7
on people’s attitudes and behaviors. Much of the work on social identity has implicitly assumed
that everyone who identifies with a group does so equally. People either identify with a group or
they do not. In part this lack of interest in identity strength stems from social identity theorists’
emphasis on context and the idea that salient identities simply reflect the situation at hand. A
person can identify as a woman in one situation and as a Twins fan in another. Identities are
fluid. But this approach to understanding group identity ignores individual differences in the
extent to which people identify with a particular group, what Huddy (2001) calls “shades of
identity” and Ellemers, Spears, and Doosje (1999b) call commitment. As Deaux (2000: 5)
argues, “Although people may share a common cognitive category, their identification with the
category can vary substantially, and these variations have important consequences for
behaviour.”
In a laboratory setting in which subjects are assigned to a group randomly or based on the
most minimal of reasons (for example, overestimating or underestimating the number of dots on
a screen), which is where social identity theory has its roots, group identity may well be fluid and
context dependent (Tajfel et al. 1971). But Huddy (2001) convincingly argues that many political
identifications are long-standing and quite stable, and when an identity is long-standing and
stable, it is more likely to be integral to a person’s sense of self and to be accessible across many
different contexts, even when the fit is poor. Strong party identifiers or strong ideologues
maintain their identifications over much of their lifetime (Jennings and Markus 1984) and these
identifications affect a variety of attitudes and behaviors. These are not transitory group
identities dependent only on context for their existence. In fact, people who strongly identify with
a group are less affected by context than people who only weakly identify with a group (Kinket
Convention