Expert Article Library

Concepts in Human Factors Engineering (Part 1 of 11)

Concepts in Human Factors Engineering (1): An Overview of Anthropomorphic Data Gathering

by Dennis R Andrews PhD, PSP, CECD

By Dennis R Andrews PhD, PSP, CECD (ExpertPages member profile page)

Human factors are the relationship between humans and their environment such as machinery, health, safety, expectancies and limitations. Anthropometrics involves the study of various sizes, shapes and limits of humans. To properly determine useful data for anthropometrical study the method of gathering this data must be consistent throughout the world. The world consists of many countries with populations that differ greatly as such exacting dimensions must be obtained. For example people of the Orient are smaller in stature than the general population of the United States, therefore expectancies and limitations must be considered.

Concepts in Human Factors Engineering is a series containing eleven articles:
  1. An Overview of Anthropomorphic Data Gathering
  2. The Fragile Skeletal System
  3. The Energy Force of Our Frame
  4. The Message Delivery System
  5. The Body as a Machine
  6. The Oxygen Machine
  7. The Body’s Transportation System
  8. Human Body Energy
  9. A Hot Workplace
  10. The Rhythm of Working
  11. The Bionic Worker of the Future
Measuring methods must be consistent and universal if the information and data gathered is to be considered accurate and meaningful. Naturally if one is using a tape measure whether it is metric or standard units the method must be documented and detailed. Measurements of certain portions of the human body made at a starting point and ending point must be explained in detail. These locations must be very well documented and described with each measurement obtained. For example if one is to measure the height of an individual the starting point may be from the bottom of the foot to the top of the head. The direction or vector of the measurement should also be documented such as vertical, horizontal, fore or aft. Since the human body is not a straight line, whether it is vertical or horizontal, the actual contour or curvature must be documented. Documented points a.k.a. locaters, orientators and positioners must be identified anatomically using universally understood terms such as sagittal, coronal, posterior, anterior, etc.

Various measurements should be taken while the subject is seated, standing, lying flat, as well as performing specific and universal tasks. Valuable data relating to postures must be exact in their documentation and measurement. Postures could have great significance for example in designing seatbacks and headrest's for motor vehicles. There are a number of instruments, which are used for measuring such as calipers and tape measures as well as others. Calipers are probably the most exacting since they can be locked in a measurement position. The type of tool used for specific measurements must also be documented along with the measurement itself, so that measurements can be accurately compared for populations throughout the world. There are disadvantages to manually measuring populations since the measurer may have preconceived or biased assumptions. A more accurate method of measuring is a technique using a laser such as the type of measurement used for animations in Hollywood movies. Documentation using a video camera is a valued asset when using a laser during dynamic measurements. The still photograph can also be a valuable tool and quite objective when documenting measuring static activities.

Most anthropomorphic data involves military personnel, since this type of data was of special interest to the government. One problem with using military data is it does not truly represent the general population. Usually military personnel are young, physically fit, and until recently contained very little data relating to females. While this data can be useful it should be viewed with caution for the reasons stated above. This is not to say that the areas with no data or very limited data of the general population are not useful. Statistics play a very crucial role of determining what one may consider as normal anthropomorphic data. Statistics deal with the average, percentile of a group, and the greater portion of a sample population. While statistics are most helpful they can also be quite deceiving and easily manipulated. In my working career it is not unusual where scientifically trained individuals attempt to bend and interpret statistics and research to their advantage. Any documentation, measuring data and method should be scrupulously documented and possibly photographed so that there will be little if any possibility of misinterpretation and manipulation whether through lack of data or purposely. Recently the Society of Engineers has released a program that has been in the works for some years, it’s called CAESAR (Civilian American and European Surface Anthropomorphic Resources). The project has data on 2400 US and 2000 European civilians. This will help when confronted with data originally taken from the military only and give better data on non military subjects.

One method, which is useful in analyzing data for determining estimations, is through correlation coefficients. This method allows the comparison of various body parts or body section measurements to be compared statistically throughout a specific population. When this method is presented in a graph or chart one can readily pick out the "outliners". These outliners can then be eliminated when statistical calculations are performed. There are many types of correlation coefficients, which can be determined from the human body, i.e. age to weight, waste circumference to shoulder circumference, etc. This method can be quite useful when comparing the population of one country to another. In this way a prediction of human limitations can be determined with a degree of accuracy for designing products for a specific end user. Since body parts of the population can differ greatly the correlation method can bring some degree of stability and confidence to the data presented. The regression analysis is very useful in determining the probability of accuracy considering the wide differences in human dimensions.

Most anthropomorphic measurements are actually static measurements. Dynamic or ranges of function is very important in design and safety. Such movements as bending, stretching, griping, manipulating knobs or handles, etc. are most important considering human dynamics of motion and work task. There are times when within a general population there are many different ethnic backgrounds and consequently many different sizes, shapes and stature. When this is prevalent in a section of a population, the type of work, geographical location of the work and the design considerations are to be considered. For example in the New York City garment section many oriental workers seek employment. Designing for the majority of the specific work population is both necessary and appropriate when considering safety. There are many types of variations of individuals both within and without specific population groups as well as age or secular differences. One consideration of little concern relating to anthropomorphic data of the human may be in designing products, which have a limited or short lifespan.

Some variations in anthropomorphic data can be attributed to sloppy sampling, gathering techniques, unmotivated individuals, etc. As with most processes a motivated person is the best asset to obtain accurate and well documented data. Anthropomorphic data is simply tool used in the design and safety process. Motor vehicle crash testing is done with the 50th percentile male crash dummy which would lead some nonscientific individuals to believe that this would be the average person, when in fact there is no average person. Since some degree of scientific input must be used, the percentile anthropomorphic dummy has been randomly chosen. This type of crash dummy is determined from anthropomorphic data but in reality there is no such person. If one uses a crash dummy to determining forces during an impact the accuracy would be acceptable since the anthropomorphic measurements do not dictate the quantitative force valued. If the anthropomorphic dummy were used to determining a threshold of injuring the outcome would be highly suspect and considered junk science since the anthropomorphic data and the strength of the materials do not replicate biological components of the human body.

Specific selection of anthropomorphic body parts is necessary when designing specific products. Anthropomorphic data should be considered when a job function requires an individual deal with knobs, gauges and levers. It may be necessary for a specific job function, range of anthropomorphic sizes or population percentile to be considered necessary to perform a task safely and satisfactorily. Precise calculations, which resemble the bell curve of normal distribution, can be determined and this range would be quite acceptable to use. As stated earlier body postures are an important value when analyzing tasks in the seating positioning i.e. driving a motor vehicle, performing a repetitious task while seated. To be considered with the posture is the head and eye angle to the objective of the task. For example for a motor vehicle the interior rearview mirror and the driver's seat must have a range of adjustments considering the seating posture as well as the head and eye angles.

One must guard against overall extrapolation from one or two body part measurements to conclude that other body parts have a correlation to specific data. In other words if an arm length is consistent with a fifth percentile value one cannot conclude that data of other body parts of the subject is consistent with all fifth percentile values. If one assumes this correlation for all subjects of the fifth percentile data when only portions of their anatomy are fifth percentile, other data may be outside of this range. Very few instances require an exacting measurement to within a centimeter; consequently a guestimate can be achieved fairly easy. One must be cautious because while one task may simply requiring a guest mate other tasks may require a more exacting measurement. Ratio scaling is basically an estimate from known dimensions and data. This method assumes a critical hypothesis that states while people vary greatly their proportions or body part correlations are similar. This may be true in some circumstances by certainly is not true in all. If one chooses to use this method care must be used so that a more exacting analysis is not needed for safety and productivity.

A high degree of coefficient of correlation must be achieved in order to even consider the ratio scaling method. The regression analysis is a method to estimate the probability and accuracy of estimates, considering linearity exists. Sample size is of course a very important factor since estimating data from a small sample is likely to have less accurate data than from a larger sampling. Some scientists who use a small sample for interpretation and presentation leave themselves open for peer criticism and possibly a fraudulent presentation.

Body parts can vary widely given a specific protocol of measuring and anatomical position. Here again statistics play an invaluable role in determining a usable and acceptable range for use. Specific statistical methods can be employed as well as a combination of anthropomorphic data can be analyzed to arrive at a more appropriate level of accuracy. Dynamics of the human body includes various degrees of freedom for which a person can move about within its immediate spatial area. Some of these dynamics include movements of the lower extremities while walking or climbing stairs. There are specific building codes which determining the construction of stairs, which consider among other things anthropomorphic data. This movement or mobility is directly related to several variables such as age, health, fitness and athletic skill. If we consider the large variable in anthropomorphic data of humans during static measurements there is an even larger variable in dynamic measurements. One reason for this is the large difference in athletic ability between individuals. Workspace areas must take into consideration the dynamics of the population ranges, which usually would occupy a specific workspace. Such dynamics for workspace areas would including but not limited to twisting, forward bending, unusual postures and long static positions such as holding one's arms above their head. Major anatomical joint movement must be considered for various work tasks so the task does not require an individual to move beyond the normal range of motion.

In summary there are many techniques for obtaining anthropomorphic measurements, some are more accurate than others. The most accurate would be the laser method, which simply scans the human body much as an MRI. The laser method is more expensive than other methods plus it is more intrusive in that for accuracy the subject must remove their clothing. Accuracy does not infer correct usage or interpretation. Proper interpretation includes knowing the source of the data and the group population measured. Obviously one would not design machinery for repetitious work for a specific population, which is substantially different than other population. Also the "average person" does not exist and if one designs for the "average person" the designing will fit no one and could be very unsafe. One must also remember that most of the engineering anthropomorphic data is based on military personnel as opposed to civilians, as such the data may not be adequate for a specific design project. If more exacting anthropomorphic data were required for a specific project it would be best to create your own engineering anthropomorphic data from a cross-section of the general population, who would be using the machinery or product. Correlation tables are easier to use than creating your own data but may not have a specific measurement using the locator or orientator of interest to you. When analyzing the engineering anthropomorphic data curvature, circumference and reach must be considered in any design.

  • Click here for the next article in the Concepts in Human Factors Engineering series