Secular mobilization?

According to the General Social Survey, from 1985 to 2004 the proportion of Canadians 15 or older without a religious affiliation jumped 7 points to 19%, and the share of the affiliated who had not attended any religious services in the previous year climbed from 19% to 25% (Clark and Schellenberg 2006). [...] Table 2 displays both the gross effect of religion on protest involvement in a comparison of 1990 and 2000 to 1981, and, in the columns on the right, those effects in a comparison of 1990 and 200015. [...] For example, the logit coefficient for ‘Religious x 1990’ in Table 3, as the estimated difference in the 1981 and 1990 slopes, can be added to the coefficient for religiousness in the same column, as the effect if religiousness in 1981, [.03 + (-.13)] to produce the logged odds of protest participation in 1990, which is reported at the top of the fourth column (with rounding error). [...] While its effect fails to reach significance in 1981 or 1990 (as evidenced at the top of the second and fourth columns in Table 4), adding the relevant increments estimated in the interaction terms to determine its 2000 effect (and reversing the coding for time to determine the logit’s significance) reveal a shift: weekly or more frequent attendance to religious ceremonies in 2000 reduces the odds [...] Significant changes in the effects of social conservatism and religiosity, finally, were only detected in the expanded net model, suggesting the utility of a decomposed measure of religiousness in a multivariate analysis.