This actually depend on campaigns to help them

This observation brings us to a
crucial point: vote choice seems to coincide
with both partisan and non-partisan
variables (as shown by the NES, Holbrook,
and Fair), but campaigns “matter” because
they create and reinforce voters’ impressions of candidates, and thus help shape their
choices. As Just and her colleagues write,
“One of the most striking conclusions that
one takes away from the in-depth interviews
is that it is impossible to categorize the
individuals in our panel as party voters,
issue voters, personal voters, and so forth.
Instead, individuals have a wide range (if
not always a large amount) of information
about the candidates, which they try to
integrate into a meaningful whole.”20 In
essence, they argue that voters accept hordes
of information from local and cable news,
the internet, advertising, and campaign
events, and they use it to infer one aspect of
a candidate from another. For example,
some voters may deduce a candidate’s issue
stances from his partisan affiliation, but
others critique a candidate’s abilities as a
campaigner. Of particular importance for
these voters is a candidate’s commitment to
a particular problem, such as the economy.
From his willingness to campaign on an
issue which, as we have seen, is salient for
many voters, one might make judgments
about a candidate’s character and ability to
manage the economy to that voter’s
benefit.21 Campaigns therefore provide the
information necessary to make both
retrospective and prospective evaluations.
In-depth interviews have proven that voters
do not simply “sense” things like the
direction of the economy, as Fair suggests,
but that they actually depend on campaigns
to help them make their decisions.22
This argument rejects another of
Fair’s assumptions, that campaigns are
always on equal footing and thus “cancel
each other out.” In reality, campaigns
convey certain messages, which are filtered
through the media and evaluated by voters.
This argument rests on its own critical
assumption, that information-gathering
influences vote choice. Holbrook, in
rejecting his own model, presents some
compelling evidence to suggest that it does.
Specifically, he cites Popkin’s work, which
says that voters demand information from
campaigns because they pay very little
attention to public affairs between elections.
Holbrook also draws on Gelman and King’s
argument that repeated swings in public
opinion indicate that voters continually
analyze campaign events, and that their
choices are therefore informed, not simply
subconscious expressions of preexisting
realties. Because their choices are informed,
they are somewhat predictable. Campaign
events, however, reduce elections’
predictability because they can sway public
opinion inequitably, especially in
battleground states.23 As we have seen, in
the 2000 election, both individual and
aggregate data indicated that Gore should
have won a close race. Instead, he lost
despite winning the popular vote, which
cannot be explained by forecasting models
but is explained by campaign and Electoral
College effects.
The fact remains, though, that
forecasting models are surprisingly accurate,
and while they may neglect a few critical
percentage points and some institutional
factors, a stronger argument is required on
campaigns’ behalf. One of Holbrook’s final
conclusions explains the success of these
models, and why it does not disparage the
role of campaigns. According to Holbrook,
campaigns are as processes which generate a
product, because, as we have seen, they
clearly influence public opinion and allow
voters to make choices based on rational
considerations.24 The positive, though
imperfect, correlation of non-campaign
factors with electoral outcomes only proves
that campaigns are successful in forcing
voters to use these rational considerations,
and that other, less-predictable factors are at
play. The work of Just and her colleagues
helps clarify this point, that presidential