Do Poorly Executed Spanish-Language Appeals Negate their Persuasive Effect?

• In Event: Candidate Evaluations and the Constraints of Race, Place, and History

Abstract

This paper examines the possible costs to a candidate when they try to show a connection with Latinx constituents in Spanish and fail or whether their mere attempts to engage the bilingual community are rewarded. Additionally, this study tests whether certain candidates are given greater latitude to make mistakes. For example, in the first Democratic primary debate for the 2020 presidential election, a nationally televised event broadcast on English-language networks, a total of four candidates (Beto O’Rourke, Cory Booker, Julian Castro, and Pete Buttigieg) chose to answer some questions in Spanish. These candidates’ Spanish-language gamble varied in quality. When they are subpar, this study argues and does find that these failed attempts can negate their persuasive intent. In fact, the analysis shows that these mistakes do have major consequences, so much so that the costs are greater than the potential reward.

To test whether there is such an adverse effect, this study fielded an experiment from a unique sample of 2,000 native Spanish-speaking Latinxs focusing on two central components. First, Latinx bilinguals were asked to evaluate two hypothetical candidates, one White and one Latinx, after highlighting the linguistic ability of each. Sometimes the White candidate was described as more fluent than the Latinx and vice versa. The second component in this experiment speaks to possible consequences of poorly executed dual-language outreach. Respondents were asked to evaluate two different candidates, one Latinx and one White, varying which candidate makes the lexical mistake. The focal analysis estimated the effects of a candidate’s Spanish-language ad when it contains a lexical mistake versus when it does not. Treatment effects between the two were then compared to see if these failed attempts are punished in equal measures. A mistake by a White candidate could be reasonably expected of an ethnic outsider. A mistake by a Latinx candidate, however, could be seen to be egregious and to go against the subjective expectations of someone within the ethnic group. The political implications of a study of this kind are clear. Candidates seeking to use multiple languages as a way of reaching specific communities will need to weigh the benefits as well as the potential costs of such strategies, which are often implemented haphazardly.

This research confirms that candidates do have much to gain by signaling bilingual ability, irrespective of their ethnicity. Interestingly, this section of results also suggests that co-ethnic candidates are more heavily penalized for limited proficiency with Spanish than are White candidates. One plausible explanation for this is that Latinx candidates are seen through an entirely different lens that carries with it certain linguistic expectations. While co-ethnic candidates can communicate such a kinship in several ways, these results suggest that White candidates can highlight their own ties to bilingual or non-English speaking constituents through language. These findings further indicate that when candidates do choose to use Spanish, linguistic missteps do not go unnoticed. These errors are markedly more detrimental when made by a White candidate. Moreover, as in the previous section of the experiment, respondents’ sense of ethnic attachment is found to have a mediational influence on evaluations of the quality of a candidate’s Spanish-language appeal, and by extension, their perceived political proximity with Latinxs.

Author

• Alejandro Flores, University of Chicago