The independent variable(IV) of a statistical model is a variable that is not dependent on the other variables in the model.

While I have been studying statistical modeling, I've kept embarrassed with the use of different names indicating the independent variable(IV) even in the same field.

It would be nice to organize the synonymous names for IV and what the origins of the names are such as the research field the name is used often. If we know this, we might be able to use a suitable term for a specific context. It's interesting to know also.

You may write down the synonymous terms to the IV and the simple explanations and their origins. Thanks.

  • 8
    $\begingroup$ Add at least explanatory variable and controlling variable. A full answer to this would be challenging. A few random comments (1) my impression is that use of "feature" signals a machine learning person and is still uncommon in statistics narrower sense (2) "exogenous" to me cries "I am an economist" but it's not a synonym for the others as much of the point is that some predictor variables are endogenous (3) "covariate" used to have a very specific meaning in ANCOVA but it has morphed into a more general meaning (4) "treatment" doesn't appeal (much) outside a medical or agricultural context $\endgroup$
    – Nick Cox
    Commented Jul 1, 2020 at 11:09
  • 7
    $\begingroup$ (ctd) (5) "factor" is not always a synonym for the others and often means (only) categorical predictor (a factor variable could also be a response or outcome) (6) "independent variable" is classic mathematical usage, widely disapproved by many statistical people, but still in common use in some communities, including under the lamentable jargon IV (which for yet other people means instrumental variable, if it doesn't mean intravenous). $\endgroup$
    – Nick Cox
    Commented Jul 1, 2020 at 11:12
  • $\begingroup$ @NickCox Thanks!! I added an explanatory variable. And updated controlled variables into a controlling variable. Not sure which one is more common. $\endgroup$ Commented Jul 1, 2020 at 12:01
  • 1
    $\begingroup$ The initial paragraph may confuse readers because its description of statistical modeling is unrecognizable, especially the phrase "decide the target variable." $\endgroup$
    – whuber
    Commented Jul 15, 2020 at 16:36
  • $\begingroup$ @whuber I'm confused. Could you suggest a better way to describe it? $\endgroup$ Commented Jul 16, 2020 at 9:39

1 Answer 1


The terms dependent variable and independent variable come from experimental research, where they represent the necessary components of an experiment:

An experiment is the systematic manipulation/variation of a independent variable (IV) to observe the influence of this manipulation on a dependent variable (DV) while controlling for confounding variables.

Thus, both have a specific definition and should not be used interchangeably with the other terms on your list. Terms such as treatment are simply more specific names for those two components. For example, treatment is usually used in medical science instead of the term independent variable.

A covariate is defined as a variable that varies with another variable (co-variation). It is the broadest term and its specific meaning varies on the statistical model. Usually, it represents a variable that is not of main interest but confounds, moderates or mediates a relationship.

The terms predictor and outcome are terms that denote components of a statistical model. A predictor is supposed to predict variation in another variable (the outcome). Prediction implies a believed causal relationship between predictor and outcome, but in fact no other assumptions or background information (in contrast to IV/DV) is implied. Every IV is a predictor but not every predictor is an IV. They are often used in contrast to IV/DV to indicate that an analysis uses observational data and not experimental data as defined above. Regressor denotes a predictor in regression models.

Feature and target are jargon for predictor and outcome in machine learning models. I can only guess their origin but I believe they were introduced by image recognition. The features of an object determine its classification.

Factor can have multiple meanings, depending on the model in which this term appears (e.g. an ANOVA model, factor analysis etc.)

  • 2
    $\begingroup$ I would say that the terms dependent variable and independent variable have a long history in pure mathematics too. Whether experimenters started using the terms earlier or later would be a good question. $\endgroup$
    – Nick Cox
    Commented Jul 13, 2020 at 8:01
  • $\begingroup$ I disagree that prediction implies believing causality. $\endgroup$
    – carlo
    Commented Sep 5, 2020 at 20:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.