The study, collection, analysis of Human Factors data was originally conducted by the military, using crude physical measurement techniques and statistical analysis including the use of "Frequency Histograms" a mathematical distribution graph or bell shaped curve.  The bell curve is created by a defining a sample population or human measurements, based on a specific measured criteria such as a static or dynamic dimension (i.e. human stature height). The individual measurements are plotted to graph represent and qualify the distribution median measurements from smallest and largest.  The curve defines a minimum and maximum range which will capture or satisfy the samples percentage of the population (5th to 95 percentile).   The highest incidents or measurements are typically grouped in a median or middle 50% range, flanked by increasing and decreasing incidents of measurements getting larger in smaller.   In product development it is standard practice to try and satisfy 90 - 95% of the population in relation to product size, fit, adjustability and comfort. 

     For example the above illustration is of the general pattern or distribution created by collected data, plotted points is somewhat predictable approximating the frequency of occurrence versus the sample group or population or magnitude measurements.  This is referenced as the 50th percentile or median measurements.   The size or measurement ranges of the anthropometric data is distributed to define the whole range signifying the 95^th percentile male – [average largest] versus the 5^th percentile female [average smallest] versus the 50^th percentile male or female [median average].  The 50^th percentile is the largest percentage or top, middle mass of the distribution curve of the sampled population with the population numbers reducing as you get to the left (smaller) or right (larger) in size.  For example the number of extremely small women in the 5^th percentile female population is much lower than the average slightly larger size of the 25^th percentile female population. 

      In general, the example distribution curve could represent any static or dynamic anthropometrical measurement. For example measuring stature or height of only the target market or a female population or both female and male, excluding children.  It should be noted that historically the sampled data (Military data) collected is from mostly males, aged 18- 24 years old, usually with no disabilities, which is not necessary representative of the global human cultural and race diversity.  There has been recent studies (data assembled) to update this data based on a broader sampling of the population.

See Sizing and Adjustability

See Universal Design: Usability - Localization & Internationalization