146 / 126
146
E
UR
A
MERICA
(Hummer & Lariscy, 2011) and used nationally-representative
American data to search for the best functional form of education-
mortality association. Their results suggested the best fitted
functional form is that the educational gradient of mortality was
much steeper after an individual received a high school diploma, but
rather flat before attaining a high school diploma with a drop in
mortality risk upon completing 12 years of schooling (Montez et al.,
2012). However, their results did not disprove the heterogeneity of
the education-mortality association after individuals have attained at
least a high school diploma or even a college/graduate degree. In fact,
a recent study from Lawrence and her colleagues has shown the
heterogeneity of significant education-mortality association for
people with associate, bachelor’s, or more advanced degrees
(Lawrence, Rogers, & Zajacova, 2016). This suggests that using the
trichotomized educational attainment measure showing step-change
reduction in mortality risk at the 12-years-of-education level
(Montez et al., 2012) may overlook the heterogeneity and
importance of education-mortality association for postsecondary
schooling in the United States (Lawrence et al., 2016). Therefore,
this study adopts the framework from Brown et al. (2012) as the
basis upon which to investigate the association between educational
attainment and mortality compression using mode measure as the
central tendency measure of death ages at the aggregate level and
estimating the variation of modal age of death. However, the study
goes beyond their trichotomized educational attainment measure to
examine whether advanced education has become increasingly
important in determining the life chances of individuals in the
United States.
II. Methods
A. Data
The data is drawn from the Health and Retirement Study (HRS),
a biennial survey spanning the years 1992 to 2012, primarily