148 / 126
148
E
UR
A
MERICA
focus of the study and was chosen in order to examine whether
advanced education plays an increasingly important role in mortality
compression in the United States. These education categories
roughly correspond to less than a high school education or its
equivalent (0-11 years), a high school degree or its equivalent (12
years), some college education or higher (13+ years), some college
but not a bachelor’s degree (13-15 years), bachelor’s degree or
higher (16+ years) (Brown et al., 2012; Hummer & Hernandez,
2013; Montez et al., 2012).
C. Analytic Approach
All-cause mortality rates are estimated from Gompertz models
that regress the risk of death on age, and models are stratified by sex
and education:
(1)
where,
x
is age, and
(2)
The hazard function
m
(
x
) is a conditional failure rate that gives
the instantaneous potential per unit time for the death incidence,
given that the respondent has survived up to time
x
(Kleinbaum &
Klein, 2005; Teachman & Hayward, 1993). The study adopts
mortality rates estimated from regression models rather than
obtaining through observation of the sample in order to smooth the
fluctuations of mortality rates by ages
.
Probability of surviving from age
x
to
x
+1:
(3)
Probability of dying from age
x
to
x
+1:
(4)
The life tables are calculated based on the estimated all-cause
mortality rates, and the probability of surviving/dying calculated
from the all-cause mortality rates above. They begin at age 50 and