Alex Jennings and Dr. Wendy Baker-Smemoe, Department of Linguistics and English Language

Linguistic profiling¹ has emerged as an area of research only within the last decade. Several researchers have shown that people who speak non-standard, ethnically-tied dialects are treated less favorably than their Standard American English (SAE)-speaking counterparts. However, no known studies have shown the relative importance of visual versus linguistic cues in profiling. The purpose of my project was to experimentally this by answering this question: Does one’s dialect have a greater bearing on racial judgments than skin color? Sadly, because this type of research has never really been done, we had to synthesize a new methodology. We did this by adapting aspects of established linguistic methodologies in order to fit our needs. This proved to be much more difficult and time-consuming than we had anticipated. Because of several setbacks with difficulties in selecting stimuli, obtaining permissions, and setting up procedures, we have not yet finished the experiment; however, we are well on our way and expect to finish our work by the end of March and hopefully publish our results in late spring.

One of the setbacks we encountered was the difficulty of collecting and preparing stimuli. Since this is a voice-face study, we needed representative portraits and sound clips from multiple populations (African American, Hispanic, and Caucasian with respective culturally-associated dialects). We tried to make this simple by using photos and recordings in the public domain, but we discovered that it is nearly impossible to find twelve portraits which are consistent in lighting, angle, background, dress, age, and facial expression. If we were to do it again, we would find models and take the photos ourselves, though we hadn’t the resources to do so this time.

Sound recordings were also tricky: all voices had to read the same script at roughly the same rate and intonation using a comparable recording quality. Luckily, the International Dialects of English Archive (IDEA) was able to provide us with the necessary clips, but then from them we learned about paperwork and the costs of obtaining permissions. It was good for undergraduates like us to learn about this and really helped me appreciate the hard work researchers go through to carry out a successful study.

Last of all, as far as the stimuli were concerned, it was hard to find people whose faces or voices were clearly recognizable as coming from a particular ethnic/cultural group. For example, we originally intended to use Middle-Eastern stimuli in the study as a guise to lead subjects to think we were comparing American to foreign stimuli; however, we discovered that the average person cannot really tell the difference between a Middle-Eastern and a Hispanic person by just voice or face alone (especially when stripped of identifying cultural markers like particular clothing). In the end, we decided to proceed without this group in order to simplify the study.

Another roadblock we encountered was dealing with the logistics of the experimental procedure. One example of this was compensation of subjects. Originally we planned on entering subjects in a drawing for one of three $25 gift cards, and later on we decided to offer $3 to each participant, but in the end we decided to do neither. The reason for this decision was that we had no idea how to go about physically getting each subject the money. Do we ask for their addresses and send them an envelope? If we do offer compensation, then how do we know they completed the survey and didn’t just click through and sign up for the compensation? We were not prepared to deal with these questions, and lacked the knowledge of how to go about it without compromising anonymity of the subjects or costing a lot of time and money. Fortunately, our mentor suggested that the experiment was interesting enough to go viral on social media without compensation, so we took her advice and decided to forgo compensation entirely

While we learned a lot from setbacks, we were also able to learn a great deal from our successes. For example, it was a blessing to have access to the survey tools available to BYU students through Qualtrics.com. This website gave us the power and flexibility to allow us to carry out the face pilot study easily, and we will continue to use it for the voice pilot and the voice-face experiment. One other miracle we had was the power of social media. We hoped to have enough subjects to make our results meaningful, but we didn’t know how to advertise effectively. We decided to make an open-enrollment Facebook group with the survey links, and invited friends. They then shared the links and within a day from the original posting we already had over twenty seven respondents. Since Phase One was just a pilot study to tighten up our selection of face stimuli, we are excited to see how many respondents our Experimental Phase will bring in!

The most rewarding aspect of our research was, of course, working with our faculty mentor, Dr. Wendy Smemoe². She proved our most valuable resource as she guided us through the process. She has ample experience in research and publication, particularly in the area of English variations and sociolinguistics. Every time we encountered one of the aforementioned roadblocks, it was she who would guide us to a solution. Yet she was not only a great resource for information—she also helped us begin to understand the “behind-the-scenes” aspects of research like IRB proposals, obtaining permissions, elimination of variables, etc. These are crucial areas which many undergraduates do not get the opportunity to experience.

The original methodology called for three phases: a pilot to help narrow down our selection of face stimuli, a pilot to help with voice clips, and then an experimental phase which pairs up the faces and voices in order to test our hypothesis. As of February 15, 2013 we have completed Phase One and are preparing to release Phase Two next week. Once we have analyzed the results of these two (and selected our final stimuli for the experiment) we will be able to release the Experimental Phase survey. Our plan is to finish collecting that data by March 15. We are excited to see the results, and plan to share them in the Journal of Language and Social Psychology in spring of this year. So far, the preliminary results seem to indicate that linguistic cues are more important than visual cues in racial profiling.

We would like to thank the Office of Research and Creative Activities as well as our donor Alan Harker for offering us the resources and opportunity to pursue this research. We also thank Dr. Smemoe for sharing her patience, guidance, and knowledge with us on this project.

References

1. We would like to thank the Office of Research and Creative Activities as well as our donor Alan Harker for offering us the resources to pursue this research. We also thank Dr. Smemoe for sharing her patience, guidance, and knowledge with us.
2. We would like to thank the Office of Research and Creative Activities as well as our donor Alan Harker for offering us the resources and opportunity to pursue this research. We also thank Dr. Smemoe for sharing her patience, guidance, and knowledge with us on this project.