Wednesday, 20 October 2010

Speakers with a foreign accent are perceived as less credible - and not just because of prejudice

Speakers with a foreign accent are perceived as less believable than native speakers. A new study shows this isn't just because of prejudice towards 'outsiders'. It also has to do with the fluency effect, one manifestation of which is our tendency to assume that how easily a message is processed is a mark of its truthfulness. The effort required to understand an accented utterance means that the same fact is judged as less credible when uttered by an accented speaker, compared with a native speaker. This remains true even if the accented speaker is merely passing on a message from a native speaker.

Shiri Lev-Ari and Boaz Keysar recruited 9 speakers to utter 45 trivia facts, such as 'A giraffe can go without water longer than a camel'. Three of the speakers were native (American) English speakers; three had mild foreign accents and originated from Poland, Turkey or Austria/Germany; and three had strong accents and were from either Korea, Turkey or Italy.

Twenty-eight undergrad participants rated the veracity of each of the spoken facts (which speakers uttered which facts varied from participant to participant in a balanced design). Crucially, participants were led to believe that the study was really about using intuition to judge facts. Also, it was made clear to them that the facts had been penned by the researchers - that the speakers were merely acting as messengers. To drill home this idea, the researchers also had the participants go through the charade of themselves uttering a few facts, ostensibly to be presented to other participants.

On a 0-14cm scale from 'definitely false' at one end to 'definitely true' at the other, the participants rated facts spoken by mild and heavily accented speakers as less believable than facts uttered by native English speakers (the mean ratings were 6.95, 6.84 and 7.59, respectively - a statistically significant difference).

What if participants are made aware that the difficulty they have processing a foreign accent could be interfering with their judgements? A second study with another 27 undergrads tested this very idea. It was similar to the first but this time participants were told the explicit aim of the study. Now, facts spoken by a speaker with a mild accent were judged to be just as credible as facts spoken by a native English speaker. However, facts spoken by a heavily accented speaker were still judged to be less true. It seems we can override our bias for assuming easily processed utterances are more truthful - but only up to a point. Also, it's worth remembering that in real life, prejudice towards foreign speakers is likely to augment the effects observed here.

'These results have important implications for how people perceive non-native speakers of a language, particularly as mobility increases in the modern world, leading millions of people to be non-native speakers of the language they use daily,' the researchers concluded. 'Accent might reduce the credibility of non-native job seekers, eye-witnesses, reporters or news anchors.'

ResearchBlogging.orgLev-Ari, S., and Keysar, B. (2010). Why don't we believe non-native speakers? The influence of accent on credibility. Journal of Experimental Social Psychology, 46 (6), 1093-1096 DOI: 10.1016/j.jesp.2010.05.025

Post written by Christian Jarrett (@psych_writer) for the BPS Research Digest.

Link to related open-access feature article in The Psychologist magazine [Hyunjin Song and Norbert Schwarz describe some fascinating findings on how fluency affects judgement, choice and processing style].


Neuroskeptic said...

I suspect though that under some conditions certain accents could increase credibility.

For example there is a stereotype (probably dating back to Einstein and the early quantum physicists) that Germans are good at physics.

I can imagine that someone explaining the Theory of Relativity in a German accent, might seem more credible than someone doing it in an English accent...

Hadas Shema said...

I agree with Neuroskeptic. In Israel, American or British accent often increases one's credibility.

Unknown said...

Hi Neuroskeptic and Hadas - good point. Would be interesting to pitch those cultural stereotypes against the influence of the fluency effect reported here.

Anonymous said...

fluency decreases the credibility is essentially a kind of prejudice I think. lack of fluency may activates the association with unfamiliar people in our mind who is not always credit. (Bo from Hong Kong)

skm said...

I am curious about the cultural bias brought up by Neuroskeptic.

I also have a couple of questions.

1) If a subject matter expert is talking to cohorts in the same subject, does that wash out the fluency result?

2) If you vary volume or static or other auditory qualities of a voice speaking in the listener's accent, do truthfulness ratings vary along with the audio quality?

Anonymous said...

Seeking Canadian women to participate in a research study on facial preferences for a 1 in 5 chance to win a $10 gift card!

We would like to invite all women currently residing in Canada to participate in a study about the perception of men’s faces. Participation involves viewing a series of photographs of men and responding to corresponding questions. Participants will also be asked to complete items assessing background information, sexual attitudes, personality characteristics, and mate preferences. The study takes approximately 20 minutes to complete. Those completing the study will have their name entered in to a drawing for a 1 in 5 chance to win a $10 gift card.

If you are interested, please click on the link below

This project is on file with the UNB Research Ethics Board
(REB #2010-100).

mich said...

Slightly related to Neuroskeptic's comment, I was wondering why, with all other things being so tightly controlled, they used "Poland, Turkey or Austria/Germany" vs "Korea, Turkey or Italy". Using Turkey twice could disentangle cultural bias if the one Turkish person clearly spoke with less of an accent than the other, except for the fact that their voices would be different on a plethora of other levels. As such, in a perfect world, I would think the study would have been of greater interest if the foreign accents would have been from the same country (or obviously, from different countries, but then for a 2 x 2 design). Perfect world, not real world, though, as this study does beautifully demonstrate!

Robby said...


I think if a foreign English speaker speaks fluently despite having accent he will be judged by his merits rather than by his accent - as far as his accent doesn't interfere with listener's understanding.

I also think the study results mentioned in this blog post aren't conclusive as the people surveyed were expressing their first impression about the truthfulness of foreigners' statements.

In real life when foreigners are involved in a longer discussion - as it will indeed happen during a job interview, for instance - they'll be rather judged on their specific knowledge, intelligence (of course, as far as their English is fluent).

My main point - fluent English doesn't necessarily mean speaking with perfect English pronunciation and vice versa.



PsychologyinRome said...

Not having read the actual research I'm going to throw this into the do we know the effect is to do with the foreign accent and not anything to do with other factors such as ingroup vs outgroup biases? As it says in the blog post "Also, it's worth remembering that in real life, prejudice towards foreign speakers is likely to augment the effects observed here"...

Unknown said...

hi PsychologyinRome
the idea was that the influence of out-group prejudice was ruled out in this study by having the accented speakers pass on facts written by the researchers - a detail that was made clear to the participants.

Alex Truck said...

This was a very interesting article for me to read. As somebody with a heavy Russian accent, I must attest that I have noticed the effects that it had on my credibility... I try to combat it by establishing my credentials and well working on improving my speech.

Anonymous said...

My wife and I are both hispanic. I grew up in the US and have a minimal accent, my wife grew up abroad. We have personally experienced this effect, in public I receive more attention and validity than my wife (she has a professional yet thick accent). We are both teachers and live fighting for our credibility on a daily basis.

Unknown said...

The title of this article is exaggerated, factually incorrect, and immediately contradicted in the first sentence of the abstract. If Elsevier wants us to pay arms & legs to be able to read their journals, and justifies the high cost with their superior quality control, how come that their peer-review let something like that pass? Is it because this way, the title sounds more sensational? Apart from not matching with the title, the differences in believability the authors found in this study are actually quite small, *especially* for the native vs. mild accent. They may be significant, but we all know (we do, right?) that there is a difference between significance and size. This type of article might contribute to the growing belief that social psychology should "clean up its act", to quote Kahnemann.

Unknown said...

I am from Jamaica and I have gotten respect from my listeners and I come across very credible
Thank you

Post a Comment

Note: only a member of this blog may post a comment.