This time, I would like to introduce two notable trends emerging in the field of internationally comparable panel surveys of middle-aged and elderly people. A key characteristic is that both help enable a more precise comparison of individuals using the same measure. The first emerging trend is the use of biomarkers. Respondents are asked not only to answer the questionnaire but also to provide their specimens by consent. This may include measuring the height, weight, and blood pressure of each respondent as well as collecting blood and saliva samples for medical analysis.
As one such practice, the measurement of handgrip strength is included in many surveys because of some research findings showing a significant correlation between handgrip strength and the probability of survival. In the case of the English Longitudinal Study of Ageing (ELSA), a panel survey conducted in the United Kingdom, interviewers are accompanied by nurses who draw blood samples for componential analysis and/or saliva samples to examine the level of cortisol, a hormone closely linked to depression. Meanwhile, the newest attempt is to collect DNA information. Although it is still in the trial-and-error stage, massive investments have been made to find which genes affect people's behavior.
The collection of biomarkers involves significant challenges. In addition to the costs involved, respondents' consent needs to be obtained, and various ethical issues—such as those regarding the use of data obtained from biomarkers—must be cleared. Yet, vigorous efforts are being made overseas to proceed with such initiatives. Indeed, biomarker data have been long utilized extensively in medical and epidemiological surveys. It would be useful to utilize the know-how accumulated in these fields. The use of biomarker data is crucial to the objective measurement of health conditions. It is thus necessary to make efforts to promote public understanding of the need to collect such information. As part of the Japanese Study of Aging and Retirement (JSTAR) survey, we ask relevant municipalities to provide annual health checkup data, subject to consent from respondents concerned. Such data can be combined with those obtained through the main JSTAR questionnaire for analysis.
The second trend is to control for individual differences in perceptions in conducting attitude surveys, which include many questions asking respondents about their "feeling." A typical question would ask respondents to assess the current state of their health on a scale of one to five. It is surely important to let respondents answer subjectively. However, some people might describe their health conditions as "not so good" simply for having a slight cold, whereas others who are hospitalized for serious illness might say their conditions are "good." As such, it would be difficult to make comparison across individuals (and across countries). Thus, as a way to avoid this problem, a new approach, anchoring vignette, which uses hypothetical questions is drawing much attention. A questionnaire under this approach would start with a question asking respondents how they would assess their health status in a specific hypothetical situation (e.g., Person A has been suffering from headaches about once per month, but the headaches are manageable with medicine and do not disturb normal daily activities), that is, if they were Person A. The questionnaire would then ask respondents about their respective personal conditions, which would be controlled for personal differences based on their answers to the hypothetical question. Currently, vigorous efforts are being made to search for appropriate hypothetical situations that can effectively isolate differences in response style across individuals.
* Translated by RIETI from the original Japanese "Yasashii Keizaigaku" column in the September 22, 2011 issue of Nihon Keizai Shimbun.
