1.4 Study Design

In the light of multiple etiology, and prior studies, a hypothesized model is established to illustrate the structural relationships of SES, social interaction, healthy lifestyle and health status, as well as survival time in Figure 1-8. It depicts the underlying direct and indirect pathways from SES to health. There are four latent variables in the ovals（SES, social interaction, healthy lifestyle and health status）and one observed variables in the rectangle（survival days）in the model. Health status and survival days are employed in the study to indicate health outcomes, which were described as quality and quantity of life, respectively. Among the influence factors, SES is a distal factor which individuals have the least control over; social interaction and healthy lifestyle are behavioral factors which individuals could change relatively easier. Single-headed arrows represent the direction of relationship between two variables. It is hypothesized that：

Hypothesis 1: SES has both positive and direct effects on health status.

Hypothesis 2: SES has both positive and direct effects on survival days.

Hypothesis 3: SES affects health status indirectly by means of social interaction.

Hypothesis 4: SES affects survival days indirectly by means of social interaction.

Hypothesis 5: SES affects health status indirectly by means of healthy lifestyle.

Hypothesis 6: SES affects survival days indirectly by means of healthy lifestyle.

Hypothesis 7: Survival days was positively associated with SES, social interaction, healthy lifestyle and health status.

Hypothesis 8: The structural relationships between SES, social interaction, healthy lifestyle, health status and survival days vary by age and gender.