19
Latinos. The only issue among this set that was significant in the probit analysis was the health
care issue, which had a negative impact. Looking at the first differences, we also see that this
effect was small, only decreasing the probability of voting by about 3 percentage points for both
Democratic and Republican ads.
These results also help solve the puzzle of why Republican Spanish language ads were
more effective than Democratic Spanish language ads. Recall from the descriptive section on the
ads that a higher percentage of the Republican Spanish language ads focused on education,
compared to the Democratic Spanish language ads.
Table 7: First Differences from the Model with the Issue Ad Variables
Type of Ad
Democratic Ads
Republican Ads
Min-Max
Min-Max
Education
.544
.899
Family
-.020
-.012
American Dream
-.004
-.011
Environment
.124
.259
Health Care
-.031
-.030
Abortion
.026
.129
Economy
.041
-.030
Discussion and Conclusion
This paper has addressed two of the limitations in extant work on the role of mobilization
in influencing Latino turnout by looking at another dimension of mobilization, namely political
advertisements, and by separating the sponsoring party of the advertisements. Our results extend
this literature by showing that ads have an impact, though the direction and magnitude is
conditional on the tone, sponsor, and type of issue being used. Furthermore, our results test
competing models of the effects of ads based on the target and tone of the message, which is
beneficial to the literature on acculturation and to the general literature on campaign effects. We
will briefly review the results, and then turn to plans for future research.
We clearly established that indirect mobilization in the form of political ads is an
important factor in influencing Latino turnout (H1), in that many of our measures of political ads
were significant in many of the models. We were also able to test competing models of whether
Spanish language or English language ads would be more effective in influencing turnout. Our
results lend support to the model of assimilation (H4) over the models from the persuasion
literature and the pluralist model (H3), in that the ads in English, particularly the indirectly
targeted ads, were more effective at increasing the likelihood of turnout.
Overall, the results supported our hypothesis (H5) that the Democratic ads would be more
likely to increase turnout than the Republican ads. However, the Republican ads in Spanish and
the indirectly targeted ads had a significant positive impact on Latino turnout. Our models that
related to the tone and issue lend some insight as to why this might be the case, in that the
Republican ads were more positive and more likely to focus on education. We hope to further
tease out these findings in future analyses by controlling for the differential impact among
Cubans and by possibly breaking down the Spanish language ads and indirectly targeted ads by
tone and issue.
In regard to the tone of the ads, we found support for Ansolabehere and Iyengar’s (1995)
hypothesis (H10b) that negative ads should decrease turnout. These results cannot be extended
Convention