Persample

Persample is a term used in data collection and analysis to denote a measurement or observation recorded for each unit included in a sample. It refers to data that are tied to individual samples rather than aggregated across features or groups.

The word is a portmanteau of per and sample, and is used primarily in informal contexts or in specialized workflows where per-sample detail matters. Its usage helps distinguish detailed, observation-level data from higher-level summaries.

In practice, persample data appear as rows in a dataset, with each row representing one sample and columns representing variables measured for that sample. It contrasts with per-feature summaries, such as mean or median across samples. Persample data can be analyzed directly in statistical models or used for diagnostic checks and quality assessments that rely on sample-level variability.

Common uses include sensor networks recording measurements for each reading, survey research collecting responses per respondent, and experimental studies tracking outcomes for each trial. Persample values enable fine-grained analyses of variability, outlier detection, and modeling that accounts for sample-specific characteristics.

Limitations include potential ambiguity and a lack of standardization in terminology. The term is not universally adopted, so when used it should be clearly defined in documentation or methods sections to avoid confusion with aggregated results or other per-sample statistics. See also sampling, data representation, and observation-level data.