Abstract: Comparing the wellbeing of groups using self reported measures of wellbeing can be challenging. The scale dependency of many summary statistics applied to arbitrary Cantril scales attached to ordinal categorical data can engender a lack of coherency in results based upon alternative, equally valid scales. Furthermore, the conditions under which results will be robust across alternative scales seldom prevail in practice. Here scale independent methods for the multilateral and multidimensional wellness measurement and comparison of groups are proposed and implemented in a study of the health-loneliness-aging-gender nexus in 21st century China. The results indicate that poor health and loneliness appears to increase with age, though not monotonically. Improved health status is always associated with better un-loneliness outcomes and improved un-loneliness status is always associated with better health outcomes. While a large portion of the population are not affected by loneliness, of those who are, ill health is generally more likely to be reported. With regard to the health - loneliness joint distribution, generally, males enjoy better joint outcomes than their female counterparts in almost every comparison and urban dwellers enjoy better outcomes than their rural counterparts.
Keywords: Wellbeing Measurement, Ordering Distributions, Ordinal Data.
JEL Classification: C14; I14; I30; I31