measures away from the means (at the tails of the distribution where the 5th and 95th percentile values occur) are not additive. Because of lack of perfect correlation of body dimensions, those people excluded by the cutoff of one variable are not the same as the people excluded by the cutoff of a second variable. In a simplistic example, the 95th percentile cutoff on one dimension could give 0.95 x 0.95 = 0.90 or about a 10 percent excluded at the large end of the distribution. If three variables are considered simultaneously, then 0.95 x 0.95 x 0.95 = 0.85, or about a 15 percent excluded at the large end of the distribution (Robinette and McConville, 1982: Moroney and Smith, 1972).
Methods for determining multivariate limits: There are several methods for evaluating multivariate accommodation. Where accommodation of a certain percentage of a population is desired, such as the central 95 percent, using the raw data from the original anthropometric survey, the critical limits can he found by making iterative eliminations with trial combinations of the critical dimensions until the desired percentage of population remains. A second method involves using a hypothetical sample of individuals having extreme dimensions on one or a combination of variables (Bittner, 1986). A dial-up data base is now available which can be used to calculate multivariate "test cases" for a number of workplace/crew station design applications. The data base is described in "User's Guide to the Anthropometric Database at the Computerized Anthropometric Research and Design (CARD) Laboratory," by J. C. Robinson, K. M. Robinette, and C. F. Zehner (1992).
Methods of verifying multivariate limits: One method involves testing with a group of subjects carefully selected so each has at least two or more of the required dimensions. Each of the subjects are tested in a mockup or the actual equipment. In practice, it is difficult to find subjects with the rare combinations of size and proportions desired. It may he necessary to use subjects with correct torso and arm dimensions to evaluate the seat-hand control accommodation, then use separation subjects with correct torso and leg dimensions to evaluate the scat-pedal accommodation. There are also techniques for "shimming" subjects to simulate the desired dimensions.
Computer models. Computer models can generate 3-D manikins which can be dimensioned to match all the subjects in a multivariate set. Whereas finding people with specific dimensions and unusual body proportion is a very difficult task, computer models allow the designer to easily define subjects of any size and proportion. These models can be used to evaluate "electronic mockups as described in Models for Ergonomic Analysis and Design: COMBIMAN and CREW CHIEF" (McDaniel, 1990) and in "The Development of Computer Models for Ergonomic Accommodation" (McDaniel 1990).
Why 5th to 95th design criteria?
The traditional design range of from 5th to 95th percentile of the variable is a trade-off between the necessity of not considering every one in the design range and the practicality of including almost everyone.
The range of human size is much greater than most people realize, with stature of adults ranging from 2 to 8 feet and the weight of adults ranging from 20 to 1,000 pounds. Most of us will never even meet people with these extremes of body size, so the designer should not be forced to consider them when designing a product.
Figure 1 is a plot of a hypothetical dimension (Y axis) to be accommodated against percentiles (X axis) for a cumulative-normal distribution. While the normal distribution has tails which theoretically go to out to infinity, the range of percentiles in this figure go from the 0.05 percentile (that is, 5 in 10,000) to the 99.95 percentile. Thus, the graph is constructed so that the range of Y, which goes from 0 to 100, covers the frequency range of the central 99.9 percent of the population, or all but 1 in 1,000.
Now, compare the extremes of the distribution with the 5th percentile measures (which equals 25 on the Y axis) and the 95th percentile measure (which equals 75 on Y axis). This shows that the included range in Y between the 5th and 95th percentile is 50 units or half the Y range. In other words, accommodating the central 90 percent (5th to 95th percentile) costs half as much as accommodating the whole range.
Examining the slope of the curve, one can also see that it turns and gets very steep beyond the 5th to 95th percentile points. So the 5th and 95th percentile marks have been chosen as the most cost effective design range. It excludes a very small part of the population, and most of those excluded can still cope with the design, since people have some ability to stretch or stoop.
The problem of the quantification of body builds (somatotyping) has been investigated for many years. One of the earliest attempts at quantification of the bodily Physique was undertaken by J. Matiegka (1921) who broke down the human physique into four (4) components: gross both weight, skeletal, dermal, and muscular components, Matiegka's study was important in that he pointed to a direction of considering the human body as capable of being measured in a functionally oriented, "dynamic" fashion.
Unfortunately. Matiegka's article was long ignored in anthropometric circles. Using more sophisticated mathematical models, Tanner, Healy and Whitehouse (1959) reconfirmed with slight modifications Matiegka's original observations by performing a factor analysis of body build and finding that there are four orthogonal (independent) factors: skeletal frame size, skeletal width, muscle width, and fat width (Tanner et al. 1959:91). (Factor analysis is a multivariate analytical procedure which identifies the independent factors that explain most of the variation in a sample in decreasing order of importance.)
It must be noted that the measurement of body composition is an important advance from the grossly inappropriate height-weight tables that are currently enjoying great popularity. Taking someone's weight and comparing the resulting deviation from the corresponding "ideal weight'' is almost biologically meaningless. One does not have any idea of the amount of muscle versus the amount off fat in the body. The visual system of "body typing" developed by W. C. Sheldon in 1940 is far from desirable when scrutinizing the human body in a functional frame of reference.
We shall briefly consider here the role of the skinfold as an indicator of "fatness vs. leanness," i.e.. as an estimator to the individual's muscularity. The relevance of this procedure to apparel design will then become apparent. However, even though at present, the skinfolds' primary area of application would be apparel design and for biomedical and physiological studies, this does not negate the potential for use in the very near future, especially with regard to dynamic anthropometric studies.
Definition and Methodology. The procedure known as "skinfold'' is used basically to determine the amount of subcutaneous fat. It must be noted that the thickness of the skin is relatively constant throughout the body (Montagu, 1960:85). However, the pattern of fat deposition is highly variable both within and between individuals and populations. As a result, the choice of the site for the taking of a skinfold measurement is critically important. At present, there are two (2) primary sites where comparable data on skinfolds can be extracted.
Triceps skinfold: This is one of the least culturally objectionable and most readily accessible sites on the body. Individuals of both sexes in most cultures can be measured at this site with the least amount of tension and conflict. The site is located at the back of the right or left (it is important to be consistent with the choice of side) upper arm over the triceps muscle at a level approximately midway between the tip of the scapular acromial process and the elbow. The forearm is conventionally flexed at a 90° angle. Once the site is located, the subject then lowers the forearm so that the arm hangs freely as the measurement is taken by the investigator.
Subscapular skinfold: The subscapular skinfold is taken with the subject standing freely. The skin is lifted most readily along a line at about a 45 degree angle, going medially upwards and laterally downwards. At this site, the thickness of the subcutaneous adipose tissue layer is fairly consistent.
Other sites: A variety of other sites, such as the midaxillary line, chest, abdomen, and legs have been pro posed, but the variability encountered with the amount of subcutaneous adipose tissue, as well as cultural discomforts, make these sites less desirable.
Summary: In order to assure comparability of skinfold measurements, the following factors should be borne in mind:
(1) the "skin" should be lifted by firmly grasping the fold between thumb and forefinger;
(2) the width of the skin enclosed between the fingers should be minimal, but still yielding a well defined fold;
(3) the depth of the caliper placement should be the minimal distance front the crest where a true fold, with surfaces approximately parallel to each other, occurs;
(4) the side chosen (right or left) and site locations should be consistent at all times; and
(5) the caliper should consistently be read approximately three seconds after the skinfold is lifted to standardize the effects of compression that inevitably occur.
Finally, note that upper arm circumference minus skin and subcutaneous adipose tissue (skinfold) is a gross measure of the muscularity of the subject.
INTENDED USE OF THE STANDARD
Static dimensions, skinfolds arid some dynamic dimensions are defined in this standard. The static dimensions which are included are those frequently referred to in design and engineering and have a well established military data base. A few of the dimensions also have an established civilian data base. There are many dimensions not included, but these are available from the secondary bibliographic sources listed. The standard is not intended to be a complete listing of available anthropometric dimensions. Several functional dimensions used in workspace design are also included.
Dimensions specific to safety in the design of furniture and toys for children have not been included, but appropriate references are included in the bibliography.
Each definition is cross-referenced to a drawing which illustrates the anatomical or body surface land marks which comprise the dimension.
Popliteal Height, Sitting (III, 25)
Dimension Name (Plate Number, Dimension Number)
A Cross-Reference list of Dimension Numbers with Dimension Names has also been included since the reader may not immediately know the dimension name by inspection of the plates. In referring to the drawings, locate the number corresponding to the dimension in which you are interested; then, refer to the Cross-Reference List to locate the name of the dimension.
A Glossary of terms is included following the Cross- Reference List for explanation of anatomic terms. The reader may find that referring to the drawings in conjuntion with use of the glossary explanation will clarify the term.
The drawings included represent only the adult population since bone maturation is completed by then. The focus of this reference is toward design for the adult population. A male skeleton has been used in the drawings, since, in most cases, the dimensions are similarly taken for male and for female, although the actual data bases are distinct, statistically speaking. For data taking in children's populations, some dimensions would require redefinition inasmuch as certain of the bony landmarks are not well-developed and thus are not palpable.
If the reader needs further information on actual data bases or methods of taking anthropometric data, he or she should consult the annotated bibliography. The bibliography is subdivided into two sections: primary and secondary. The primary bibliography lists those sources from which the dimensions in this standard were selected. In some cases, the dimension definition has not been changed at all; changes were made only in those instances where it was necessary for clarity. i.e. by providing language understandable by the layperson. The drawings, however, in this standard do include the skeleton as a reference for the dimensions whereas the figures shown in the primary sources did not provide skeletal reference.
The secondary bibliography lists sources of static and dynamic anthropometry which include a more extensive listing of dimensions as well as a data base specific to children's dimensions. Other sources included present methodological information and practical evaluations of the use of static and dynamic anthropometry.
< Previous | Next >