The institutions which drive social development are by nature difficult to detect, given that they rest upon tacit norms, beliefs, and practices which lack explicit formalization. Previous quantitative studies of social institutions have therefore largely relied upon using either proxies based upon causes or consequences (such as using daily newspaper circulation as a proxy for the extent to which citizens take an active interest in local politics, or linguistic fractionalization as a proxy for cohesion or otherwise between social groups), or survey responses to questions regarding social attitudes.
Not all survey data is perceptionsbased, however, and can often be behavioral, as when respondents are asked whether they have been the victim of crime, whether they have signed a petition, or whether they have contacted a local representative. Both proxy variables and survey items are used in these indices, and both correlate to an exceptionally large extent.
For example, a country’s reported level of social trust is strongly predicted by a country’s homicide rate, while the correlation between the proportion of managers who say men have a greater priority than women to a job, and the ratio between male and female labor force participation, is likewise high. To some extent, this reflects the fact that perceptions and attitudes are not simply the result of social institutions, but are the institution, to a substantial degree.