Latent Variable
- A latent variable is a hidden or unobserved construct that is not measured directly but inferred from observed variables.
- It helps explain relationships among observed measures (for example, why taller people tend to weigh more).
- Researchers use latent variable analysis to uncover underlying mechanisms and to improve predictions of future behaviors or outcomes.
Definition
Section titled “Definition”A latent variable is a hidden or unobserved construct that is believed to explain the relationship between observed variables. These latent variables are not directly measured, but rather inferred from the observed variables.
Explanation
Section titled “Explanation”Latent variables are inferred from patterns in observed data rather than measured directly. By positing a latent construct that links multiple observed measures, researchers can better describe and interpret the relationships among those measures. Using latent variable analysis allows researchers to understand underlying mechanisms and to make predictions about future behaviors and outcomes.
Examples
Section titled “Examples”Body composition explaining height–weight relationship
Section titled “Body composition explaining height–weight relationship”Taller individuals tend to have a higher weight than shorter individuals. A latent variable such as body composition can explain this relationship: individuals with a higher percentage of muscle mass may be taller and have a higher weight, even if they have a lower body fat percentage. In this example, body composition is the latent variable that explains the relationship between height and weight.
Intelligence inferred from cognitive tasks
Section titled “Intelligence inferred from cognitive tasks”Intelligence is not directly measured but is inferred from various observable behaviors and performance on cognitive tasks. For example, a person’s intelligence may be inferred from their performance on standardized tests, their problem-solving ability, and their overall academic achievement. In this case, intelligence is the latent variable that explains the relationship among these observed behaviors and task performances.
Use cases
Section titled “Use cases”- Understanding underlying mechanisms and relationships between observed variables.
- Making predictions about future behaviors and outcomes by modeling latent constructs.
- Applying latent variable analysis to infer unobserved constructs from observed measures.
Related terms
Section titled “Related terms”- Observed variable
- Latent variable analysis