The presidential nomination process continues to change, and so too does our understanding of which factors drive nomination outcomes. The most recent changes have come in the form of drastic adjustments to the schedule of state contests. Nomination campaigns have become increasingly front-loaded, compressed, and nationalized (Mayer and Busch 2004), making the preprimary campaign, or what is also termed the invisible primary or the exhibition season, a powerful force in the selection of party nominees for president. Indeed candidates for nomination often withdraw in this period, and party nominations are clinched within, what was until recently, mostly shorter and shorter time periods (Haynes et al. 2004; Norrander 2006). Recent scholarship now emphasizes that a candidate's level of funding and national political support at the beginning of the primary season largely determine who wins the nomination (Adkins and Dowdle 2001, 2005; Cohen et al. 2008; Mayer 1996, 2003; Steger et al. 2004; Steger 2007). As a result, long shots have less of a chance to compete with early front-runners (Steger 2000), leading some scholars to conclude that parties have essentially turned back previous electoral reforms (Aldrich 2009; Cohen et al. 2008).
While the national invisible primary appears to hold greater importance, the role of the traditional early state bellwethers, Iowa and New Hampshire, is now in doubt. Recent analyses of what predicts nomination success have painted the contribution of New Hampshire as small and Iowa as nonexistent (Adkins and Dowdle 2001; Steger et al. 2004; Steger 2007), to the point some scholars liken these contests to "bumps in the road" (Adkins and Dowdle 2001, 2004). The minor role of these states thus pales in comparison to that of national party endorsements, campaign cash reserves, and national poll standings at the end of the invisible primary period.
Although convincing, these findings present a puzzle: why then are candidates still devoting a substantial amount of attention and resources to these contests? (1) We propose that Iowa and New Hampshire continue to be powerful players within nomination contests, because the increasingly visible invisible primary grants these states a different means of influence. National levels of candidate viability and exposure are responsive to the dynamics of these state contests, since long-shot candidates and the media predominately focus on these two states throughout the invisible primary. As a result, the happenings of the Iowa and New Hampshire campaign have an important say within the national invisible primary before the votes in these states are actually tallied.
Using data from the 2008 presidential nomination campaign, we evaluate these claims with daily measures of candidate prominence in national news coverage, national polls, and Iowa and New Hampshire polls during the last half of 2007. We accommodate for missing data, sampling error, and reciprocal influences by estimating a Bayesian state space vector autoregression model. We find that a candidate's early performance within Iowa and New Hampshire is equally if not more of a contributing factor to national polls and national media coverage as they are a reaction to them. The subsequent analysis of weekly campaign contribution totals also shows that polling performance within these early states has a strong influence on campaign contributions during the invisible primary that equals the contribution of national polling performance. Although these findings do not dispute the importance of candidates needing sufficient national support entering an election year to win a party's nomination, they provide evidence that states like Iowa and New Hampshire still have a profound but indirect influence on the process. To the extent a candidate can gain early support in these states, they will likely increase their national popular support and campaign contributions before any election is held.
Front-loading and the National Invisible Primary
As presidential nominations have become more front-loaded the influence of early states within the selection process has been questioned. Initially, some scholars proposed that front-loading would bolster the influence of Iowa and New Hampshire within nominations by allowing a winner's momentum to affect more contests within shorter periods (Bartels 1988). However, front-loading has also forced candidates to have a campaign organization in place that can compete across numerous states almost contemporaneously. Thus, despite Iowa and New Hampshire's ability to propel candidates onto the national stage, candidates are unable to reap the benefits of an early victory if they lack a strong preexisting national organization (Aldrich 2009; Butler 2004).
Not only has a compressed and front-loaded primary limited the opportunity for early state momentum effects, but studies now suggest that Iowa and New Hampshire victories have little benefits compared to establishing a large national campaign organization. The bulk of the literature has suggested that a candidate's campaign funds, national poll standing, and party organization support entering election season are the strongest predictors of nomination success. Mayer (1996, 2003) has consistently shown that national poll standings at the beginning of the election year are powerful predictors of nomination success. Likewise, other scholars find a campaign's cash reserves entering the election year can determine who is the nominee (Adkins and Dowdle 2001). Cohen et al. (2008) and Steger (2007) have also recently added to this literature by pointing to the role of party insider support measured through endorsements as key predictors of success. Throughout their detailed analysis, Cohen et al. (2008) find that party endorsements are as powerful a predictor of nomination success as national poll standing, and it outperforms other factors like campaign funds and national media coverage.
Moreover, when included as a predictor of primary success, most analyses fail to demonstrate that Iowa or New Hampshire outcomes make a significant contribution. Adkins and Dowdle (2001) found that after controlling for national poll standings and fund-raising numbers, New Hampshire and Iowa outcomes mostly improve predictions of candidate placement but do not modify the winner. At most, Steger et al. (2004) and Steger (2007) find that New Hampshire has a "correcting" effect on candidate performance, where New Hampshire's contribution is much more influential among Democrats. In their analysis Cohen et al. (2008) found that early state victories make significant contributions, but they also found these victories reflect national invisible primary forces such as a campaign's overall level of funding.
Why Early States Still Matter
These findings demonstrate that candidates need a prominent national presence and insider party support when entering the contest period, but they also question the necessity of campaigning in the early state contests of Iowa and New Hampshire. We believe such analyses have yet to fully recognize the complete nature of Iowa's and New Hampshire's influences within the nomination process. As the invisible primary has gained importance and attention within nomination contests, these states have remained a focal point for candidate competition during the invisible primary. In fact, we propose that Iowa and New Hampshire are such an integral part of the exhibition season that many national forces, like media coverage, polls, and campaign funds, partially reflect the performance of candidates within these states before their votes are actually cast.
Iowa and New Hampshire possess a prominent role within the invisible primary partly because long-shot candidates essentially have to win these states if they hope to win the nomination. Front-runners may have enough money and a large enough organization to ignore these early states, but long shots have less resources to remain relevant without early success. Since they cannot divide their resources across multiple contests, these candidates focus on the early states in the hopes that an impressive showing will provide them with positive national media exposure and the resulting necessary resources and popular support to continue their candidacy. Indeed, recent long-shot candidates have explicitly stated their belief that by concentrating their efforts in Iowa or New Hampshire they can gain national support and slingshot their way through to victories in the immediately following contests (Will 2007).
With many candidates concentrating their efforts in these states and given their past role as first-in-the-nation contests, journalists covering the exhibition season often look to these states for indications of a candidate's potential success. Front-loading and greater candidate activity has increased journalist incentives to cover the invisible primary. Beyond national polls and campaign finance reports, the early competitive happenings in Iowa and New Hampshire are easily placed within journalists' predilection for a horserace narrative. Iowa and New Hampshire also provide familiar settings with established campaign events that occur throughout the years preceding the election, including debates, banquets and town hall meetings. Moreover, because many candidates and journalists concentrate on these states, front-runner candidates, who may not need to win these states, might still find that competing in these states provides beneficial levels of media exposure and momentum. These states are thus able to maintain their prominence within the exhibition season of the nomination campaign despite continued doubts about the necessity of winning such states.
Buell (1996) previously found that early activities in these first states were a prominent component of news media coverage and candidate activities prior to the election year. Even prior to the 2008 nomination, when...