This post at the Latino Decisions blog drills down into the Florida International University robo-poll from earlier this month which claimed that Florida Latinos were breaking for Obama by only 50.7% to 44.2%
The post offers two things to consider.
1. The poll did not include cell phones, and as most here already know, we now have a wealth of information that this means that non-Cuban Hispanics are likely to be undersampled.
2. It does not account for the political attitudes of Cuban-American voters as distinct from other Latino groups, for example Puerto Ricans. As the post notes:
shows that 39% of Cuban Americans intend to vote for Obama (50% for Romney), compared to 71% of all other Latinos in the poll (22% for Romney). The issue here is that the FIU results were not “weighted” to account for the possibility that Cuban Americans—i.e., votes for Romney—were overrepresented in the poll relative to their actual share of Florida’s Latino electorate.
The author of the post, Casey A. Klofstad, is an associate professor of political science at the University of Miami. It is his opinion, not officially representing Latino Decisions, but there is no doubt that they agree with him by letting him make this guest post, which provides lots of data.
Thought it was worth ensuring that people were aware of it.
Peace.