That is why simple post-election tallies miss the real impact that a group has on election outcomes during the weeks and months of the campaign. Nevada exemplified this impact in 2008. Exactly two weeks before the election, the Politico/Insider Advantage poll put Nevada at 47% Obama, 47% McCain, and 6% undecided. The campaigns spent $13 million on television advertising alone in the state, which was inundated with television and radio ads, candidate appearances and events, and voter outreach efforts.31 Ultimately, Obama won Nevada by twelve points, with an estimated 76% vote from Latinos, up from 60% for Kerry in 2004. This is Latino influence, whichever way we want to count it.
Voter traits and trends, of course, are standard measures of influence, but there is more to deciding whether a given group is influencing the political landscape. Voters, the media, and campaigns tell each other how close a race has become. Campaigns rely heavily on polling and are especially attuned to short-term, recent trends in turnout, partisanship, margins of victory, voter registration, and demographic composition. Using this information, they decide where to spend money and how best to get the vote out. National and regional media communicate to both voters and campaigns the closeness of the race and the importance of particular issues and groups of voters. When voters see these reports on how close the election is, the increasing attention and excitement has been shown to affect turnout.32
In this vein, news stories that highlight the importance of the Latino vote convey to campaign staffs and the broader electorate the importance of Latinos in creating statewide competitiveness and winning coalitions. Online media provided another unique contribution to assessing and publicizing campaign competitiveness in 2008. Three sites—Five ThirtyEight.com, Pollster.com, and RealClearPolitics.com—developed a national following for their regularly updated (at weekly and daily intervals), empirically derived predictions of state-by-state election outcomes, and FiveThirtyEight.com’s Nate Silver even became a star eventually. National, state, and local news outlets regularly sourced the “RCP average” and the “538 prediction,” based on the survey and poll results that were posted and analyzed by both websites, as authoritative measures of national and state-level campaign competitiveness in the weeks leading up to election day.
Of course, voters are also influenced by direct mobilization: television, print and radio advertising, mailers, phone calls, and online mobilization efforts signal to voters that their state is in play.33 Latino voters are no exception, though as we pointed out in Chapter 4, they are less likely than other groups to be the target of outreach, particularly from parties and candidates. A spate of recent research points to the effectiveness of targeted campaign appeals to Latinos.34
The Obama campaign brought peer-level innovation to online mobilization and incorporated this technology into unique Latino outreach strategies already in place.35 As anyone who ever contacted the 2008 Obama campaign knows, it revolutionized the use of electronic communication and social media for political purposes. The campaign website facilitated extensive contact in two directions: (1) directly from the campaign to voters, and (2) voter to voter. Those who provided contact information to the campaign regularly received text messages and emails encouraging their participation (as voters, contributors, or volunteers) in the primary and general elections, which were consistently described as “tight races” and “tough battles.” Individuals were also encouraged to self-identify with multiple online peer groups (for example, “Latinos for Obama,” “Ohioans for Obama,” or “Obama-mamas”), each of which had its own Web-based organizational arm. Every one of these organizations conducted outreach activities aimed exclusively at its particular affinity group in key states.
As we discussed earlier, it is often easier to get Latinos mobilized by resorting to friends and family rather than spokespeople, famous or otherwise. For average Latino voters—many of whom are never contacted at all by candidates or parties—making it easier for them to interact repeatedly with influential familiars and to be contacted by them was an important and potentially pivotal strategy shift.
Sí SE PUEDE? MEASURES OF LATINO INFLUENCE IN 2008
In the aftermath of Obama’s victory, many efforts were made to assess the relevance of the Latino vote. Using nothing more than post-election tallies to evaluate whether the Latino vote caused a state to be won or lost, analysts found the Latino influence to be weak. In addition, Obama had won many states by a wider margin than expected, and that made it difficult to find the math proving that Latinos cast the deciding ballots. In our own assessment, we find strong and consistent evidence across our three key areas—group size, electoral patterns, and mobilization—that Latinos in key states did influence the 2008 election. Further, our data may foreshadow which states will merit the attention of pundits down the road as the Latino influence grows in new regions and new states (two strong candidates being Montana and Georgia).
The traditional post-hoc election result tally is one of the measures of influence cited by de la Garza and DeSipio in their quadrennial analysis of the Latino vote in presidential elections, though they (correctly) dismiss it as too unrealistic.36 Latinos may have influenced the 2008 election if the margin they provided for the winner was larger than the overall margin of victory. In other words, if no Latino had voted, would Obama have won?
TABLE 6.2Did Latino Votes Provide the Margin of Victory in the 2008 Election?
Note: BO = Barack Obama and JM = John McCain. Electoral outcomes are from statement of vote in each state. Latino estimates are based on exit poll estimates multiplied by total votes cast.
a. Because the North Carolina and Ohio state polls are not available, the national average from the Latino Decisions poll has been substituted for each.
Looking at Table 6.2, we find three instances in which the overall state victory margin for Obama was smaller than the vote margin provided to him by Latinos alone. By this crude measure, it is possible to argue that Latinos directly influenced the election results in North Carolina, Indiana, and New Mexico.
However, this measure dismisses the influence that Latinos probably had in other states, such as Nevada, Florida, and Colorado, where the overall victory margin was too great for Latinos alone to have provided it. This type of analysis is problematic for several reasons. First, it offers no leverage in predictive research questions about the conditions prior to election day and about the states that will matter because it is based solely on tallies of election results. Second, it ignores the states where there may have been Latino influence during the campaign through outreach, advertising, or mobilization but the election results do not back this up. Third, this analysis may artificially categorize a state as an “influence” state just because the overall margin was razor-thin there. As we have noted, it is more than possible for a state to be perceived as a close contest but to be deemed noncompetitive after the votes are counted.
Group Size and Growth
A prerequisite for a group to have influence is that it meets a minimum group size and, preferably, it should be cohesive or mobilized. If the presidential election in Maine or North Dakota is very close, it is impossible for Latinos to influence the outcome because their group size is too small in those states and it’s not growing rapidly. Thus, a simple starting point for any analysis of a minority group’s influence is to assess the share of all registered voters that it represents. Data from the 2006 Current Population Survey, which provide the best estimate for Latino percentages among registered voters for all fifty states, show a range from a low of 0.1% in Maine to a high of 30.4% in New Mexico. In particular, it would be very difficult to ever witness Latino influence in a statewide election in a state that is less than 2% Latino among registered voters. According to the CPS data, twenty-five states were 2% or less Latino among citizens registered to vote in 2006. The patterns depicted in Figure 6.6 are predictable and consistent with Latino population figures that are now well known. States in the Southwest and the Mountain West have significant Latino registered-voter popul
ations, as do Florida and some states in the Northeast.
FIGURE 6.6Percentage Latino among Registered Voters, 2006
In addition to group size, the growth rate among registered voters is particularly important as a gauge of influence. Figure 6.7 reports the change in the Latino-to-white voter registration share over an eight-year period, 1998 to 2006. This estimate gives us a sense of the absolute gains in Latino voter presence vis-à-vis the largest group in the state (whites). The states depicted in lighter shading reported little to no change in the Latino-to-white comparison. That is, if Latinos were 10% of all registered voters in 1998, they still made up about 10% of registered voters in 2006. In contrast, the states depicted in darker shades experienced accelerated Latino registration growth. For example, in 1998 the Nevada electorate was 86% white and 5% Latino; by 2006 that had changed to 75% white and 10% Latino, resulting in an eleven-point difference for whites and a positive five-point change for Latinos and yielding a net increase for Latinos of +16. Other states, such as Wyoming, Missouri, Ohio, Maryland, and Massachusetts also witnessed a net increase in registered Latino voters of over 7%. These sizable shifts in ethnic composition within the electorate are remarkable because they occurred in less than a decade.
FIGURE 6.7The Growth in Latino Registration Relative to White Registration, 1998–2006
Of course, we would expect Latinos to exert some influence where they are a sizable share of the population. They may also have influence in states where they are relatively small in number but are becoming a rapidly increasing share of the electorate, signaling the demography of the future voting public. Growth measures alone may miss the influence of Latino voters in places where there is a large and relatively stable share of Latinos in the electorate. New Mexico is a case in point: the growth in the Latino percentage of the population between 1998 and 2006 is unremarkable, but nonetheless nearly one voter in three in New Mexico is Latino, and the state has a long history of electing Latinos to office, including the US Senate and the governor’s office. Indeed, the last two governors of New Mexico have been Latino—former Democratic governor Bill Richardson and the current governor, Susana Martinez, a Republican. So it is the combination of population size and growth rate that more realistically captures a group’s opportunity for influence in a state election.
Electoral Patterns and Volatility
For a group to demonstrate electoral influence, election-specific factors are of obvious importance. Here we focus on two: the degree of voting cohesiveness among Latinos, and the degree of expected competitiveness of the state election.
FIGURE 6.8The Latino Democratic Vote Relative to the Non-Latino Democratic Vote, 2000–2006
Voting cohesiveness is measured as the average Democratic vote among Latinos from 2000 to 2006 minus the average Democratic vote among non-Latinos. For Latinos to have influenced a state’s election, they ought to have demonstrated somewhat different voting patterns than non-Latinos in the state. Using the National Election Pool (NEP) state polls for the 2000, 2002, 2004, and 2006 elections, we created a measure for average Democratic vote for Latinos and non-Latinos by state. Figure 6.8 shows the Latino Democratic vote differential for all states. States shaded dark are those where Latinos vote much more consistently Democratic than do non-Latinos in the state, while states shaded lighter are those where Latinos and non-Latinos demonstrate very similar partisan vote preferences. Latinos tend to vote more Democratic than non-Latinos do throughout the United States, and this is most pronounced in the Southwest and Mountain West, where four states—Texas, Arizona, Colorado, and Utah—have a Latino population that is about twenty points more Democratic. There is a notable pattern of Democratic vote cohesion among Latinos throughout the entire West.
FIGURE 6.9Average Competitiveness Level by State, October 2008: Real Clear Politics
The second important piece of electoral information is the expected competitiveness in the state election. This is one of the most important indicators of group influence. Without a doubt, it is much more difficult—though not impossible—to influence an election in a state that is completely uncompetitive. Although it is worth pointing out that some states, like California, are uncompetitive largely because Latinos are influential.
However, the traditional measure of looking to post-election results is not a complete guide to competitiveness. A group has influence, not after the election, but during the active campaign, most likely in the last thirty days. We took the average poll rating one month before the election from Real Clear Politics’ state poll average in 2008.
The map in Figure 6.9 is familiar to most readers: it shows the anticipated closeness of the 2008 presidential election in each state. The most darkly shaded states were those with very close pre-election poll averages, while the very lightly shaded states were not expected to be close at all. Coupled with the data reported in Figure 6.8, it is possible to sort out the states with more or less Latino influence. For example, a state like Ohio was expected to be very competitive, and it had a Latino electorate that voted considerably more Democratic.
FIGURE 6.10Changes in the Media Coverage of the Latino Vote, 2004–2008
As we add in additional factors, such as those collected in Figures 6.6 and 6.7, the overall influence story begins to take shape. We turn next to the last piece of this puzzle: the components of mobilization.
MOBILIZATION AND RESOURCES
The final set of criteria that we believe to be important is how a campaign engages—or fails to engage—the Latino community. The campaign itself must have taken note of Latinos as a potential influence group. That is, as the candidate campaigns and the media covers the campaign, both must have paid attention to the Latino vote as a crucial bloc. We assess this by measuring three factors: the change in media coverage of the Latino vote; the campaign ad buys targeting Latinos; and the campaign’s ethnic mobilization. Using Lexis-Nexis content data on the Latino vote and the 2004 and 2008 presidential elections, we amassed data on both the rate of news stories on the Latino vote and the change in this rate from 2004. These data are normalized and combined in Figure 6.10. The darker states are those where media coverage of Latino voters increased; as the map shows, almost every state saw a steady increase in the coverage of the Latino vote during the presidential campaign.
Spanish-language television ads are an easily collectable proxy for outreach to the Latino community. TV ads are important because they are costly; the decision to spend finite resources on the Latino community would have emerged from an important campaign calculus. However, Spanish-language ads are not the only way in which campaigns targeted Latino voters. Unfortunately, data on the content of English-language ads, collected by the Campaign Media Analysis Group (CMAG) project, are not available until two years after the election. In contrast, Spanish ad data are available in real time from the public disclosures on campaign spending. We do not include a figure on Spanish TV ad expenditures because only four states ran both Obama and McCain campaign ads: Florida, Nevada, Colorado, and New Mexico. The fact that none of the other states had such ads puts significant weight on these four states.
In addition to TV ad spending, we gathered data on Latino group mobilization by the presidential campaigns. Although data were available only from the Obama campaign, they were very rich. Again, for each state we collected the number of members of Latinos for Obama groups, and also the amount of money raised by these groups, both normalized over the state’s total Latino registered voter population. In Figure 6.11, we map the Obama campaign’s Latino mobilization. Before we can accept or dismiss the claim that Latinos had influence in a particular state, we should assess the degree to which they were mobilized. For example, Figure 6.11 shows that Latino mobilization by the Obama campaign in 2008 was fairly strong in Virginia, North Carolina, Indiana, and Nevada. It also shows that Latinos in Missouri and Arizona were less likely to be mobilized by the Obama campaign; Obama narrowly lost these two states, but might have won them had his campaign engaged in str
onger Latino mobilization. Latino activity was brisk in noncompetitive states like Illinois, New York, and Texas. The online participation and monetary contributions of Latinos in these states especially matter to the political elites and fund-raisers, irrespective of the electoral competitiveness of their state.
We can depict Latino influence by combining all of these data points into a single model. Drawing on these three categories, we include group size and growth, electoral volatility, and mobilization to explain Latino influence in the 2008 election. The final Latino influence map (Figure 6.12) combines all prior metrics. The darker states had higher Latino influence composite scores, while states that are white had practically no Latino influence. Theoretically, a state with the absolute strongest Latino political influence meets the following conditions: it has a large Latino population, rapid growth in Latino voter registration, record increased rates of partisan cohesiveness compared to non-Latinos, a competitive electoral environment, media focus on the Latino vote, and extensive campaign outreach to and mobilization of Latinos. Anecdotally, the darkest states on the map seemed to have had the greatest Latino influence: Florida, Nevada, Colorado, and New Mexico. Texas and Arizona also score high because on a number of metrics they demonstrated considerable Latino influence; however, they did lack a key element in 2008, namely, competitiveness and/or campaign outreach. The states shaded gray matched our expectations: Virginia, Indiana, and Ohio all demonstrated high Latino influence in 2008. Missouri, Minnesota, California, and Washington were also Latino-influence states, but they did not score consistently high across all dimensions.
Latino America: How America's Most Dynamic Population is Poised to Transform the Politics of the Nation Page 11
