AI 용어집
인공지능 완전 사전
Mediating variable
An intermediate variable that transmits the effect of an independent variable to a dependent variable, explaining the underlying causal process.
Baron-Kenny method
A classic four-step approach to test mediation, although now superseded by more robust methods like bootstrapping.
Bootstrapping in mediation
A non-parametric resampling technique to estimate confidence intervals for indirect effects, particularly useful for small samples.
Moderated mediation
A model where the mediation effect varies depending on the levels of a moderator variable, introducing a complex interaction between mediation and moderation.
Proportion mediation effect
The ratio of the indirect effect to the total effect, indicating the proportion of the total effect that is mediated by the intermediate variable.
Mediation criteria
A set of statistical and theoretical conditions that must be satisfied to establish the presence of a valid mediation effect.
Simple mediation model
A basic structure involving an independent variable, a single mediating variable, and a dependent variable to analyze a single causal mechanism.
Multiple mediation model
Analysis involving multiple simultaneous mediators, which can be parallel or serial, to capture complex causal mechanisms.
Sobel Test
Classical statistical test to assess the significance of the indirect effect, based on the normal approximation of the product of mediation coefficients.
Counterfactual Causal Mediation
Theoretical framework using the language of counterfactual potentials to rigorously define and estimate mediation effects.