Replication data for: Reply to "Shirking in the Contemporary Congress: A Reappraisal"

In adding a fixed-effects component to our original model, Carson et al. (2004) suggest that our findings of shirking by departing legislators are spurious. While we might debate whether we require Congress-specific effects for our analysis, we will adopt another approach here. We show that properly incorporating fixed effects using a heteroskedastic regression model, controlling for Congress-specific changes in both the mean and the variance of ideological change, again uncovers shirking.