latent variable

English

Etymology

The variables are latent (“hidden”) because they are not directly observed, but instead inferred from other variables.

Noun

latent variable (plural latent variables)

1. (statistics) A variable that is not directly measured but is inferred by means of a mathematical model from one or more directly observed variables.
   • 1995, J. Tague-Sutcliffe, L. Vaughan, C. Sylvain, “Using LISREL to Measure the Impact of Information on Development: London Site Pilot Study”, in Paul McConnell, editor, Making a Difference: Measuring the Impact of Information on Development, Proceedings of a Workshop, page 141:

     Each of these is a latent variable, which is measured by one or more observable variables.

   • 2002, Cavid J. Bartholomew, 1: Old and New Approaches to Latent Variable Modelling, George A. Marcoulides, Irini Moustaki (editors), Latent Variable and Latent Structure Models, 2012, page 1,

     The earliest use of latent variable ideas in time series appears to have been due to Wiggins (1973) but, as so often happens, it was not followed up. […] The basis for this remark can be explained by reference to a simple example which, at first sight, appears to have little to do with latent variables.

   • 2008, Donna Harrington, Confirmatory Factor Analysis, page 26:

     As stated earlier, in addition to having df greater than 0, the second condition for model identification is that the latent variables have to be scaled.

Synonyms

• (variable not directly measured): hidden variable, hypothetical construct, hypothetical variable

See also

• latent space