Dialect on trial: use of African American Vernacular English influences juror appraisals^*

* Supplemental data for this article can be accessed https://doi.org/10.1080/1068316X.2019.1597086.

ABSTRACT

We investigated the effect of dialect and race on juror decision making. Mock jurors read a summary of an ambiguous criminal case, which included audio of a defense witness (Study 1) or defendant (Study 2). Both speaker dialect [General American English (GAE)/African American Vernacular English (AAVE)] and race (White/Black) were crossed; Study 2 also included three levels of case (Ambiguous/Pro-Prosecution/Pro-Defense) to evaluate any effects of evidentiary context. In both studies, jurors who listened to the AAVE recording found the AAVE-speaking witness to be less professional and less educated than their GAE-speaking counterparts. Interestingly, jurors in Study 2 who heard the defendant use GAE were more likely to find him guilty and found him less credible when the case favored the prosecution, hinting that ingroup biases such as the black sheep effect may also play a role in perceptions of dialect. Secondary analyses found that AAVE predicted more negative overall evaluations of the speaker, and these negative evaluations were associated with an increase in guilty verdicts. Together, these findings suggest that dialect plays an under-investigated role in the courtroom, and that bias against AAVE negatively impacts juror appraisals of its speakers and can potentially influence juror decision making.

KEYWORDS:

• Dialect
• race
• juror decision making
• appraisals
• attitudes
• bias

Acknowledgements

The authors would like to acknowledge Sam Henderson, Erika Arvidson, Ann Iftikhar, Elysse Reyes, Rachel Sangster, Tierra Carter, and Gabi Fowler for their invaluable help on this project. We also extend our thanks to the anonymous reviewers of this manuscript for their helpful comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Courtney A. Kurinec http://orcid.org/0000-0001-5800-1610

Notes

* Supplemental data for this article can be accessed https://doi.org/10.1080/1068316X.2019.1597086.

1 Unlike dialects, accents refer to only the pronunciation of a given language.

2 Restricting the sample to only those participants who correctly identified both the race and dialect of the witness in Study 1 only revealed a marginal effect of dialect on likeability ratings, F(1, 73) = 2.97, p = .089, partial η² = .04. Otherwise, the pattern of reported results remained the same.

3 We also asked participants to indicate whether the dialects they heard from the defendant and officer were ‘standard’ or ‘non-standard’ to provide support for the claim that the AAVE recording was viewed as non-standard compared to the GAE recording. As expected, participants perceived the GAE accents for the defendant and officers to be ‘standard’ American English (all χ² p < .001), but this did not reach significance for AAVE, χ²(1) = 0.64, p = .423. Nearly all participants identified the GAE-speaking defendant (92.3%) and officers (94.9% with the GAE suspect, 97.4% with the AAVE suspect) to be speaking standard American English, while only 56.4% of those listening to the AAVE-speaking defendant believed him to be speaking a standard dialect.

4 Thanks to Sara Appleby for sharing the photos with us.

5 Restricting the sample to only those participants who correctly identified the race and dialect of the defendant in Study 2 revealed an overall similar pattern of results. However, our interaction between case and dialect on verdict decision now also indicated that those in the pro-defense condition were more likely to find the AAVE-speaking defendant guilty, χ² = 7.63, p = .006, ϕ = -.29. Additionally, the case by dialect interaction on credibility ratings became marginal, F(2, 281) = 2.61, p = .075, as did the effect of case on professionalism ratings, F(1, 281) = 2.54, p = .081, and our race bias measure was no longer a significant covariate for verdict, ratings of professionalism, or level of education. Regarding our secondary analyses, using this restricted sample rendered our case by dialect interaction marginally significant, F(2, 281) = 2.73, p = .067, and revealed a marginal effect of case on overall evaluations, F(2, 281) = 2.35, p = .097, partial η² = .02. However, adding in our covariates resulted in the same pattern of results as reported in the text, although race bias was not a significant covariate. Notably, the effect size of dialect on overall defendant evaluations was much larger with this restricted sample and more comparable to that in Study 2 (partial η² = .08-.09). Note, we were unable to run our logistic regressions on verdict using our full model due to quasi-complete separation in this restricted sample. Therefore, we included only our main effects and expected case by dialect interaction, which allowed us to run the model.

6 To examine the density of the AAVE employed in our studies, we coded the AAVE speaker’s audio for 28 features associated with AAVE: the 23 morphosyntactic features used by Craig and Grogger (2012) and five phonological features coded by Rickford et al. (2015). The total number of occurrences of the features were divided by the total number of words to create our dialect density measure. The AAVE dialect in Study 1 was denser than that used in Study 2 (Study 1: 0.11 features per word, Study 2: 0.05–0.07 features per word). The difference in density likely reflects the difference in speech contexts (witness statement versus police interview). Of note, the density of morphosyntactic features in Study 1 (0.03) is higher than the average reported for adult African Americans by Craig and Grogger (2012; M = 0.02), whereas the density for Study 2a is more comparable (0.01; changes made to the cases in Study 2b removed the few morphosyntactic features in the audio).

Additional information

Funding

This work was supported by an American Psychology-Law Society Grant in Aid. The authors declare that there are no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Results from this project were presented at the American Psychology-Law Society Conference in 2017 and 2018.