The results of a recently commissioned poll by the Las Vegas Review Journal were probably warmly received by the Nevada state Democratic organization.  Indeed, polls indicating a tightening of the Democratic race nationally, in the other early voting states, and in Nevada suggest that Hillary Clinton may not have the cake-walk through the nomination process that many expected just a month or so ago.  These same polls also indicate a Republican contest that is increasingly more fluid by the day and open to a myriad of scenarios.  All of this, in turn, should elevate the importance of what happens in Nevada on January 19th.

At the same time, the Review Journal’s poll of 300 likely Democratic and Republican caucus voters belies the difficulty of polling in caucus states.  First, polls with sample sizes of 300 tend to lack precision, resulting in large margins of error.  In this case, the margin of error in the RJ poll for both the Democratic and Republican fields is plus or minus six percent!   This means on the Democratic side that Hillary Clinton’s lead in Nevada over Barack Obama could be anywhere between two and 20 percent.  Similarly, Rudy Guillani could either be in a statistical tie with Mitt Romney or lead by as much as 17 points in the Republican contest.

Second, the poll seeks to identify respondents who are considered “likely caucus voters.”   Typically, this is done via a series of screening questions that depending upon how a person contacted by the poll responds determines if he or she is included in the actual sample.  Each person who was initially contacted by the poll, but was then excluded from the final sample due to their responses to the screening questions may undermine the representativeness of the eventual sample.  This is the case because these people were randomly selected to be in the poll and their exclusion means that they needed to be replaced by someone else who initially was not selected to be in the sample.  Systematic differences between these two groups can enter bias into the results. Unfortunately, the RJ did not report how many calls were actually made by its pollster to obtain the samples of 300.

Beyond the nuts and bolts of how a poll is conducted, polling for caucuses is problematic because caucuses are very different animals from the more typical primary or general elections.  Primary and general elections, of course, require much less effort from participants, as all they need to do is walk into their polling place and pull a lever (or touch a box on a computer screen) and then leave. 

In contrast, caucus participation requires a much greater investment of time and energy as caucus voters need to not only show up in the correct location at the appropriate time, but then be willing to either convince (or be convinced by) other voters from their precinct which candidate they should or should not support.  Given that intra-party contests typically draw very small numbers of potential voters (the dirty secret of the Iowa caucus is that in 2004 less than six percent of eligible voters participated) and when push comes to shove, even politically engaged citizens may be unwilling to spend a Saturday during a three day weekend participating in an odd and unfamiliar process, actually identifying who will turn out on January 19th is difficult at best. 

On the Democratic side, matters become even less definitive due to the party’s decision to allow voters to register and participate on the 19th (GOP voters must be registered a month prior to the primary).  Same day registration opens the door for potential raiding by campaigns that are able to mobilize a large number of previously unengaged voters into the process at the last minute.  While state law requires all participants to be Nevada citizens, in all likelihood their will be no mechanism in place to validate these same day registrations prior to the actual caucuses creating a potential free for all.

The Democrats also are implementing precinct level threshold requirements.  These requirements mean that they are actually two rounds of voting at the precinct level in the Democratic caucuses.  In the first round, those in attendance cast what might be thought of as a sincere vote for the candidate they really want.  For voters whose candidates fail to meet a predetermined threshold, their candidates are eliminated from further consideration and they can be recruited by supporters of other candidates.  In practice what this means is that barring a breakthrough in Iowa or New Hampshire by a second-tier candidate, voters whose first choices are not Hillary Clinton or Barack Obama are likely to become free agents in the second round of voting and can throw their support to which ever these candidate they prefer or remain undecided.

As we see, trying to figure out who a likely caucus voter is and who these likely caucus voters are going to support is no easy task.  Moreover, the tool typically used to address these concerns - the public opinion poll - is of little help in a caucus setting.  So for those who are interested in predicting what will happen in Nevada on January 19th, what available information can be used to make an informed guess? 

The answer to this question actually is straightforward.  Quite simply, caucuses are won and lost by the quality of the campaigns’ grass roots organizations.  Campaigns that have developed the infrastructure needed to identify, repeatedly contact, and mobilize likely supporters on the 19th are likely to do well.  Politics of this type place a premium on the strategic allocation of manpower, unrelenting one on one persuasion of undecided and soft supporters, and a willingness to exploit the rules and procedures of the caucus process. 

Candidates who are banking on name recognition (Giuliani) or favorable media attention (Mike Huckabee and Ron Paul), last minute ad buys, or the movement of assets into Nevada after the Iowa and New Hampshire (John Edwards and Bill Richardson) contests will quickly find their campaigns outgunned and out maneuvered on the 19th by the campaigns that have been consistently playing for keeps in Nevada.  On the Democratic side this means Hillary Clinton and to a lesser extent, Barack Obama.  For the Republicans, the only candidate with the semblance of organization in the state is Mitt Romney, who may now be counting on a Nevada win to buffer his campaign against early losses in Iowa and South Carolina.