That is one question explored by Nate Silver in this post, which starts by exploring new polling data out of AZ that shows Obama marginally ahead.
In the post, Nate explains why he thinks that is unlikely, but then drills down more deeply, noting that the poll in question conducted bilingual polls. He then writes:
Polling firms such as Latino Decisions that have conducted interviews in Spanish have shown Mr. Obama with a larger advantage among Hispanic voters than those which interview in English only. The most recent Latino Decisions poll, for example, had Mr. Obama ahead 72-20 among Hispanic voters. This poll is not an outlier; other polling firms that have conducted Spanish-language interviews have found similar results.
He notes that some think this offset because those who speak only Spanish or who are inclined to prefer to speak Spanish are less like to vote. Here I might disagree - I have a brother-in-law who grew up speaking Spanish at home and not learning English until he began school, but his family has been in Northern New Mexico for several hundred years. I also note that one can arrive in the states from Puerto Rico, know next to no English, and still be entitled to vote, because someone born in PR is a natural-born citizen.
But Nate to his credit goes further, as i will explain beneath the squiggle.
Nate acknowledges the possible difference that could be made if polls were understating the participation of Hispanics who prefer to be interviewed in Spanish. He offers this:
On Saturday, I ran an alternate version of the FiveThirtyEight simulation in which I assumed that Mr. Obama would in fact win Hispanic voters by 50 percentage points, his edge in the Latino Decisions poll, as opposed to the roughly 35-point margin he’s had on average in polls that were conducted in English only.
In many states that would not make a difference - increasing Obama's share of Hispanic voters in TX adds only 3 points, not enough to change the state.
So far in AZ based on his modeling it would only increase Obama's chances of carrying the state from 4% to 8%.
But then consider this:
However, the adjustment increased Mr. Obama’s win probability in Colorado to 57 percent from 44 percent, in Florida to 53 percent from 35 percent, and in Nevada to 77 percent from 62 percent. It even helped him slightly in Virginia, where about 5 percent of voters identified as Hispanic in 2008 exit polls.
Overall, Mr. Obama’s chances of winning the Electoral College rose to 69 percent from 63 percent.
Nate reminds people that the only two Senate races his model got wrong in 2010 were Nevada and Colorado, both states with relatively heavy Hispanic populations.
It worth adding to Nate's comments several observations
1. There has been a major effort to register younger Hispanics this cycle
2. While many in the Latino community are not happy with the administration on deportations to date, they are ecstatic with the executive order implementing parts of the Dream Act agenda, and they despite Republican positions on immigration not only because of Arizona's Show Me Your Papers law and Joe Arpaio, but because Mitt Romney and other Republicans are in their mind associated with this approach
3. In both NC and VA, the proportion of the overall vote that will be Hispanic is going up - in Virginia it may reach 7% this time.
4. In Florida a good portion of the increase in Hispanic voters comes from Puerto Ricans in the I-4 corridor.
For a variety of reasons I have argued for much of the past few months that the polls were likely to be somewhat understating Obama's support in some states precisely because of the Latino vote.
Then again, we will know some of this in about 3 1/2 weeks. Except, given the lack of exit polling in for example TX, NY and CA because they are not considered in contention presidentially, yet among the three hold a substantial portion of the nation's Hispanic population, we may not be sure of the impact because we will not know what the actual proportion of the votes in those states are Hispanic.
Still, I commend Nate Silver for being willing to examine this topic and point out the possible difference that it could represent in what may be happening.