The theoretical model seemed to predict longevity for the rural elderly. The Cox
analysis found that risk of death was best predicted by age, male gender, poor health
and no participation in formal organizations. Interestingly, children/relative participation
and friend networks were not included in the predictor variable model. The Time Dependent Covariate analysis to asses the effect that changes in the predictor
variables had on relative risk supported the Cox model, Positive changes in formal
participation were better for survival than having no participation or going from
participation to no participation. As formal social networks increase or decrease
over one's life, risk of death decreases or increases accordingly, controlling for age.