Opening the book…
A sample is a portion selected for examination. Its value depends on how well it supports the question being
asked. A representative sample reflects relevant characteristics of the material or population for that purpose. Random selection can reduce selection bias, but randomness alone does not guarantee that a particular sample is representative.
Bias is a systematic tendency for selection or measurement to favor certain results. Sampling variation is the chance difference that occurs because only part of a population is examined. A small sample can have high uncertainty without necessarily being systematically biased.
For example, one vial selected from a five-day production run of 150,000 vials provides little information about changes across the run. Selecting only an easily reached vial from the beginning adds a selection problem. A justified plan might cover relevant times, locations, containers, or other sources of variability. The plan must fit the material, risks, and decision rather than relying on a fixed percentage for every situation.
Clinical samples also depend on collection conditions. A blood result may be affected by fasting, collection time, posture, exercise, collection technique, or sample handling. The appropriate sampling procedure defines which conditions need control for the analyte and purpose.
Simple random sampling gives each eligible unit an equal chance of selection. Other valid approaches include stratified sampling, which selects units within defined groups such as production periods or storage locations. A statistician can help design plans for complex or high-risk decisions.
For individuals or containers, define the population and identifiers before selection and use an appropriate random procedure. For an area, a numbered grid may support selection, but access, sample size, and spatial variation must also be considered. Avoid substituting convenient locations without documenting the effect on the plan.
For a liquid or suspension, mix sufficiently to distribute the target material without damaging it or changing the measurement. Mix frozen samples after thawing as the approved method requires; avoid unnecessary freeze-thaw cycles. Do not assume that every biological sample should be fully thawed and mixed in the same way.
For solids, the plan may require samples from multiple locations, depths, or containers. A composite sample combines portions and can estimate an average property, but it may dilute a localized contaminant or hide variability. Use individual samples when the decision requires information about differences between units.
This page titled 15.3: The Sample was last modified on Mon, 09 Nov 2026 00:06:12 GMT and is shared under a CC BY-SA 4.0 license and was authored, remixed, and/or curated by .
The book could not be loaded just now.