Stories
Slash Boxes
Comments

SoylentNews is people

posted by CoolHand on Friday May 29 2015, @06:05AM   Printer-friendly
from the could-you-please-speak-in-english dept.

Racial stereotypes and expectations can impact the way we communicate and understand others, according to UBC research. The new study, published in the Journal of the Acoustical Society of America, highlights how non-verbal "social cues" - such as photographs of Chinese Canadians - can affect how we comprehend speech.

"This research brings to light our internal biases, and the role of experience and stereotypes, in how we listen to and hear each other," says Molly Babel, the paper's lead author and an assistant professor with UBC's Department of Linguistics.

One of the study's tasks involved participants from the UBC community transcribing pre-recorded sentences amid background static. The sentences were recorded by 12 native speakers of Canadian English. Half of the speakers self-identified as White, and the other half self-identified as Chinese. All speakers were born and raised in Richmond, B.C., which is south of Vancouver.

The pre-recorded sentences were accompanied by either black and white photos of the speakers, or by an image of three crosses. Overall, listeners found the Chinese Canadians more difficult to understand than the White Canadians - but only when they were made aware that the speaker was Chinese Canadian due to the photo prompt.

Participants were also asked to rate the strength of the accents of the speakers. They were asked to listen to two sentences from each speaker - one accompanied by the speaker's photo, the other by an image of crosses. "Once participants were aware that they were listening to a White Canadian, suddenly the candidate was perceived as having less of a foreign accent and sounding more like a native speaker of Canadian English," says Babel.

"It tells us as listeners that we need to be sensitive about the stereotypes that we carry," notes Jamie Russell, the study's co-author who was an undergraduate honours student in UBC's Department of Linguistics during the project.

http://phys.org/news/2015-05-racial-stereotypes-impact.html

[Abstract]: http://scitation.aip.org/content/asa/journal/jasa/137/5/10.1121/1.4919317

[Source]: http://news.ubc.ca/2015/05/26/how-racial-stereotypes-impact-the-way-we-communicate/


[Editor's Comment: Original Submission]

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Interesting) by theluggage on Friday May 29 2015, @10:31AM

    by theluggage (1797) on Friday May 29 2015, @10:31AM (#189603)

    Half of the speakers self-identified as White, and the other half self-identified as Chinese

    [Actually, Chinese Canadian, non Chinese - from TFP]. One of the major stumbling blocks to this sort of research: of course, ethically, you have to accept the subjects self-identification of their ethnicity. Unfortunately, that leaves you with no solid basis for assuming that their self-identified ethnicity reflects their actual accent or facial features.

    but only when they were made aware that the speaker was Chinese Canadian

    Checked this in TFP - the intelligibility figures for audio-only are indeed within standard error, while the audio+photo ones are a couple of standard errors apart. Surprising - if there was any genuine difference between their accents you'd expect some sort of effect in the audio-only, with the 'photo' effect showing up as a bias.

    If you specifically ask people to identify a difference that isn't there, is it surprising that they'll latch on to whatever evidence you offer them, however slight or irrelevant? Does this actually have any significance in real-life situations?

    Actual paper is TLDNR for a quick soylent post, but interesting wrinkle that caught my eye was the ethnic make-up of the 40 trial participants:

    (Asian = 14, Asian and Pacific Islander = 2, Asian and White = 4, Black = 1, Pacific Islander = 1, South Asian = 3, White = 15)

    Problem with a N=40 trial is that while you might scrape by with 40 data points, you don't get enough of any one ethnic group to dig into that.

    Starting Score:    1  point
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 0) by Anonymous Coward on Friday May 29 2015, @01:43PM

    by Anonymous Coward on Friday May 29 2015, @01:43PM (#189653)

    [Actually, Chinese Canadian, non Chinese - from TFP]. One of the major stumbling blocks to this sort of research: of course, ethically, you have to accept the subjects self-identification of their ethnicity.

    This is debatable at best. Is it really ethical to avoid hurting people's feelings if more accurate results could be used to better society, including the lives of the very people who get upset?

    • (Score: 0) by Anonymous Coward on Friday May 29 2015, @06:46PM

      by Anonymous Coward on Friday May 29 2015, @06:46PM (#189777)

      The ends rarely justify the means. So long as nobody gets hurt, directly or indirectly, its probably fine, but even that can start you down a slippery slope.

  • (Score: 2) by theluggage on Friday May 29 2015, @04:28PM

    by theluggage (1797) on Friday May 29 2015, @04:28PM (#189723)

    Should read "[Actually, Chinese Canadian, not Chinese - from TFP]" - i.e. the self-identification is about ethnicity, not nationality.