What kind of social bonding are you referring to that atheists and agnostics have? I'm not trying to be insulting. I agree they have a strong identity. But from my experience they tend to be more individualistic, willing to stand apart from groups for the sake of "the truth." IOW there really isn't anything in atheism or agnosticism that is comparable to the religious community to be bonded to. So I'm wondering what you are referring to instead.
I'll copy and past a large segment, because there is evidence to back it up. Of course, this should be taken, like all studies, with scepticism.
Cameron (2004) outlines three non-mutually exclusive factors that constitute social identity. The
first factor, centrality, is defined as the frequency at which the given group “comes to mind” in daily
interactions and the subjective importance of the group to one’s self-concept. In a recent qualitative
assessment of the identity formation process of American atheists, Smith (2011) found that social
identification as a non-believer is an important aspect of crystallizing one’s identity, and that as part of
the process of crystallization, the atheist identity is moved to the top of the identity hierarchy. In accord
with Stryker’s (1968) position that higher-order identities are of greater importance to definitions of
the self, understanding of social events and processes, and shaping behaviors, Smith’s (2011) research
clearly indicates that the atheist identity is one of the (if not the) most salient identity for non-believers.
Further, “coming out” as a non-believer, similar to the process of “coming out” as a homosexual
(O’Brien 2004), creates social pain and discomfort that further solidifies identity and serves as a
constant reminder of one’s deviant status (Smith 2011).
In-group affect encompasses the emotional aspects of group identity (Cameron 2004; Rosenberg
1979). Tajfel and Turner (1979) claim that emotional connections associated with group membership are
central to social identity theory, as positive or negative evaluations of group membership structure the
degree of affective ties to the group, group identity salience, and related behavioral outcomes. Often
referred to as private collective self-esteem (see Luhtanen and Crocker 1992), in-group affect refers to the
subjective emotional evaluation of group membership (i.e., “I feel good about the religious group I belong
to”; see Ysseldyk et al. 2011). In a study of religious identification among American Atheists, Catholics,
Jews, Muslims, and Protestants, Ysseldyk et al. (2011) found that atheists express statistically similar
levels of private collective self-esteem when compared to the other aforementioned religiously oriented
groups. In other words, atheists generally “feel good” about their group membership, which falls in line
with the degree of in-group affect expressed by members of established religious groups.
Lastly, in-group ties reference “the psychological ties that bind the self to the group” (Cameron
2004, 242). This emotional closeness to the group, or group cohesion, reflects a sense of belonging
and bonding for group members (Brown et al. 1986; Cameron and Lalonde 2001; Phinney 1992).
When using the “feeling thermometer” (see Cairns et al. 2006) to measure in-group ties among
atheists and established religious groups, Ysseldyk et al. (2011) found that non-believers showed the
highest level of emotional closeness to their group and achieved a level that was higher than all of the
other religious groups questioned (i.e., Catholics, Jews, Muslims, and Protestants).
Taken together, the evidence suggests that non-believers hold a strong social identity and
likely experience benefits from their strong group identification, similar to the benefits of strong
group identification for religious believers (see Baumeister and Leary 1995).
So, part of their identity is atheism or agnosticism and it's quite strong. However, this is not good enough to explain why atheists or agnostics do not have similar crime association to the uncertain believers, but it's beyond the scope of this study. I'm sure there are numerous explanations why and perhaps uncertain belief is the cause of something else, like trauma. Hence, drifting like this may be a coping mechanism. Just as a note, social identity theory or social control theory was not measured in this study.
