Turning to the duration models, Table 4 presents the results of pooled time-series
poisson regressions where the dependent variable is a count of the number of continuous years of crisis The table entries are incidence-rate ratios. Ratios less
than 1 reflect a negative effect of the variable of interest; ratios above 1 reflect
positive effects. Three general observations are in order. First, in comparison with
the results reported in Table 2, our political variables provide less insight into the
factors that lengthen crises than the emergence of a crisis in the first place. Fewer of
the political variables achieve statistical significance, and as will be seen below,
when they are significant, their effects are less profound than in the crisis event
models.