To evaluate an electoral system
under some set of counterfactual conditions (such as if incumbents did not run
for reelection or if campaign spending was the same for each party), follow the
directions under Evaluation, in the previous paragraph. Then (without
specifying a new YVOTE command) add an XNEW command following XVARS to indicate
the factual or counterfactual conditions you wish to analyze. For example, say
you wanted to gauge the partisan bias in an electoral system independent of
each party's fundraising ability. You might run an initial evaluation with an
explanatory variable (input with XVARS) representing the difference in spending
between the two parties' candidates. Then you would run the counterfactual
evaluation, with XNEW to replace the spending variable with a vector containing
all zeros (so that the analysis will assume no differences in funding levels).
Other variables would be identical in XVARS and XNEW. See below for the
commands used to choose one or more desired forms of statistical output.
