Comparans

Comparans is a term used in some scholarly contexts to denote the items that are the subject of a comparison in an analysis, study, or evaluation. The word derives from the Latin comparare, “to compare.” In practice, comparans can be data points, physical objects, stimuli, or theoretical constructs, depending on the field. They are distinguished from the comparator (the standard or reference against which the comparans are judged) and from the broader comparison class (the set of items considered interchangeable for the purpose of the analysis).

Although not universally standardized, the term appears in discussions across philosophy of science, statistics, data science, and cognitive science to emphasize the role of the items being evaluated. In benchmarking and experimental design, comparans are the entities subjected to measurement or modeling, while the outcomes and metrics summarize performance across the comparans. This framing helps clarify what is being assessed and how generalizable the conclusions are.

Examples of usage include machine learning benchmarking, where a set of test instances serves as comparans to compare model predictions, or perceptual studies in psychology, where a set of stimuli are the comparans used to assess perceptual differences. The concept can also apply to theoretical analyses, such as comparing alternative models or hypotheses by evaluating how well each explains the comparans.

See also: comparandum, comparison, benchmark, reference sample, standard.