18
Similar findings obtain for Democrats when we examine the appeals of both
parties on issues owned by the Democratic Party. The distribution of this variable differs
somewhat from the volume variable even though they are highly correlated. The range is
somewhat truncated, going from -506 (Albuquerque-Santa-Fe, late September/early
November) to 520 (Philadelphia, early September). On average Democrats make more
appeals on these issues than Republicans (µ=-8.57, s=117.19, 75
th
percentile=7). Under
these circumstances, the impact of moving from the minimum to the maximum decreases
the probability all Democrats support Gore from roughly 91% to about 79%. The decline
in Gore support is even more substantial among peripheral Democrats, moving from
roughly 89-percent under Democratic-favoring conditions to 75-percent when the
Republicans make considerably more appeals on Democratic issues.
Although these effects are not by themselves determinative of election outcomes,
they are fairly substantial. First, it is quite possible that these effects are in part
underestimates. This is because there is substantial similarity in the number of appeals
that both parties offer on Democratic and Republican-owned issues, particularly when the
level of advertising becomes intense. Such a situation potentially masks important
dynamics in the voter distribution by supplementing losses from one group with gains by
the other. Second, campaign effects are often quite small relative to the impact of
individual-level characteristics, though the ones here are still fairly substantial. For
example, among Democrats the impact of moving from being very liberal to very
conservative is to decrease the probability of a Gore vote by 38%. The impact of
campaign advertising is roughly 1/3 and 1/2 (depending upon the circumstances) the
Convention