The variable which impacts the relationship between an independent variable…

Concept. A research design names each variable by the role it plays in a presumed cause–effect link. The independent variable is the presumed cause and the dependent variable the presumed effect. Any further factor capable of acting on the link between them is an extraneous factor, and a sound design does not leave it free to vary: the researcher holds it constant, matches on it, or statistically adjusts for it so that it cannot supply a rival explanation for the result. A factor treated in that way — one that bears on the independent–dependent relationship and is therefore kept in check — is called a control variable.

Application. The stem describes exactly that third factor: not the presumed cause, not the presumed effect, but the variable whose presence impacts how the independent and dependent variables relate. Under the standard taxonomy that role is the control variable. In an experiment on teaching method (independent) and test score (dependent), for instance, prior achievement affects how strongly method and score co-vary; the researcher therefore equalises prior achievement across groups so its effect cannot be mistaken for the method's — that equalised factor is the control variable.

Contrast with the neighbouring roles.

• antecedent variable — a variable placed earlier in the causal chain, acting on the independent variable itself; the sequence runs antecedent → independent → dependent, so it explains where the cause came from rather than conditioning the cause–effect link.

• predictor variable — the regression and correlational label for the input side of a model, i.e. the variable whose observed values are used to estimate the outcome; it is the independent variable renamed in the language of prediction.

• precedent variable — 'precedent' names an earlier case or ruling that later ones are referred back to; the standard scheme of named variable roles in research design contains no role under this label.

• control variable — the outside factor bearing on the independent–dependent relationship that the researcher deliberately holds constant, matches, or adjusts for, so that its influence on that relationship is neutralised.

Result. The variable that impacts the relationship between an independent and a dependent variable is the control variable.