Vahevarjatud

Vahevarjatud is a term used in Estonian statistics and data analysis to denote an intermediate hidden variable that lies between observed variables in a probabilistic or causal model. Such a variable is not directly observed; its presence is inferred from the patterns in the data through methods such as factor analysis, structural equation modeling, or Bayesian networks. The vahevarjatud variable often acts as a mediator or mediating factor, helping to explain how one observed variable influences another.

In practice, identifying vahevarjatud variables involves specifying a model that includes latent constructs and estimating it with approaches such as maximum likelihood, Bayesian inference, or other iterative estimation techniques. The concept is closely related to the broader idea of latent variables and to mediation analysis, but the emphasis of vahevarjatud is on its intermediary role within a causal or statistical chain.

Examples commonly cited in discussions of vahevarjatud include latent cognitive ability factors that mediate the relationship between study behavior and test performance, or unobserved biological processes that mediate the effect of an exposure on health outcomes. In Estonian-language literature, the term is used alongside peidetud muutuja or varjatud muutuja to describe constructs that are not directly measured but are inferred from the data.

See also: latent variable, mediation analysis, structural equation modeling, Bayesian network.