Reference Groups Manajemen | Fakultas Ekonomi Universitas Maritim Raja Ali Haji 591.full

where µ r and σ r are the reference group mean and standard deviation. While this meas- ure captures distance, it is different from the Yitzhaki-based measures in that it is sen- sitive to changes in the income distribution below the individual’s income. Further, while average Yitzhaki-based relative deprivation increases as income inequality goes up, the mean and standard deviation of z-score are insensitive to changes in the distri- bution of income.

IV. Reference Groups

In order to address the relative deprivation hypothesis, one must con- sider how individuals define reference groups. The social psychology literature sug- gests that members of one’s reference group are typically selected on the basis of either similarity or geographic proximity Singer 1981. While geographic proximity is relatively easy to determine, “similarity” is a more nebulous concept. Various stud- ies report that individuals define reference groups along demographic lines such lines as sex, education, and race Merton and Kitt 1950; Singer 1981; Bylsma and Major 1994. However, it is well acknowledged that there is no perfect formula for deter- mining reference groups. Critics assert that the “Achilles heel” of social evaluation theory is the “failure to explain adequately how the relevant comparisons are selected in the first place” Pettigrew 1978; p. 36. Perhaps because of the difficulty of determining reference groups according to “similarity,” most studies dealing with health and inequality define relative depriva- tion within the context of geographical location. U.S.-based studies of income inequality and health typically use state or MSA of residence as the implicit reference group. Restricting inequality measures to geographic boundaries makes sense if we expect that inequality affects health through its impact on public investment in human and social capital. However, if Wilkinson’s psychosocial pathways are the more prob- able culprits, then it’s not clear that reference groups should be limited to geographi- cal confines. Individuals may compare themselves to others of similar demographic backgrounds, regardless of geographical location Frank 1985. Deaton 1999 addresses the issue of “similar circumstances” by using birth cohorts to define refer- ence groups. In this study, we construct reference groups based on observable demo- graphic characteristics such as state of residence, race, education, and age. Groups defined using such characteristics do not necessarily constitute the unobservable true reference groups. Yet, members of such groups have a high degree of similarity and are likely to contain a high proportion of relevant reference people.

V. Empirical Analysis