are negative in all models in each survey year. Those who have recently expe-
rienced trauma and feel that life is hopeless and nothing can cheer them up
are not the people who are attending religious services. At the same time,
those attending church are less likely to give up hope and more likely to
report higher levels of subjective well-being. People who join others in reli-
gious activities are generally more likely to believe in life after death. Such
beliefs are likely to promote provisions of meaning that afford people the
opportunity to place their lives in a context that makes them feel comfort-
able. The question that we cannot answer in this analysis is which comes
first. Does participation over the life course (or some part of it) buffer the
effects of trauma when it occurs and hence provide a framework for coping
with the problem? If this were the case, then we would expect continued
attendance at the onset of various negative life events. However, higher atten-
dance rates may not be the case if the order is reversed. That is, do people
attend religious services as a reaction to psychological trauma? Our results
suggest that those who experience negative psychological experiences do not
necessarily turn to places of worship as coping mechanism or reaction to
these negative life events. Instead, they appear less likely to attend.
Clearly, employing data such as the General Social Surveys has its strengths
and limitations. The GSS is one of the few data sources that addresses a wide
range of issues and taps a number of attitudinal and behavioral dimensions.
The GSS includes a number of items concerning religiosity, subjective well-
being, and other sociodemographic information. The GSS also spans an approx-
imate twenty-year time period and affords opportunity to examine trends in
attitudes across these years. This analysis addresses attendance at religious
services as an indicator of the participatory dimension of religiosity. The same
question is asked in all years included in the study. We are also able to tap
the psychological trauma dimension across this time period as well. While a
slightly different measure is employed in 1998 than in 1978 and 1988, the two
measures are highly correlated, indicating that both measures are address-
ing similar concerns. Finally, the GSS includes consistent items across the
years concerning sociodemographic characteristics.
There are of course limitations to using these types of data. Our depen-
dent variable is limited to attendance at religious services as a proxy for pub-
lic participation in church or synagogue. Other items that tap ideational or
devotional dimensions were not available in all survey years. In addition, the
combination of psychological trauma and religiosity items were not included
in the same years or asked to the same respondents.
Operationalizing the Critical Theory of Religion • 351