This is an excerpt from ACSM's Body Composition Assessment With Web Resource by American College of Sports Medicine,Timothy Lohman & Laurie Milliken.
By Vinson Lee, MS; Leslie Jerome Brandon, PhD, FACSM; and Timothy G. Lohman, PhD
Systematic and Random Error
Systematic errors are measurement biases in one direction, which lead to measured values that are consistently higher or lower from the actual value. All measurements are prone to systematic errors, often of different types. Sources of systematic error may be imperfect calibration of measurement instruments, changes in the environment that interfere with the measurement process, and imperfect protocols or methods of observation. An example of systematic measurement error is when a protocol calls for a specific measurement of abdominal circumference and a different protocol is followed, leading to consistently higher or lower results. Systematic errors can be reduced by ensuring that all equipment is properly calibrated and taking meticulous measurements using standardized protocols.
Random errors are errors caused by the lack of predictability (uncertainty) that is characteristic of the measurement process and variation in the variable being measured. These errors fluctuate around the true value and, unlike systematic errors, are unavoidable. Random errors can be reduced by performing repeated measurements. Both systematic and random errors are inherent to the errors discussed in the next two sections.
Systematic errors occur when a bias is introduced in the measurement, causing it to be consistently higher or lower than the actual value. An example of how this could occur is when a technician incorrectly reads a scale on a skinfold caliper as each line equaling 2 cm when each line really equals 1 cm. Random errors occur equally in both directions and do not cause a bias in the measurement. They are not consistently in one direction or another. Better training and more practice can reduce both kinds of errors. It is also recommended that your measurements be certified against a trained professional's measurements so that you are sure you are making accurate assessments.
Intra- and Interobserver Variability
Intraobserver variability is the error or difference obtained by the same investigator when completing the same assessment with the same participant using the same equipment and employing the same techniques. These errors are assessed two ways for this discussion: repeated assessments within the same day (within day) and assessments between days (interday). When assessments are made on different days, other sources of variation, such as hydration status of the participant, may contribute to intraobserver measurement error, whereas fewer sources of error are present when repeated assessments are performed within the same day.
Interobserver variability or error is a measure used to assess the degree of agreement among investigators. It provides a score that indicates the consensus among investigators, often called objectivity of the measurement. Assessment of interobserver variability is necessary because different investigators typically experience measurement error regardless of their efforts to ensure a high degree of precision. This is especially true with inexperienced investigators completing body composition measurements. To better understand and effectively compare body composition assessments of people from different research laboratories across the world, standardized procedures that yield minimum errors are needed.
Following standardized protocols should yield acceptable intra- and interobserver variation for a given body composition measurement method.