I found Sean Trende’s discussion of poll results at RealClearPolitics pretty interesting:
For my regressions, I used the same controls for each aspect I examined: gender, age, race/ethnicity, and income.
With controls in place, Hispanic voters, for example, are less likely to consider themselves highly patriotic than white voters, but not nearly to the extent of black or Asian voters. Being Hispanic increases your odds, all other things being equal, of giving a less patriotic response than whites by 17%, compared to 48% for blacks and 89% for Asians.
Likewise, with controls in place, Hispanics were more likely to be Democrats than similarly situated whites, but not overwhelmingly so. With controls in place, the odds that a voter who was Hispanic would be a Democrat were 11% higher than that he or she would be Republican, compared to 62% for blacks. This is encouraging for Republicans, as it suggests that as Hispanic voters continue to climb the socioeconomic ladder, they may continue to pick up GOP voting habits.
Interestingly, and perhaps counterintuitively, with controls in place the racial groups in the survey exhibited no statistically significant divergence from whites on the question of whether illegal immigration is a major problem, a minor problem, or no problem at all. Likewise, with controls in place, Hispanics are no more likely than whites to say that immigration has done more good than harm. Blacks, however, are more likely than whites to say that immigration has done more harm than good, with controls in place.
While there is more Hispanic support for the DREAM Act than there is among whites, things like a pathway to citizenship show no significant differences between racial groups, once controls are in place.
The survey also asked about Hispanic subgroups. While the number of respondents is low, the regression analysis finds no significant differences between subgroups in levels of Republicanism. This is surprising, but is consistent with interpretations of Cuban American levels of Republicanism being attributable largely to socioeconomic status. Cuban Americans were, however, significantly more likely than other subgroups to respond that immigration did more harm than good. This could reflect the unique status of Cuban Americans with respect to immigration laws, or intra-Hispanic attitudes toward other groups. Cuban Americans were also more likely to support fining employers who employ illegal immigrants.
More than anything else it suggests to me that both Democrats and Republicans are making some pretty poor assumptions.
