Promises and Pitfalls of Anchoring Vignettes in Health Survey Research

Data harmonization is a topic of growing importance to demographers, who increasingly conduct domestic or international comparative research. Many self-reported survey items cannot be directly compared across demographic groups or countries because these groups differ in how they use subjective response categories. Anchoring vignettes, already appearing in numerous surveys worldwide, promise to overcome this problem. However, many anchoring vignettes have not been formally evaluated for adherence to the key measurement assumptions of vignette equivalence and response consistency. This article tests these assumptions in some of the most widely fielded anchoring vignettes in the world: the health vignettes in the World Health Organization (WHO) Study on Global AGEing and Adult Health (SAGE) and World Health Survey (WHS) (representing 10 countries; n = 52,388), as well as similar vignettes in the Health and Retirement Study (HRS) (n = 4,528). Findings are encouraging regarding adherence to response consistency, but reveal substantial violations of vignette equivalence both cross-nationally and across socioeconomic groups. That is, members of different sociocultural groups appear to interpret vignettes as depicting fundamentally different levels of health. The evaluated anchoring vignettes do not fulfill their promise of providing interpersonally comparable measures of health. Recommendations for improving future implementations of vignettes are discussed.

Health systems' responsiveness and reporting behaviour: Multilevel analysis of the influence of individual-level factors in 64 countries

Health systems' responsiveness encompasses attributes of health system encounters valued by people and measured from the user's perspective in eight domains: dignity, autonomy, confidentiality, communication, prompt attention, social support, quality of basic amenities and choice. The literature advocates for adjusting responsiveness measures for reporting behaviour heterogeneity, which refers to differential use of the response scale by survey respondents. Reporting behaviour heterogeneity between individual respondents compromises comparability between countries and population subgroups. It can be studied through analysing responses to pre-defined vignettes - hypothetical scenarios recounting a third person's experience in a health care setting. This paper describes the first comprehensive approach to studying reporting behaviour heterogeneity using vignettes. Individual-level variables affecting reporting behaviour are grouped into three categories: (1) sociodemographic, (2) health-related and (3) health value system. We use cross-sectional data from 150,000 respondents in 64 countries from the World Health Organization's World Health Survey (2002-03). Our approach classifies effect patterns for the scale as a whole, in terms of strength and in relation to the domains. For the final eight variables selected (sex; age; education; marital status; use of inpatient services; perceived health (own); caring for close family or friends with a chronic illness; the importance of responsiveness), the strongest effects were present for education, health, caring for friends or relatives with chronic health conditions, and the importance of responsiveness. Patterns of scale elongation or contraction were more common than uniform scale shifts and were usually constant for a particular factor across domains. The dependency of individual-level reporting behaviour heterogeneity on country is greatest for prompt attention, quality of basic amenities and confidentiality domains.