Synonyms

Centenarian study; Chinese study; Psychological resilience; Psychological traits

Definition

This entry aims to introduce the centenarian subsample of the Chinese Longitudinal Healthy Longevity Survey (CLHLS) and present some key findings on psychological traits of centenarians.

Background: Centenarian Studies in the World

Tireless efforts are made to explore the secret of human longevity throughout history. However, it is not until very recent that science-based studies on the mechanism of longevity appeared with sufficient sample sizes and multidisciplinary perspectives have been launched (Poon and Cheung 2012). One shortcoming of most existing longevity research projects is that little research has been done for those who survive to age 100 and beyond, namely centenarians (Poon and Cheung 2012; Zeng 2012).

The urgent call to study centenarians is largely due to the increasing importance of this special subpopulation. Because of the steady decline of mortality at very old ages (Vaupel et al. 1998; Wilmoth et al. 2000), the number of centenarians is booming in the world (Robine et al. 2010; Wilcox et al. 2008, 2010) and is projected to exceed three million by 2050 and possibly 20 million by 2100 in a conservative estimation of the United Nations Population Division (2015). More importantly, with the world population aging, centenarians come to be considered as a model of successful aging or healthy aging (Andersen-Ranberg et al. 2001; Poon et al. 2010). But why could some people live up to age 100 and beyond, while others die at much younger ages? Why could some people live so long but still remain healthy? Although there has been a consensus among researchers that socioeconomic, behavioral, environmental, and biological factors jointly determine one’s longevity and health, to what extent and how exactly these factors contribute to centenarians’ exceptional long and healthy life is mostly unknown.

There have been a number of centenarian studies around the world to attempt to address such research questions. For example, the longest ongoing centenarian study in the contemporary world is the Okinawa Centenarian Study (OCS), which was launched in 1975. The OCS has heretofore collected over 900 centenarians and several thousands of their siblings of septuagenarians, octogenarians, and nonagenarians in Okinawa, Japan. The Georgia Centenarian Study (GCS) is the longest centenarian study in the USA, which started in 1988. In the Phase I (1988–1992), the GCS collected 76 centenarians with 92 octogenarians and 89 sexagenarian as comparisons; 250 centenarians were further included with 80 octogenarians as comparison in the Phase III (2001–2009). The largest centenarian study in the USA is the New England Centenarian Study (NECS), which was launched in 1995. The NECS has collected data from about 1,600 centenarians in the USA with 500 children (in their 70s and 80s) and 300 younger controls since 1995. Several European countries have also launched centenarian studies since the early 1990s such as the Italian Multi-center Study on Centenarians (IMUSCE) (around 2,000 centenarians) and the Longitudinal Danish Centenarian Study (about 300 centenarians) (Koenig 2001; Poon and Cheung 2012). These centenarians and all other relevant studies have resulted in a boom in centenarian studies and improved understanding about their secret of longevity.

However, nearly all centenarian studies are from developed countries. There was almost no scientific research project with a sufficient sample size of centenarians in developing countries before the late 1990s (Zeng et al. 2001). Because the contributions of sociodemographics, psychological factors, and behavioral factors to longevity vary in different cultures and societies with different development stages (Poon et al. 2010; Kolovou et al. 2014; Willcox et al. 2006), it would be interesting to study centenarians from developing countries where the socioeconomic resources, healthcare service, and technology are limited. Furthermore, while there were about 50 centenarians per million in Western Europe (Jeune and Vaupel 1995; United Nations Population Division 2015), there were less than three centenarians per million in China in the 1990s (United Nations Population Division 2015). The genomes of long-lived individuals from China may be more enriched for disease-preventive genes than their counterparts in the West, because they survived the brutal mortality regimes of the past when famine, civil wars, and starvation affected their birth cohorts of many millions. In addition, the genetic composition of the Han Chinese ethnic group is relatively homogeneous. Unlike Western countries that received many immigrants from other parts of the world and thus provide relatively heterogeneous genetic compositions even within the same ethnic group, China received very few international immigrants. Consequently, the Han Chinese are relatively genetically homogenous, compared to the Western counterparts. For example, it was estimated that “the average of genetic differences between Han Chinese population samples (FST = 0.002) was much lower than that among European populations (FST = 0.009)” (Xu and Jin 2008). This is a comparative advantage to increase statistical power for studying effects of genetic and GxE interactions on healthy aging.

Another major limitation of existing literature on centenarian studies has been the lack of surveys with large sample sizes. To address above research questions, including investigating genetic variations in longevity and examining gene-environment interaction effects on longevity and health, large samples are required. Small sample sizes of surveys often produce results with insufficient statistical power or poor robustness; and in some cases, small-sized surveys on centenarians often lack representativeness when the size of underlying centenarian population is relatively large, such as in China. Yet, with few exceptions, the sample sizes of most centenarian studies around the world are less than 1,000 centenarians (Koenig 2001; Poon and Cheung 2012). To promote centenarians studies, there is thus a need for studies with large representative samples in developing societies, such as in China which homes about 1.3 billion population or about 19% of the world total population.

Research Objectives of the CLHLS

Launch of the CLHLS

While it is very useful and important to uncover the secrets of human longevity to study centenarians, it is also equally important or even more prominent to study the oldest-old population aged 80 or older. This is because the remarkable increase in the number of oldest-old population in the recent years and near future presents a serious public health challenge to promote the quality of life. Because their large consumptions of social and medical care services and benefits of research on them are far out of proportion to their size, the oldest-old population in aging and longevity studies has received increasing attention over the past decades. In this context, Drs. Yi Zeng and James W. Vaupel launched a nationwide project in China on determinants of healthy longevity in 1998, titled as the Chinese Longitudinal Healthy Longevity Survey (CLHLS). This project received financial supports from the National Institute on Aging, the National Natural Science and Social Sciences Foundations of China, UNFPA, and other resources.

The CLHLS aims to collect extensive data on a large sample of the oldest-old aged 80 years and older with a comparison group of younger elders aged 65–79. The project also collected information on the offspring of the elderly in 2002 and 2005 to better investigate the role of intergenerational transfers and its association with human longevity. Starting in 2009, adult children of centenarians and controls of nonrelatives of centenarians in seven longevity areas (later becoming eight longevity areas in 2012 and 2014) were included in the CLHLS (see section “Centenarian Sub-Sample in the CLHLS” below). More specifically, the objectives of the CLHLS research project are threefold: (1) to shed light on the determinants of healthy longevity and to discover social, behavioral, environmental, and biological factors that may have an influence on the healthy longevity of human beings, as well as to answer questions such as why some people survive to very old age without much suffering while others suffer considerably; (2) to fill in the data gap and gain a better understanding of demographic and socioeconomic conditions, as well as of the health status and care-giving needs of the oldest-old population; and (3) to provide a scientific base for sound policy making and implementation, so as to improve the system of care-giving services and, ultimately, the quality of life of the elderly.

Sampling Strategy of the CLHLS

The CLHLS is conducted in a randomly selected half of the counties and cities in 22 of China’s 31 provinces. The 22 provinces are Liaoning, Jilin, Heilongjiang, Hebei, Beijing, Tianjing, Shanxi, Shaanxi, Shanghai, Jiangsu, Zhejiang, Anhui, Fujian, Jiangxi, Shandong, Henan, Hubei, Hunan, Guangdong, Guangxi, Sichuan, and Chongqing (see Fig. 1). The exclusion of nine provinces in the North-West parts of China, where ethnic minorities represent a high proportion of total population, was based on concerns about the inaccuracy of age-reporting among local elders. Previous studies have evidenced major inaccuracy (mainly exaggeration) in age reporting at old ages in these nine provinces (Coale and Li 1991; Huang 1993). In contrast, in the 22 provinces as chosen, local people, mostly Han, tend to use the Chinese lunar calendar and/or Western solar calendar to specify their birthdays, which largely reduces the inaccuracy of age reporting. The accuracy and reliability of age reporting for Han Chinese is related to the fact of their cultural tradition that the exact date of birth is significant for them in making decisions on important life events such as matchmaking for marriage, date of marriage, and the date to start building a house, among other events (Coale and Li 1991; Zeng 2012). The total population of the survey areas constituted about 85% of the total population in China in 2000 and 82% in 2010. So far, seven waves in 1998, 2000, 2002, 2005, 2008/09, 2011/12, and 2014 have been conducted.

Chinese Longitudinal Healthy Longevity Study, Fig. 1
figure 31figure 31

Spatial distribution of the sampled counties/cities in the CLHLS, the 2008 wave. Note: This map was made by the authors based on a county boundary map from the National Bureau of the Statistics of China. The designations employed and the presentation of material on this map do not imply the expression of any opinion whatsoever on the part of the Secretariat of the United Nations concerning the legal status of any country, territory, or area or of its authorities or concerning the delimitation of its frontiers or boundaries

Full size image

In the sampling areas, the CLHLS aims to interview all centenarians who voluntarily agreed to participate in the study. For each centenarian interviewee in each wave, the CLHLS interviewed one nearby octogenarian (aged 80–89 years) and one nearby nonagenarian (aged 90–99 years) with predefined age and sex. “Nearby” is loosely defined – it could be in the same village or in the same street, if available, or in the same town or in the same sampled county or city district. The predefined age and sex are randomly determined, based on the randomly assigned code numbers of the centenarians, to have comparable numbers of males and females at each age group. In the first two waves (1998 and 2000), the CLHLS did not collect data from elders aged 65–79 years. Since the 2002 wave, the CLHLS extended its sample to include elders aged 65–79 under same sampling strategy with approximately three nearby elders aged 65–79 of predefined age and sex in conjunction with every two centenarians. Respondents who were younger than age 100 at an interview but subsequently died before a subsequent wave or resettled or refused to be interviewed at a subsequent wave were replaced by new interviewees of the same sex and age (or within the same 5-year age group). However, such a strategy was not applied to the sixth and seventh waves where only follow-ups were performed due to shortage of budget, except the eight longevity areas where new participants were recruited to replace the deceased or the refusals.

To avoid the problem of small subsample sizes at the more advanced ages, the CLHLS oversampled respondents at more advanced ages, especially among male elders, in addition to recruiting all centenarians with a consent agreement. Consequently, appropriate weights were generated based on the age-sex-rural/urban-specific population distribution in the census. The method for computing the age-sex-rural/urban-specific weights and the associated discussions are presented in Zeng et al. (2008) and available at the CLHLS web page.

The questionnaire design was based on international standards and was adapted to the Chinese cultural/social context and carefully tested by pilot studies. The CLHLS collects various information covering demographics, socioeconomic conditions, psychological traits, health practice, and various health condition. All data were collected via in-home visits. The basic physical capacity tests were performed by a local doctor, a nurse, or a medical student.

Centenarian Subsample in the CLHLS

Subsample of the Centenarian Interviewees

In the research design of the CLHLS, the group of centenarians is one of the major components. As shown in Table 1, the CLHLS from 1998 to 2014 interviewed 10,804 centenarians in total with 2,130 male centenarians and 8,674 female centenarians. The total number of interviews of these centenarians is 16,582, of which 3,876 centenarians have two interviews and 1,360 centenarians have three interviews; only 372, 117, 39, and 14 have 4, 5, 6, and 7 interviews, respectively.

Chinese Longitudinal Healthy Longevity Study, Table 1 Sample distributions of centenarians in the 1998, 2000, 2002, 2005, 2008–2009, 2011/12, and 2014 waves of the CLHLS
Full size table

Data Quality of the Centenarians

Accurate age reporting is crucial in centenarian studies. The CLHLS has employed different methods to verify centenarians’ ages, including birth and marriage certificates if available; household registration information; ages of their siblings, children, and relatives; genealogical record; any relevant document from local communities if available; and reported ages in Chinese zodiac. (The Chinese zodiac is a repeating cycle of 12 years, with each year being represented by an animal according to the Chinese lunar calendar. These zodiac animals are used to record one’s date of birth). Based on the solid comparisons of various demographic indices, it was concluded that although the age reporting quality of centenarians of Han Chinese was not as good as in Sweden, Japan, England, and Wales, it is almost as good as in Australia and Canada, slightly better than in the USA (white, black, and other races combined), and much better than in Chile (see Zeng et al. 2008).

The systematic assessment of data quality of the CLHLS indicates that there was no substantial underreporting of deaths, and most variables or items in the questionnaire were in high quality. However, the causes of death of centenarians reported by next-of-kin might not be reliable, because nearly 60% of reported deaths had no information on causes of death (Zeng et al. 2008). This might be due to that significant portion of the centenarians did not go to the hospital to diagnose/treat the disease prior to death or they in fact died without specific disease.

In-Depth Study of Longevity Areas Including Adult Children of Centenarians

The CLHLS launched a subproject for an in-depth study in seven longevity areas where the density of centenarians is exceptionally high in 2009 as part of the 5th wave of the CLHLS, and in eight longevity areas (the previous seven plus a new one) in 2012 and 2014 as part of the 6th and 7th waves of the CLHLS, to investigate why some areas have a much higher proportion of healthy and long-lived individuals than other areas. The seven areas in 2009 were Chenmai County (Hainan Province), Yongfu County (Guangxi Province), Mayang County (Hunan Province), Zhongxiang City (Hubei province), Xiayi County (Henan Province), Sanshui City (Guangdong Province), and Laizhou City (Shandong Province). Rudong County (Jiangsu Province) was added since 2012. The criteria of section for longevity areas come from the Committee of the China’s Longevity Areas associated with the Chinese Society of Gerontology, including high density of centenarians and nonagenarians, high life expectancy, and a series of within-area consistency checks including good health status and good environment quality, etc. One biological child of each centenarian interviewee in the longevity areas was recruited since the 6th wave. The purpose of such design is to collect data on factors associated with longevity by comparing longevity transmission between families with and without centenarians. In addition to the regular home-interviews, the in-depth study on these longevity areas includes more sophisticated health exams and blood and urine sample collections for biomarker analysis.

In 2002, with support from the Taiwan Academy Sinica and Mainland China Social Sciences Academy, the CLHLS collected a sample of 4,478 adult children aged 35–65 of the elderly interviewees in eight provinces out of the 22 CLHLS sampled provinces: Guangdong, Jiangsu, Fujian, Zhejiang, Shandong, Shanghai, Beijing, and Guangxi (mostly eastern coastal provinces). Of 4,478 dyadic pairs of data, there are 440 pairs for centenarians and their adult children in these eight provinces. Unlike the dyadic pairs of dataset in the longevity areas which deals with familial transmission of longevity, this dyadic dataset focused on the family dynamics of adult children and their intergenerational transferring. One follow-up survey for these 4,478 adult children was conducted in the 2005 wave. Such a study design is rare and valuable, as these dyadic datasets are particularly useful for studying familial factors that are associated with healthy aging.

Deceased Centenarian Interviewees Between Surveys

One unique feature of the CLHLS is the relatively comprehensive information collection on the extent of disability and suffering before dying of each centenarian (also of each respondent of other age groups) who died between two adjacent waves. The information was retrospectively collected from the next-of-kin or the primary caregiver of those deceased centenarians as well as other died respondent. The information includes dates and causes of death, and health and healthcare conditions from the last interview to the time of death, such as chronic diseases, activities of daily living (ADLs), number of hospitalizations, whether the centenarian had been bedridden, and whether the subject had been able to obtain adequate medical treatment when he/she was sick. Data on how many days before death the elder did not go outside and how many days before death the elder spent more time in bed than out of bed were collected as well. Information on socioeconomic and demographic characteristics, such as marital status, family structure, caregivers, financial situation, and living arrangement before death, as well as the caregiving costs within 1 month before the death were also collected.

Table 2 presents the number of the decreased centenarians between two adjacent waves from 1998 to 2014 in the CLHLS, which was 9,853 centenarians with 2,159 males and 7,694 females, for whom the data in the 2 years prior to death have been collected.

Chinese Longitudinal Healthy Longevity Study, Table 2 Distributions of deceased centenarians between adjacent waves from 1998 to 2014, CLHLS
Full size table

DNA Samples and Home-Based Health Examinations

The CLHLS collected DNA samples from 4,849 centenarians in addition to 5,190 nonagenarians, 5,274 octogenarians, 4,770 aged 65–79, and 4,609 aged 40–64. Health exams for a total of 2,035, 2,862, and 2,651 participants in the longevity areas were performed in 2008/09, 2011/12, and 2014, respectively, by local certified doctors and nurses who are affiliated with the China Center for Disease Control and Prevention (CDC) as contracted for this project. The medical personnel used standard instruments to check heart, lungs, breast, waist, lymph, limbs, and thyroid of the participants. They also wrote down impressions and symptoms of disorder if any, and furthermore enquired about the participants’ family disease history and current medications.

In sum, the large population-based sample size, the focus on healthy longevity (rather than on a specific disease or disorder), the simultaneous consideration of various risk factors, and the use of analytical strategies based on demographic concepts make the CLHLS as an innovative project of demographic data collection and research (Zeng 2012).

Psychological Traits of Chinese Centenarians

Variables of Psychological Traits

In addition to the internationally standardized mini-mental status examination (MMSE) of cognitive function tests, the CLHLS contains seven variables relevant to psychological traits: (1) Do you look on the bright side of things? (being optimistic) (2) Do you keep things neat and clean? (3) Can you make your own decisions concerning your personal affairs? (self-determination) (4) Do you feel as happy as when you were young? (5) Do you feel fearful or anxious? (6) Do you feel lonely and isolated? (7) Do you feel useless? Each question above has six response options: always, often, sometimes, seldom, never, and unable to answer; proxy responses were not allowed. The first four questions reflect positive affect of psychological traits, while the latter three questions refer to the negative affect.

These questions are mainly derived from the Positive Affect and Negative Affect schedule (PANAS) scale and could also be considered a short version of a recently developed Scale of Positive and Negative Experience (SPANE). Both PANAS and SPANE scales mainly focus on the general adult population (see Diener and Biswas-Diener 2009). Different from the SPANE and PANAS scales, psychological traits questions in the CLHLS contain an option “unable to answer” for each question, which aims to accounting for the possibility that some oldest-olds may not be able to answer the question due to, for example, various health problems or difficulties in making up their minds. Based on the CLHLS data from the 1998 wave to the 2011/12 wave, this has been well justified by the fact that about 75% of the oldest-old respondents who were unable to answer these questions were due to health problems (this proportion was about 95% among centenarians who were unable to answer).

In order to better quantify the contribution of these psychological traits to exceptional longevity, these seven variables were dichotomized (coding 1 for answering “always” and 0 otherwise for positive affect called as “always positive affect”, whereas coding 1 for answering “never” and 0 otherwise for negative affect called as “never negative affect”) and then generated an index of always positive and never negative affect (abbreviated as APNNA) by summing these seven dummies, which ranges from 0 to 7. Because the wording of psychological traits questions in the 1998 wave is slightly different from that of the other waves and because the 2014 wave is not publicly available yet, in this section the focus of analyses of psychological trait of Chinese centenarians was on the waves from 2000 to 2011/12.

Positive and Negative Affect in Centenarians

Figure 2 shows that there was a clear decreasing trend with age in the score of the APNNA index. The overall mean scores of the index in centenarians were significantly lower than those in other age groups (Fig. 2). However, when demographics (age, urban–rural residence), socioeconomic status (education, primary lifetime occupation, economic independence), family and social support (marital status, coresidence with children), health practice (smoking, alcoholic intake, exercising), and health condition (ADLs, instrumental ADLs, cognitive function) were controlled, the pattern was reversed (results not shown). That is, centenarians had the highest mean scores of the APNNA, followed by nonagenarians and octogenarians, whereas the elders aged 65–79 had the lowest mean score. The difference between centenarians and older adults aged 65–79 was significant (p < 0.01) for males but not for females.

Chinese Longitudinal Healthy Longevity Study, Fig. 2
figure 32figure 32

Mean scores of the APNNA index and their 95% confidence intervals for centenarians by sex in comparison with other ages, CLHLS 2002–2011/12. Note: The score of the APNNA index ranges from 0 to 7, which includes four always positive affect variables and three never negative affect variables

Full size image

Table 3 reveals that with few exceptions (e.g., self-determination (column 3) and loneliness (column 6)), the difference in percentage was not large for both male and female centenarians in comparison with their younger counterparts. Furthermore, Fig. 3 reveals that centenarians were likely more psychologically robust in terms of APNNA than their same cohort peers who died between ages 90 and 99. These results suggest that presence of positive affect and absence of negative affect among centenarians may have contributed to their exceptional longevity.

Chinese Longitudinal Healthy Longevity Study, Table 3 Percentage distribution of always positive affect and never negative affect among centenarians by sex in comparison with other age groups, CLHLS 2000–2011/12
Full size table
Chinese Longitudinal Healthy Longevity Study, Fig. 3
figure 33figure 33

Mean scores of the APNNA index by birth cohort, sex, and whether the respondents interviewed in 2000 who survived to age 100 (as indicated by yes for survival to age 100 and no for those who deceased before age 100) from 2000 to 2011/12. Note: (1) The score of the APNNA index ranges from 0 to 7, which includes four always positive affect variables and three never negative affect variables. (2) n, sample size. (3) Only the results from birth cohorts 1901–1904 were significant

Full size image

Psychological Traits and Healthy Longevity

Table 4 shows that each additional increase in the APNNA index (or each additional positive possession out of the seven psychological traits) among centenarians reduced mortality risk by 4–5% (Model I to Model III). Even after controlling for baseline health (function in activities of daily living, cognitive function, and chronic conditions), such a protective effect of psychological traits was still significant, although the effective size was reduced to 2% (Model IV). The protective effect of psychological traits on mortality was larger in other ages: the younger the age group, the greater the protective effect of psychological traits. The results further reveal (not shown) that the protective effect of psychological traits on mortality were the same for both males and females and for both centenarians and other age groups.

Chinese Longitudinal Healthy Longevity Study, Table 4 Relative mortality hazards of the APNNA index of centenarians in comparison with other age groups, CLHLS 2000–2011/12
Full size table

The left panel in Table 5 further reveals that each additional increase in the APNNA index among female centenarians reduced the odds of onset of ADL disability by 6%. Such a reduction was similar to those in female nonagenarians and octogenarians and persists even after controlling for a rich set of covariates. However, such a reduction was not significant in male centenarians, possibly due to the smaller sample size. The reduction in other, younger age groups of males was significant and mostly larger than in the corresponding age groups of females. The right panel of Table 5 shows that the reduction in onset of cognitive impairment due to one additional point of the APNNA index was about 11% in female centenarians and did not change in presence of covariates. Females in other age groups had a similar pattern. In contrast, for male centenarians, such a reduction was slightly larger than in female centenarians and other age groups of males, although the significance was weakened when covariates were added.

Chinese Longitudinal Healthy Longevity Study, Table 5 Odds ratios of onsets of ADL disability and cognitive impairment for the APNNA index for centenarians by sex in comparison with other age groups, CLHLS 2000–2011/12
Full size table

Concluding Remarks

Using data from more than 10,000 centenarians of mainland China, the largest centenarian sample in the contemporary world, this entry presents a summarized introduction of the CLHLS and a brief description on psychological traits of centenarians in comparison with other older adults. We find that centenarians were more psychologically robust than noncentenarian peers of the same birth cohorts when they were all in ages of 90s and further report a significant association between possession of positive psychological traits and mortality and health worsening in centenarians. These findings suggest that centenarians are better able to handle stress, depression, or other unfavorable condition than their cohort peers, which is in line with many other centenarian studies and that maintaining a good psychological well-being is an important pathway to reach age 100 (e.g., Gondo et al. 2006; Jopp and Rott 2006; Perls 2006; Poon et al. 2010).

The findings of the present study are also similar to one recent study by Zeng and Shen (2010) that applied a concept of psychological resilience to Chinese centenarians based on questions (1), (3), (5), (6), and (7) in section “Variables of Psychological Traits” and two other variables in the CLHLS (to whom the respondent usually talks most frequently in daily life? and who does the respondent ask first for help when having problems/difficulties?). That study reports that centenarians are more psychologically resilient than elders of young ages and that psychological resilience positively contributes to exceptional longevity. Indeed, when further accounting for the response option “unable to answer” for these questions, one recent study found that there are still about 6–9% of centenarians whose psychological well-being is as good as those elders aged 65–79 years (Gu and Feng 2016).

Recently, there is a call among scholars in studies of exceptional longevity, emphasizing the importance of both quantitative and qualitative methodologies, replication of mechanisms, interdisciplinary and systems perspectives, and generalizability of results (Poon and Cheung 2012). Among these new directions of the future studies in centenarians, scholars particularly concern about instable results from small sample sizes (Willcox et al. 2006). Moreover, more studies are encouraged to examine the association between psychosocial traits and longevity, relative to biologically based studies (Poon and Perls 2007). This entry uses the large-sized sample of the CLHLS to investigate how psychological traits are associated with longevity and subsequent health condition, which echoes the new initiatives above and adds new evidence highlighting the importance of psychological factors to exceptional longevity.

However, due to space limit, details about the role of other psychosocial factors to longevity were not discussed, although many of them were already included in the models as covariates. The interactions between psychological traits, environmental factors, and genetics in determining longevity were also not investigated. As Poon and Cheung (2012) pointed out, to eventually unearth the secrets of longevity, there is still much unexplored on what, how, and why some individuals survive to age 100 with good health. By the end of 2015, only a very small portion of studies focuses on centenarians out of 450 peer-reviewed publications in English, Chinese, and 76 Ph.D. and M.A. theses/dissertations that used the CLHLS data since 1998.

One limitation of the present study in analyzing centenarians’ psychological traits and its association between subsequent survival is the way of coding of those who were not able to answer the psychological trait questions. In the case of always positive affect, they were classified into the group of those who did not always experience positive affect. In the case of never negative affect, they were classified into the group of those who never experienced negative affect. Such a coding system may somewhat underestimate the psychological traits in centenarians. As a consequence, the association between good psychological traits and subsequent survival among the centenarians may be somewhat biased. Nevertheless, as the majority of these respondents were in a very poor health condition, such biases would be only mild. Some researchers adopted an alternative approach by excluding those who were unable to answer the questions in the analyses (e.g., Zeng and Shen 2010). However, since those who were not able to answer the questions were not missing at complete random, the exclusion approach may overestimate centenarians’ good psychological traits to some extent. More research on this issue is clearly warranted.

In sum, more studies on centenarians are warranted in this field, and the CLHLS has been becoming an important resource for scholars in this field with a large and representative sample size of respondents at extremely old ages in a longitudinal context plus the voluminous psychosocial and biological data.

Cross-References