When we look at all Mexican Americans together, can we account for partisan choice,
and can we detect differences by generation of immigration? The results, presented in Table 6,
shows that there is some predictability to variations in Mexican American partisanship. It also
suggests that there are systematic differences between the generations. And some of the findings
contradict conclusions from earlier work, possibly because the mechanisms of Latino
partisanship have in fact changed.
Models 1 and 3 include income, specified separately for men and women, while models 2
and 4 exclude income for the reasons sketched out above. Models 3 and 4 combine men and
women who see the economy as the most important issue influencing their vote, while models 1
and 2 separate this variable by gender. Many of the results are the same across the four models;
we note differences below.
Religion has much the same effect as in earlier work: Protestants are more likely to be
Republican than are Catholics, who are more likely to identify as Democrats. While this
religious variable may reflect level of integration into the ethnic community, the total irrelevance
of Spanish language may indicate that instead it is picking up the religious orientations of the
2004 campaign, although we are loathe to push that point given the idiosyncrasies of this
language measure. The greater proclivity of Texans to identify as Republican is consistent with
the candidacy of a former governor and extensive Republican organizing among Latinos in the
state. In runs not reported here, we found that residence in California had no effect on party
choice across a wide range of specifications. Consistent with earlier work (e.g. Cain, Kiewiet,
Uhlaner, 1991) we find that the older members of the third generation are more likely to identify
as Democrats than the young. Our other demographic results, however, contradict the earlier
work. Age has no effect for the second generation. The immigrant generation is more likely to
identify as Democrats than the second generation, but this is counterbalanced by the fact that
29