Table 4

Estimated relative mortality risks by race and education according to age group, U.S. 1979–1985

GroupMale Female
 0–8 years1.41–1.551.16–1.381.07–1.151.00–1.631.35–1.501.07–1.10
 (0–8 midrange)(1.48)(1.27)(1.11)(1.32)(1.43)(1.09)
 9–11 years1.381.211.111.511.291.06
 (9–12 weighted avg.)(1.03)(1.02)(1.01)(1.04)(1.02)(1.00)
 12 years1.
 More than 12 years0.48–0.920.60–0.910.76–0.970.64–0.850.81–1.010.82–0.96
 (>12 midrange)(0.70)(0.76)(0.87)(0.74)(0.91)(0.89)
  • Educational categories noted by Sorlie et al. (4) were consolidated into the categories 0–8 years and more than 12 years. The categories 9–11 years and 12 years were consolidated by using a weighted average (with weights of 0.08 and 0.92, since this is the breakdown of 90–94 year olds in these two educational categories in the 1990 census) and are shown in parentheses between the values for these two categories. The numbers derived in this fashion provide rough estimates of relative mortality levels by age and sex for the three educational categories in Table 3. Using the midrange of values for the most and least educated categories may have the effect of exaggerating the magnitude of mortality differentials by education, because it gives equal weight to small, extreme categories. This is the preferred strategy, however, in order to err on the side of exaggerating the amount of the sibling mortality advantage that could be explained by the sociodemographic characteristics of the study population.