[Formation of study population for causal inference].

Area of Science:

• Epidemiology
• Causal Inference
• Observational Studies

Background:

• Epidemiological analysis is crucial for causal inference, with study population formation being the initial step.
• Understanding population cross-sections at individual and population levels is fundamental.
• Key assumptions for measurements in observational studies include attribute stability, variable independence, and individual independence.

Purpose of the Study:

• To define population cross-sections and measurement assumptions for epidemiological studies.
• To unify causal inference research based on the timing of cause or exposure.
• To propose a classification of research designs based on causal thinking and population cross-sections.

Main Methods:

• Defining concepts of individual and population level cross-sections.
• Introducing three essential assumptions for measurements in observational studies.
• Classifying research designs into history reconstruction and future exploration based on causal thinking.

Main Results:

• Established definitions for population cross-sections and measurement assumptions.
• Proposed a temporal basis for unifying causal inference research.
• Introduced a novel classification of research designs: history reconstruction and future exploration.

Conclusions:

• Causal thinking provides a foundation for classifying epidemiological research designs.
• The proposed classification aids in understanding the relationship between estimated effects and research designs.
• Further research is needed to refine effect parameters for different designs.