Page 13 - IJPS-1-1

Zachary Zimmer, Mira Hidajat, and Yasuhiko Saito

disability free, it implies the proportion of disability-free life is increasing, which is in line
with the compression of morbidity scenario.
The current study uses national longitudinal panel data from China to test whether evi-
dence of a status quo, a compression of morbidity, or heterogeneous compression exists
among older persons during a period of rapid socio-economic and demographic change.
The study first examines this by age and sex. Over a very short period of time China has
become more urbanized and more educated, and both of these factors could be influencing
disability. Thus, second, the study examines changes across four sub-groups: higher edu-
cated urban, lower educated urban, higher educated rural, lower educated rural. The study
asks the questions, do recent changes in total and disability-free life expectancy in China
suggest that older Chinese are living both longer and healthier lives portending a compres-
sion of morbidity and is the result consistent across sexes, rural/urban residence and levels
of education?
2 Methods

2.1 Data

Data come from four waves of the China Longitudinal Healthy Longevity Survey
(CLHLS), a multi-wave panel study of elderly Chinese. It is conducted in 22 provinces
which account for over 85% of China’s population. CLHLS baseline data was collected in
1998. Follow-ups were conducted in 2000, 2002, 2005, 2008 and 2011. Follow-ups com-
bined re-interviews with add-in observations to account for both mortality and non-mor-
tality attrition. When the study was launched the age of the sample was 80+. In 2002 the
study expanded coverage to those 65 and older. Because of the more extensive age range,
the current study begins with the 2002 sample. The sampling strategy involved oversam-
pling oldest-old (age 80+) and as such the study contains a considerable number of octo-
genarians, nonagenarians, and centenarians. The current analysis involves two separate
samples, the first with baseline data collected in 2002 and follow-up in 2005 and the
second baseline in 2008 and follow-up in 2011. Sample sizes are 15,641 and 15,622 re-
spectively.
Assessments of data reliability have been conducted (Zeng et al., 2001 and 2002; Gu,
2007). There is a volume that reviews fundamental findings of the survey (Zeng et al.,
2008). Detailed information and assessments of data quality can be found on the study’s
website (http://centerforaging.duke.edu/chinese-longitudinal-healthy-longevity-survey). The
website lists hundreds of peer-reviewed publications in both English and Chinese, Ph.D.,
M.S. and M.A. theses and other notable publications. A number of studies (Dupre et al.,
2008; Yi et al., 2001 and 2002; Gu, 2009) relate to life expectancy, disability and healthy
longevity.

2.2 Measures
The measure of disability is based on survey items originally employed by Katz et al.
(1963) in their Activities of Daily Living (ADL) scale. Specifically, individuals are asked
whether they can conduct the following tasks on their own and without assistive devices:
bathing, dressing, toileting, getting up from a bed and chair, and eating. Answers for inter-
viewees that are unable to respond due to cognitive and other health problems are provided
by proxy respondents. As in the common conceptualization of the disablement process, the
inability to conduct one of these ADL tasks independently is considered to be a disability
(Verbrugge and Jette, 1994). Age and sex are recorded for each respondent when they first
enter the panel. A rural/urban measure is based on the place of residence variable included
in CLHLS data (Zeng, 2001 and 2002). Education is based on a survey question about

International Journal of Population Studies | 2015, Volume 1, Issue 1 7