By their very nature cross-national investigations lend themselves to searching for
contextual effects. Indeed, the very concept of a civic culture in Almond and Verba’s classic
work (1963, ch. 11) was predicated in part on the existence of different socialization contexts
across their five-nation study. In the same vein, small scale studies of developed (e.g., Adelson
and O’Neil 1966) and developing countries (e.g. Koff, von der Muhll, and Prewitt 1973) took
contextual features into account. More studies have emerged in recent times (e.g., Macek,
Flanagan, et al., 1998, Torney-Purta, Barber, and Richardson 2004). Perhaps the most self-
conscious line of cross-national research incorporating contextual features deals with the
intergenerational transmission of partisanship, pioneered by Converse and Dupeux (1962), and
followed up by, inter alia, Percheron and Jennings ( 1981) , Westholm and Niemi (1999) , and
Ventura (2001).
Small scale studies, whether in single or multiple countries, can provide an intimate,
process-oriented look at contextual effects. However, they are limited in terms of generalization
and also often suffer from an inability to specify the effects of particular contexts over and above
or interacting with individual characteristics of those being socialized. Political socialization
inquiries have recently begun to employ multilevel models in analyzing larger-scale databases in
an effort to specify and understand the contribution of contextual features. Given the presence of
nested designs, commonly employed in studies of pre-adults wherein respondents are clustered
with schools or classrooms, multilevel modeling is the most appropriate statistical approach.
Two recent reports about adolescents in the United States exemplify this development.
Both move beyond using individual student and familial characteristics as determinants of
socialization outcomes by employing features of the communities and schools in which the
students were ”nested.” Gimpel, Lay, and Schuknecht (2003) utilized census and electoral data to
characterize the school catchment areas and hence the socio-political contexts in which the
students lived. Their results indicate that socio-political diversity elevated information holding
and participation while homogeneous and uncompetitive environments dampened various
2
Convention