You are on page 1of 16

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/230162703

A meta‐analysis of the effects of speakers' accents on interpersonal


evaluations

Article  in  European Journal of Social Psychology · February 2012


DOI: 10.1002/ejsp.862

CITATIONS READS

168 2,411

5 authors, including:

Jairo N. Fuertes William Gottdiener


Adelphi University City University of New York - John Jay College of Criminal Justice
74 PUBLICATIONS   2,493 CITATIONS    41 PUBLICATIONS   1,010 CITATIONS   

SEE PROFILE SEE PROFILE

Howard NOTE: SCOtton is not the co-author Giles


University of California, Santa Barbara
406 PUBLICATIONS   19,516 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Evaluating factors underlying schizoid detachment View project

communication accommodation theory, language attitudes, comm & aging, police-community relations, intergroup comm & dance...and dress styles View project

All content following this page was uploaded by Howard NOTE: SCOtton is not the co-author Giles on 10 September 2020.

The user has requested enhancement of the downloaded file.


European Journal of Social Psychology, Eur. J. Soc. Psychol. 42, 120–133 (2012)
Published online 26 October 2011 in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/ejsp.862

Research article
A meta-analysis of the effects of speakers’ accents on interpersonal evaluations

JAIRO N. FUERTES1*, WILLIAM H. GOTTDIENER2,3, HELENA MARTIN4, TRACEY C. GILBERT 5 AND


HOWARD GILES6
1
Adelphi University, Garden City, USA; 2John Jay College, City University of New York, New York, USA;
3
St. Luke’s-Roosevelt Medical Center, New York, USA; 4University of Maryland, College Park, USA; 5Fordham
University, New York, USA; 6University of California, Santa Barbara, USA

Abstract

This paper reports a meta-analysis of the empirical literature on the effects of speakers’ accents on interpersonal evaluations.
Our review of the published literature uncovered 20 studies that have compared the effects of standard accents (i.e., the accepted
accent of the majority population) versus non-standard accents (i.e., accents that are considered foreign or spoken by minorities)
on evaluations about the speakers. These 20 studies yielded 116 independent effect sizes on an array of characteristics that were
selected by the original researchers. We classified each of the characteristics as belonging to one of three domains that have been
traditionally discussed in this area, namely status (e.g., intelligence, social class), solidarity (trustworthiness, in-group–out-group
member), and dynamism (level of activity and liveliness). The effect was particularly strong when American Network accented
speakers were compared with non–standard-accented speakers. These results underscore prior research showing that speakers’
accents have powerful effects on how others perceive them. These and other results are discussed in the context of the literature
along with implications for future research in this area. Copyright © 2011 John Wiley & Sons, Ltd.

Communication is power (Ng & Bradac, 1993); the reader Giles & Bourhis, 1973), experimental procedures have typi-
may have heard this adage at some point. Decoding language cally been adopted—particularly the matched-guise technique
varieties is essential to listeners forming impressions of speakers, (Lambert, 1967)—in manipulating accent use by the same
and listeners use a plethora of vocal features, such as pitch, into- bi-dialectal speakers for control purposes. A simplistic, yet
nation, and speech rate in assessing speakers’ backgrounds, representative study using the matched-guise technique can be
characters, and personalities (Bradac, 1990; Scherer & Giles, found in the literature is as follows. A group of participants
1979). The present report is a meta-analysis of studies that have are randomly assigned to watch a videotape of someone speak-
examined variations in speech accents and their effect on ing. Participants in the experimental group watch a videotape of
interpersonal evaluations. Research pertaining to this topic has a person speaking with a non-standard accent. Participants in
traditionally compared the effects of standard accents (i.e., the the control group watch a videotape of a person speaking with
accepted accent of the majority population) versus non-standard a standard accent. The persons in the videotape are matched
accents (i.e., accents that are considered foreign or spoken by on various characteristics including gender, race, age, etc. After
minorities) on evaluations about the speakers (Cargile & Bradac, watching their respective videotapes, the participants then
2001; Giles & Powesland, 1975; Ryan & Giles, 1982). In the complete a series of questionnaires that measure the constructs
current meta-analysis, we identified the statistics published in of solidarity, status, and/or dynamism.
20 studies and computed 116 independent effect sizes. We also Beyond personality judgments, variations in language use
classified each as belonging to one of three domains that have have also been shown to affect compliance-gaining, social
been traditionally investigated in this area (see Giles, 1970; Giles decision making, and others’ behavioral reactions (e.g., coop-
& Billings, 2004). erativeness) toward accented speakers (for a review, see Giles
& Billings, 2004). Social evaluations are triggered by the
speaker’s use of language, and the listener uses as a reference
Background and Rationale the normative and dominant accent in the country in which the
communication takes place. Evidence of the power of verbal
Although invoking heated debate over the years in socio- cues in shaping perceptions and evaluations of speakers has
linguistic circles (e.g., Garrett, Coupland, & Williams, 2003; been documented, for example, Krauss, Freyberg, and

*Correspondence to: Jairo N. Fuertes, PhD, ABPP, The Derner Institute of Advanced Psychological Studies, Adelphi University, Hy Weinberg Center-Rm 319
158 Cambridge Avenue, Garden City, NY 11530, USA.
E-mail: Jfuertes@adelphi.edu

Copyright © 2011 John Wiley & Sons, Ltd. Received 23 April 2010, Accepted 21 September 2011
Effects of speakers’ accents 121

Morsella (2002) showed that listeners can distil personal been documented in various settings such as work, school,
information about speakers based on their recorded speech, housing, medical, counseling, and legal contexts (Lippi-Green,
even their physical characteristics, such as age and height. 1997; Ryan & Giles, 1982). For example, Blair and Connor
In this study, we examine the extant research conducted (1978) examined the relationship between employers’ percep-
thus far on speech accents effects, which represent variations tions of speech style and evaluations of job competence.
in the pronunciation of the same language. As mentioned Results indicated that African-Americans speaking “Black
earlier, the standard accent is generally considered that which English” were rated 17%–28% lower on perceived job compe-
is spoken by the majority of the population and/or is associated tence than “standard-accented” Black individuals, and 24%–39%
with high socioeconomic status, power, and use in the media lower than Whites who spoke with a standard accent. Thus,
in any given country (Giles & Billings, 2004). The evaluation use of a non-standard accent negatively affect employee evalua-
made by the listener is not only about the accent itself, for tions and contribute to decisions that are discriminatory and
example, in terms of its aesthetics (see Giles & Niedzieksi, possibly even illegal.
1998), but is used as an immediate marker by the listener to Yet at the same time, some studies have shown non-standard
evaluate the person who is speaking and/or the social group accents (such as Appalachian English and regional varieties
that is represented by an accent variety. Vocal cues can trigger of British English) to have “covert prestige” (Trudgill, 1974)
idiosyncratic personal attributions about a speaker and be in that, on occasion, speakers with these accents have been
invoked by listeners to reduce interpersonal uncertainty and upgraded on solidarity traits and in-group cohesion (e.g.,
guide impression management strategies toward the speaker Marlow & Giles, 2008). In addition, studies have demonstrated
(Berger & Bradac, 1982). However, when it comes to speak- that the social context (e.g., informal versus formal situations)
ers’ accents, there is a consensus among language attitude in which accent use is adopted has also been shown to
scholars (e.g., Giles & Ryan, 1982; Gudykunst & Ting-Toomey, affect social evaluations; for instance, speakers of standard
1990) that these provide the listener with an intergroup frame accented varieties are often perceived as socially less
for interpreting the situation, that is, more in terms of speak- appealing when talking about casual topics in informal
ers’ social than their personal identities (see Giles, Reid, & settings (Ryan, 1979).
Harwood, 2010; Tajfel & Turner, 1986). Indeed, the initial There is a long-standing literature showing that others’
development of the matched-guise technique presented by voices are potent cues for listeners to socially categorize,
Lambert (e.g., 1967) was built on the following theoretical stereotype, and form impressions of speakers (e.g., Ko,
premise: language features ! social categorization ! trait Judd, & Stapel, 2009; Krauss et al., 2002; Scherer, 1979).
attributions (see also, Robinson, 1972). Of all the vocal cues investigated (e.g., pitch and speech
The non-standard accent is generally considered a foreign rate), accents have received, arguably, the most empirical at-
accent or one used by a minority or lower socioeconomic tention and been accorded perhaps the most applied social
group. For example, the standard accent in the US is the one significance across cultures. It has been found that many
spoken by TV news anchors and is considered a derivative children are already sensitive to dialectal variations by the
of a mid-western accent. An example of a non-standard accent age of 5 months (Nazzi, Jusczyk, & Johnson, 2000) and that,
in the US would be a Spanish accent. Prior research has shown other speakers’ accents are evaluated more saliently, even
that accents affect listeners’ perceptions of speakers along than their perceived race, by 5-year-olds and 6-year-olds
three (variably labeled) main dimensions: status, solidarity, (Kinzler, Dupoux, & Spelke, 2007; Kinzler, Shutte, DeJesus,
and dynamism (Giles & Billings, 2004). We used these dimen- & Spelke, 2009). Referring to their own empirical work with
sions to organize the current meta-analysis. Perceived status adults, Rakić, Steffens, and Mummendey (2011) found that
includes evaluations about the speaker’s intelligence, compe- ethnic categorization is influenced substantially by speech
tence, ambition, education, and social class. Solidarity accents and that subjects can categorize accurately the eth-
includes evaluations of the speaker’s similarity to the listener, nicity of speakers on the basis of accent alone, even when
attractiveness, benevolence, and trustworthiness. Dynamism they are presented with discordant visual stimuli about the
refers to speaker’s level of activity and liveliness (see Giles ethnicity of the speakers. Rakić et al (2011) noted that
& Billings, 2004; Mulac, Hanley, & Prigge, 1974; Zahn & “it was rather irrelevant for participants what targets looked
Hopper, 1985). like; it mainly mattered whether they were speaking with
Standard accents, such as Standard British English (i.e., an accent or not. In this case it was almost as if partici-
also called the “Queen’s accent,” “BBC accent,” and in the pants became blind to the visual category information in the
literature “Received Pronunciation accent” or RP), Standard presence of more meaningful auditory category information
American Network, Parisian French, and Castillian Spanish, (p. 24)”.
have been shown to positively affect listeners’ evaluations of A considerable portion of experimental work on social
speakers’ education, success, social status, attractiveness, evaluations of accents has been devoted to the interdependent
personality, and the similarity with the listener (e.g., Giles, variable dimension, speaker’s standard–non-standardness (see
1970; Giles & Sassoon, 1983; Stewart & Ryan, 1985). They Ryan et al., 1982; for a discussion of standard–non-standard
have also been shown to affect listeners’ recall and compre- distinctions, see Edwards and Jacobsen, 1987, and for differ-
hension of content, even when speech is intelligible (Ryan & ent standards evolving in different contexts, see Kristiansen,
Sebastian, 1980). In the US and other English-speaking 2001). Generally, standard accents grant people access to
countries, accents have been shown to stimulate stereotypes political, economic, and educational forums and opportunities,
and instigate discriminatory behavior against accented speak- whereas non-standard accents impart stigma upon speakers of
ers (Ryan, Giles, & Sebastian, 1982). These behaviors have them (Gluszek & Dovidio, 2010b).

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
122 Jairo N. Fuertes et al.

The early work conducted by Lambert (1967) invoked the effects of accents in interpersonal evaluations. This is a
dependent variables of competence (or status) on the one hand pertinent issue given the lack of attention afforded to accents
and solidarity (or social attractiveness) on the other for judging in academic reviews of interpersonal communication and
speakers on audiotape. For many years, researchers adopted impression formation where, at best, the topic is given
these rating dimensions as really the only bases of social short-shrift. Despite the 40 years or so of cross-discipline work
judgment (see, however, Bourhis, Giles, & Tajfel, 1973). on this topic, a mere perusal of the indexes of recent
Later, and with the aim of determining which socially Handbooks and Encyclopedias in social psychology (e.g.,
evaluative dimensions were predominant, Zahn and Hopper Craighead & Nemeroff, 2004; Hogg & Cooper, 2003) and
(1985) conducted a comprehensive language attitudes study communication science (Burleson & Greene, 2003; Donsbach,
involving virtually all the rating scales used in prior research. 2008; Sanders & Fitch, 2004; Ventola & Antos, 2008) show
Although they confirmed the competence–solidarity division— little evidence of interest in this topic. Inexplicably, neither
which can now, arguably, be seen as a precursor to the the Handbook of Social Psychology (Delamater, 2003) or the
competence versus warmth dichotomy inherent in the stereotype Handbook of Nonverbal Behavior (Manusov & Patterson,
content model (see Cuddy, Fiske, & Glick, 2008, and Fiske, 2006) index this topic, nor do recent monographs on the
Cuddy, Glick, & Xu, 2002), they established a third orthogonal sociopsychological facets of language usage (e.g., Holtgraves,
dimension, namely dynamism (see also Mulac, 1975; Mulac 2002; Robinson, 2003).
et al., 1974) that has also been featured somewhat in more The potential impact of the findings from this meta-analysis
recent work. cannot be undervalued. If communication is power, then how
Research using the matched-guise technique has found that a person speaks and to whom that person is speaking can have
non-standard language varieties are evaluated less favorably considerable implications for almost all interpersonal engage-
than their standard counterparts on competence (e.g., intelli- ment from teacher–student dyads, to interactions between
gence, ambition, confidence) as well as dynamism (e.g., lively, physicians and patients, and politicians and their constituents.
enthusiastic, talkative) traits (for overviews, see Giles &
Billings, 2004; Giles & Edwards, 2010; Lippi-Green, 1997). Hypotheses
Indeed, the stronger or broader the perceived non-standard
accent, the more negative the social evaluations (Gluszek, The first hypothesis of the current meta-analysis was that there
Newheiser, & Dovidio, 2011; Ryan, Carranza, & Moffie, would be significant differences in ratings of speakers across
1977) that ensue, with the perceived quality of the message the three dimensions combined, with standard-accented speakers
content and its comprehensibility being downgraded as well being rated significantly higher than non–standard-accented
(Giles, 1973a; Gluszek & Dovidio, 2010b). On the contrary, speakers. The second hypothesis follows from the first hypothe-
non-standard speakers have been upgraded, albeit not always sis but is more specific and was that standard-accented speakers
consistently, on solidarity traits (e.g., social attractiveness, would be rated significantly higher than the non–standard-
benevolence). accented speakers on each of the three dimensions of status,
Although individual studies in a variety of nations provide solidarity, and dynamism. The third hypothesis is that the RP
evidence that accents affect a wide variety of interpersonal and standard accent would be rated higher than the American
intergroup attitudes and behaviors, stronger evidence of such standard accent when compared against non-standard accents.
an effect can be accomplished by conducting a meta-analysis Two additional questions that were exploratory in nature but
across studies (Rosenthal & DiMatteo, 2001). A meta-analysis not sufficiently clear to warrant hypotheses were about the
is a statistical review of the literature in which a researcher differences in ratings of standard versus non-standard accents
reviews all, or a sample, of the research in the area of interest as a function of setting (i.e., the different settings studied by
and then computes standardized statistics, called effect sizes (a researchers) and the country where the study took place. By
measure of how strong the relation is between two variables) settings we mean experimental conditions such as education,
from the studies. A meta-analysis of studies examining the employment, sales, and job interviews, and the countries in
effects of accents on interpersonal evaluations is needed now which the studies were conducted included the United States,
for several reasons. First, a meta-analysis will allows us to Australia, Argentina, Turkey, and the United Kingdom.
quantitatively determine if speakers’ accent affects listener
evaluations across the extant empirical literature on this topic.
Second, a meta-analysis will enable us to determine to what
METHOD
degree speaker accents affects listener evaluations. Third, a
meta-analysis might resolve conflicting findings that exist
across the empirical literature (e.g., if status and solidarity Literature Search
ratings are equally affected by non-standard accents). Fourth
and finally, a meta-analysis will help determine if there are The goal of the literature search was to identify all relevant
variables that moderate the effects of speakers’ accents on studies on how accents affect interpersonal evaluations and
listeners’ evaluations (e.g., the experimental social setting include those with statistics that could yield an effect size.
examined by the researchers). None of these are possible to The studies came from a wide variety of subject areas and
do with a qualitative review of the literature. databases. Computer databases in the areas of business and
The use of multiple published studies from across a wide economics, communication, education, medicine and health,
span of years, journals, and disciplines will then provide psychology, sociology, and social work produced relevant
standardized data about the direction and strength of the articles for the current meta-analysis. Specifically, the

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Effects of speakers’ accents 123

following subject databases were used: Business Source Two independent graduate students were trained as raters to
Premier, Communication and Mass Media Complete, Communi- collect and code the studies. To reduce the potential for human
cation Abstracts, Educational Abstracts, Educational Resources error, the two raters compared the data obtained after every
Information Center, Health Source: Nursing/Academic Edition, five studies were coded. Inter-rater agreement was achieved
PsycINFO, PsycARTICLES, Social Science Citation Index, via consensus discussion. At each interval, the coding results
Sociological Abstracts, and Social Work Abstracts. Large were compared, and very few errors were identified, and they
general collections and scholarly journal collections including were corrected during this process.
Academic Search Premier, JSTOR, and LEXIS-NEXIS were
also searched. Key words and phrases used to identify the subject Effect Sizes
of interest and population of interest (that being English-
speaking individuals) were English accent(s) or dialect(s). Fur- Cohen’s d was the statistic used to compute effect sizes in the
ther narrowing key words or phrases used to identify studies current study, and it is the difference in means between groups
pertaining to interpersonal judgments were evaluation(s) and when the difference is divided by the pooled standard
its derivatives (e.g., attitude, reactions, social class, social attrac- deviation of the groups (Cohen, 1988). Cohen (1988, p. 9–10)
tiveness, perceptions, judgments, competence, solidarity). In defines an effect size as “the degree to which the phenomenon
addition, the subcategory of comprehension and its derivatives is present in the population,” or “the degree to which the null
(e.g., learning and teaching) were also searched in certain hypothesis is false. . . .the null hypothesis always means that
databases. The references for some obtained journal articles the effect size is zero.” We calculated effect sizes in the standard
were also reviewed to identify studies fitting our criterion for manner directly using descriptive statistics or indirectly from
selection or inclusion. Whenever possible, hand searches of omnibus statistical tests or p values when a report did not
the journals that frequently published accent studies were also contain sufficient descriptive statistical information to calculate
conducted. Furthermore, the extensive bibliographic lists from effect sizes directly. The d statistic can be derived algebraically
certain language, sociolinguistic, and communication books from omnibus statistics and p values. Sometimes results were
were searched. reported as simply being non-significant. Following Cooper
(1989), we coded these statistics to have effect sizes of zero.
Inclusion Criteria Inexact p values such as p < 0.01 were reported at the next
lowest value (e.g., p = 0.009).
Each of the studies included in this meta-analysis had to meet Cohen (1988) defined effects sizes (d statistic) as small if
certain inclusion criteria. The first criterion is that the study they are below 0.2 (or below a 58th percentile standing),
had to address accented English. We chose the English lan- medium if they are between 0.2 and 0.7 (0.7 would indicate
guage studies for several reasons. First, English is the primary a 76th percentile standing), and large if they are above 0.8
language of the current team of researchers. Second, despite (or above the 79th percentile). The percentile standing refers
the historical duration of interest in the effects of speech accents to differences between the groups. For example, an effect size
on interpersonal perceptions, the total population of studies that of 0.0 would indicate absolutely no differences between
exists is approximately 153, and almost all scholarly work that groups A and B and that the distribution of their scores is
has been conducted on this topic has been conducted in English. identical (meaning that the bell-shaped distributions of the
And third, a potential problem arises when integrating studies scores for groups A and B have 100% overlap and have the
with other languages. Language is strongly linked with culture, same mean). An effect size of 0.8 indicates that the mean of
and if we were to test our hypotheses using studies conducted group A is at the 79th percentile of group B (whose normal
with other languages, then cultural differences might add a distribution would have a mean at the 50th percentile), and
problematic confound in interpreting the findings. The second there is only 53% overlap in the distribution of scores. An
criterion was that the study must include statistical information effect size of 1.0 indicates that the mean of group A is at the
sufficient to compute an effect size (e.g., means, standard devia- 84th percentile of group B, and there is less than 45% overlap
tions, group sizes, F values, t values, r values). in the distribution of scores for groups A and B.
The higher the percentile for group A, the more powerful
Coding Procedure the mean differences between it and group B and the lower
the overlap in distribution of scores for groups A and B.
Following the guidelines published by Lipsey and Wilson Thus, more weight can be given to whatever factor is con-
(2001), a coding manual (see Appendix A that contains sam- tributing to the higher mean score for group A. In the current
ple items) was created that included information from each study, a higher effect size indicates the extent to which
study and that was entered into a database. Samples of the standard accent is preferred along the dimensions on which it
characteristics rated in each study are presented in Table 1, is rated, in comparison with ratings on the same dimensions
along with our classification of each as being a status, soli- for non-standard accents. In other words, positive effect sizes
darity, or dynamism dimension. In our classification process indicate that the standard accent was preferred by listeners
of grouping characteristics into one of three dimensions, we and negative effect sizes indicate that the non-standard accent
were led by Howard Giles in making a first categorization, was preferred by listeners.
given his extensive experience in this area of research. How- Most research is multivariate, and this creates potential
ever, the first and third authors examined his categorization problems of maintaining independence across effect sizes in
independently and made minor modifications to which the a meta-analysis. We were interested in three potential
three authors agreed. outcomes—solidarity, status, and dynamism—which have

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Table 1. Individual Effects Sizes by Characteristic
124
Non-standard
Authors & country Characteristics rated d Significance level Context Dimension accent Standard accent
Arthur et al., 1974 USA Carefree-ambitious 0.79 0.00 No context Solidarity “Chicano English” “Standard English”
Arthur et al., 1974 Dependable–unreliable R 0.88 0.00 No context Solidarity
Arthur et al., 1974 Dishonest–honest 0.84 0.00 No context Solidarity
Arthur et al., 1974 Educated–uneducated R 1.89 0.00 No context Status
Arthur et al., 1974 Friendly–hostile R 0.44 0.04 No context Dynamism
Arthur et al., 1974 Hardworking–lazy R 0.71 0.00 No context Dynamism
Jairo N. Fuertes et al.

Arthur et al., 1974 Hi self-image–lo self-image R 1.56 0.00 No context Solidarity


Arthur et al., 1974 Intelligent–stupid R 1.52 0.00 No context Status
Arthur et al., 1974 Openminded–narrowminded R 1.13 0.00 No context Solidarity

Copyright © 2011 John Wiley & Sons, Ltd.


Arthur et al., 1974 Out of it–with it 1.51 0.00 No context Solidarity
Arthur et al., 1974 Quiet–talkative 0.28 0.18 No context Dynamism
Arthur et al., 1974 Serious–fun R 0.60 0.00 No context Solidarity
Arthur et al., 1974 Strong–gentle R 0.56 0.01 No context Dynamism –
Arthur et al., 1974 Untalented–gifted 1.27 0.00 No context Status
Arthur et al., 1974 Working class–upper class 1.92 0.00 No context Status
Brown et al., 1985 UK Active–passive 0.29 0.05 Education Dynamism “South East Welsh” “Standard British Accent”
Brown et al., 1985 Ambitious–unambitious 0.52 0.00 Education Status
Brown et al., 1985 Good looking–plain 0.40 0.01 Education Status
Brown et al., 1985 Happy–sad 0.31 0.03 Education Solidarity
Brown et al., 1985 Intelligent–not intelligent 0.45 0.00 Education Status
Brown et al., 1985 Just–unjust 0.34 0.02 Education Solidarity
Brown et al., 1985 Religious–unreligious 0.40 0.01 Education Status
Callan et al., 1983 Australia Solidarity 0.00 1.00 no context Solidarity “Greek-Australian” “Anglo-Australian”
Callan et al., 1983 Status 0.85 0.00 no context Status
Cargile & Giles, 1998 USA Attractiveness 0.36 0.00 Education Status “Moderate Japanese” “Standard American”
Fuertes & Gelso, 2000 USA Counselor rating 0.02 0.90 Counseling Status “Spanish” “Standard American”
Fuertes & Gelso, 2000 Willingness 0.20 0.18 Counseling Solidarity
Fuertes & Gelso, 2000 Working alliance 0.13 0.37 Counseling Solidarity
Gallois et al., 1984 Australia Solidarity 0.00 1.00 no context Solidarity “Aboriginal” “White Accent Speakers”
Gallois et al., 1984 Status 0.45 0.00 no context Status
Giles et al., 1975 UK Intelligent–dull 0.87 0.00 Employment Status “Birmingham” “RP”
Giles et al., 1995 USA Attractiveness 0.00 1.00 Education Status “Hispanic accent” “Anglo American”
Giles et al., 1995 Dynamism 0.00 1.00 Education Dynamism
Giles et al., 1995 Superiority 0.61 0.00 Education Status
Giles et al., 1981 UK Aggressiveness 0.81 0.03 Employment Dynamism “South Welsh” “RP”
Giles et al., 1981 Generosity 0.90 0.02 Employment Solidarity
Giles et al., 1981 Good naturedness 1.21 0.00 Employment Solidarity
Giles et al., 1981 Intelligence 0.69 0.07 Employment Status
Giles et al., 1981 Kind heartedness 0.73 0.05 Employment Solidarity
Giles et al., 1981 Self-assurance 0.58 0.12 Employment Solidarity
Giles et al., 1981 Sincerity 1.06 0.01 Employment Solidarity
Giles et al., 1981 Social status 0.90 0.02 Employment Status
Gill, 1994 USA Aesthetic 0.52 0.05 Education Status “Malaysian” “Standard North American English”

(Continues)

Eur. J. Soc. Psychol. 42, 120–133 (2012)


Table 1. (Continued )
Non-standard
Authors & country Characteristics rated d Significance level Context Dimension accent Standard accent
Gill, 1994 Dynamism 0.52 0.05 Education Dynamism
Gill, 1994 Sociointellectual 0.52 0.05 Education Status
Gill & Badzinski, 1992 USA Favorable rating 0.68 0.01 Education Solidarity “Egypt/Australian” “American”
Green, Scott, Argentina Ranking 1.34 0.00 no context Status “Estuary” “RP English”
vs& Rosewarne, 2002
Hensley, 1972 USA Ambition 1.08 0.00 no context Status “Non-standard Negro “Standard English”
English”
Hensley, 1972 Considerate of others 1.15 0.00 no context Solidarity
Hensley, 1972 Easy going 0.67 0.00 no context Solidarity

Copyright © 2011 John Wiley & Sons, Ltd.


Hensley, 1972 Friendly 1.10 0.00 no context Dynamism
Hensley, 1972 Good looking 0.82 0.00 no context Status
Hensley, 1972 Happy 0.52 0.00 no context Solidarity
Hensley, 1972 Hardworking 1.25 0.00 no context Dynamism
Hensley, 1972 Honest 1.23 0.00 no context Solidarity
Hensley, 1972 Intelligent 1.63 0.00 no context Status
Hensley, 1972 Loyal 1.26 0.00 no context Solidarity
Hensley, 1972 Lucky 0.97 0.00 no context Solidarity
Hensley, 1972 Things in common 0.40 0.01 no context Solidarity
Hensley, 1972 Unselfish 0.67 0.00 no context Solidarity
Kamisli & Dugan, 1997 Turkey Ambitious–laissez fare R 0.71 0.00 Employment Status “African-American “Eastern American English”
English”
Kamisli & Dugan, 1997 Determined–unsure R 1.16 0.00 Employment Status
Kamisli & Dugan, 1997 Dishonest–honest 0.66 0.00 Employment Solidarity
Kamisli & Dugan, 1997 Entertaining–boring R 0.70 0.00 Employment Solidarity
Kamisli & Dugan, 1997 Industrious–lazy R 0.96 0.00 Employment Dynamism
Kamisli & Dugan, 1997 Intelligent–dull R 0.70 0.00 Employment Status
Kamisli & Dugan, 1997 Irritable–good natured 0.40 0.03 Employment Solidarity
Kamisli & Dugan, 1997 Kindhearted–cold hearted R 0.79 0.00 Employment Solidarity
Kamisli & Dugan, 1997 Self-confident–shy R 0.97 0.00 Employment Dynamism
Kamisli & Dugan, 1997 Sociable–unsociable R 1.00 0.00 Employment Solidarity
Kamisli & Dugan, 1997 Talkative–restrained R 0.97 0.00 Employment Dynamism
Kamisli & Dugan, 1997 Unpopular–popular 0.69 0.00 Employment Solidarity
Kamisli & Dugan, 1997 Unreliable–reliable 0.40 0.04 Employment Solidarity
Levin et al., 1994 UK Ambitious 0.78 0.00 No context Status “Mild South East “RP”
Welsh”
Levin et al., 1994 Colloquial 1.43 0.00 No context Status
Levin et al., 1994 Difficult to understand 0.69 0.00 No context Status
Levin et al., 1994 Dominant 0.54 0.01 No context Status
Levin et al., 1994 Easy to listen to 0.57 0.01 No context Solidarity
Levin et al., 1994 Formal 1.26 0.00 No context Status
Levin et al., 1994 Intelligent 1.20 0.00 No context Status
Levin et al., 1994 Less relaxed 0.49 0.02 No context Solidarity
Levin et al., 1994 Long words 1.02 0.00 No context Status
Effects of speakers’ accents

(Continues)
125

Eur. J. Soc. Psychol. 42, 120–133 (2012)


126

Table 1. (Continued )
Non-standard
Authors & country Characteristics rated d Significance level Context Dimension accent Standard accent

Levin et al., 1994 Pay attention 0.84 0.00 No context Solidarity


Jairo N. Fuertes et al.

Levin et al., 1994 Source of advice 0.51 0.01 No context Solidarity


Levin et al., 1994 Speech formal 1.02 0.00 No context Status
Levin et al., 1994 Tried to impress 0.57 0.01 No context Status

Copyright © 2011 John Wiley & Sons, Ltd.


Levin et al., 1994 Words fancy 0.53 0.01 No context Status
Light et al., 1978 USA Appearance: pretty–ugly 1.39 0.00 Education Status “Black Southern” “Standard American Accent”
Light et al., 1978 Socioeconomic status: rich–poor 0.87 0.00 Education Status
Light et al., 1978 Intelligence: dumb–smart 0.42 0.05 Education Status
Light et al., 1978 Personality: nice–mean 0.39 0.07 Education Solidarity
Lindemann, 2003 USA Solidarity 0.00 0.99 No context Solidarity “Korean” “North Midwestern US English”
Lindemann, 2003 Status 0.95 0.00 No context Status
Piche et al., 1977 USA Scholastic achievement 0.41 0.10 Education Status “Black non-standard “White Standard English”
English”
Powesland & Giles, 1975 UK Sincerity 0.67 0.00 Other Solidarity “Bristol” “RP”
Powesland & Giles, 1975 Social class 3.67 0.00 Other Status
Tsalikis et al., 1991 USA Buy from him 1.74 0.00 Sales Solidarity “Greek-accented English” “American”
Tsalikis et al., 1991 Cheerful 1.81 0.00 Sales Dynamism
Tsalikis et al., 1991 Clear 2.26 0.00 Sales Status
Tsalikis et al., 1991 Competent 1.66 0.00 Sales Status
Tsalikis et al., 1991 Convincing 1.79 0.00 Sales Dynamism
Tsalikis et al., 1991 Credible 1.55 0.00 Sales Solidarity
Tsalikis et al., 1991 Dependable 1.39 0.00 Sales Solidarity
Tsalikis et al., 1991 Effectiveness 1.77 0.00 Sales Dynamism
Tsalikis et al., 1991 Friendly 1.33 0.00 Sales Dynamism
Tsalikis et al., 1991 Helpful 1.04 0.00 Sales Solidarity
Tsalikis et al., 1991 Honesty 1.22 0.00 Sales Solidarity
Tsalikis et al., 1991 Humble 0.50 0.00 Sales Solidarity
Tsalikis et al., 1991 Image to product 1.99 0.00 Sales Status
Tsalikis et al., 1991 Image to store 2.16 0.00 Sales Status
Tsalikis et al., 1991 Intelligent 1.47 0.00 Sales Status
Tsalikis et al., 1991 Knowledgeable 1.50 0.00 Sales Status
Tsalikis et al., 1991 Professional 1.67 0.00 Sales Status
Tsalikis et al., 1991 Recommend 1.77 0.00 Sales Solidarity
Tsalikis et al., 1991 Self-confident 1.73 0.00 Sales Dynamism
Tsalikis et al., 1991 Status 1.53 0.00 Sales Status

Eur. J. Soc. Psychol. 42, 120–133 (2012)


Effects of speakers’ accents 127

been measured in numerous ways across the extant body of Hypothesis 2: We hypothesized that standard-accented speak-
literature. In addition, there is no one clearly defined method ers would be rated higher on the dimensions of status, solidar-
of distinguishing these three concepts psychometrically. As a ity, and dynamism than non–standard-accented speakers. We
result, we decided to include effect sizes of all measures of obtained a mean effect size of d = 0.99, n = 50, p < 0.001 (95%
these concepts in the meta-analysis. Each effect size that was CI = 0.83–1.16) for status; a mean effect size of d = 0.52,
calculated for each of these concepts in a given study was used n = 48, p < 0.001 (95% CI = 0.33–0.70) for solidarity; and a
to obtain a mean effect size that represented the outcomes of a mean effect size of d = 0.86, n = 18, p < 0.001 (95% CI = 0.57–
given study. Hence, each study contributed one and only one 1.16) for dynamism. These findings also support our hypothesis:
effect size to the meta-analysis, thus, enabling us to maintain standard-accented speakers are accorded higher ratings of status,
independence of effect across studies. solidarity, and dynamism than non–standard-accented speakers.
We used the software “Comprehensive Meta-analysis—
Version 2” (Biostat, 2000) to calculate effect sizes and to Hypothesis 3: We hypothesized that the RP accent would be
calculate our meta-analysis statistical tests. We employed a rated higher than the American accent when compared against
random effects model because the data from the primary non-standard accents. This hypothesis was not supported, and
studies that we retrieved could not be assumed to come from the reverse was found to be the case. Our results show that the
the same population because studies were conducted in differ- American accent has a strong effect of d = 0.94, p < 0.001
ent geographical locations and accents vary enough across (95% CI = 0.79–1.09), whereas the RP accent had a more
study sites (Borenstein, Hedges, Higgins, & Rothstein, 2009; moderate effect of d = 0.56, p < 0.001 (95% CI = 0.37–0.76).
Hunter & Schmidt, 1990).

RESULTS Exploratory Research Questions

We explored if there would be differences in ratings of standard


Descriptive Statistics versus non-standard accents in the distinct experimental settings
examined by researchers and differences in ratings in the vari-
In all, 153 studies were identified as having discussed or ous countries where the research has taken place. With respect
examined speech effects or accents (this list is available from to settings examined, the results show that standard-accented
the first author). A review of these reports then indicated that speakers are rated significantly higher than non–standard-accented
approximately 70 articles were research studies in which data speakers in education, employment, and sales settings. In
were analyzed and included in the publication, therefore poten- counseling, the mean effect size was in the direction of
tially representing sources for inclusion in the meta-analysis. favoring the standard-accent but did not reach statistical signifi-
A subsequent evaluation of the 70 articles indicated that only cance. The mean effect sizes for the various settings were as
20 articles provided the statistics necessary to calculate effect follows: counseling d = 0.12, n = 3, p < 0.17 (95% CI = 0.05
sizes. From the 20 independent studies included for synthesis to 0.29); education d = 0.40, n = 20, p < 0.001 (95% CI = 0.25–
in this meta-analysis, we were able to compute 116 indepen- 0.54); employment d = 0.53, n = 22, p < 0.001 (95% CI = 0.32–
dent effect sizes. These studies included a total of 1738 partici- 0.74); and sales d = 1.59, n = 20, p < 0.001 (95% CI = 1.41–1.77).
pants: 666 men and 802 women. Five of the 20 studies did not There have been studies that have examined ratings of
report the gender of the participants. The 20 studies included English speakers in various countries and compared evaluations
in the meta-analysis are identified in the reference list with of standard and non-standard accents (e.g., business executives
an asterisk. All used the matched-guise technique or slight fluent in English were compared on the basis of standard versus
variations from the original (Lambert, 1967) (*). non-standard accents). The results show that standard speakers
Table 1 presents the 116 individual effect sizes for the 20 are rated significantly higher than non–standard-accented
studies examined. The table is organized in terms of the speakers in the United States, Argentina, Turkey, and the
author(s) last name and year of publication, the characteristics United Kingdom. The results for studies conducted in Australia
examined by the researcher(s) in the study, effect size d statistic were in the direction of favoring standard-accented speakers
and level of significance, context as defined/described by the but did not reach statistical significance. The mean effect
researcher(s), dimensions of status, solidarity, or dynamism as sizes are as follows: United States d = 1.04, n = 66, p < 0.001
determined by the current team of authors, and the type of (95% CI = 0.89–1.19); Argentina d = 1.34, n = 1, p < 0.001
accents that were examined. (95% CI = 1.03–1.65); Australia d = 0.32, n = 4, p < 0.031
(95% CI = 0.31 to 0.96); Turkey d = 0.44, n = 13, p < 0.001
Hypotheses (95% CI = 0.6–0.82); and the United Kingdom d = 0.56, n = 32,
p < 0.001 (95% CI = 0.35–0.77).
Hypothesis 1: Speakers with standard accents are rated
higher compared with speakers with non-standard accents. Instances When Non-Standard Accents were Preferred
We found a mean effect size of d = 0.82, n = 116, p < 0.01 over Standard Accents
(95% confidence interval [CI] = 0.70–0.95). This finding
supports our hypothesis and indicates that speakers with As the data above show, in almost all instances, standard-accented
standard accents are rated more positively than speakers with speakers are preferred over non–standard-accented speakers.
non-standard accents. But, as noted in the introduction, there are instances in which

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
128 Jairo N. Fuertes et al.

non-standard accents are preferred by listeners over standard acts as an immediate marker for the listeners on which to
accents. There were seven individual negative effect sizes within downgrade the non–standard-accented speaker. Although
specific studies, where the non-standard accent was preferred to legislation has been drafted to prohibit discrimination on the
the standard accent on specific outcome variables. These effect basis of gender or race, it may be that use of language is as
sizes can be seen in Table 1. In two studies conducted in the powerful a trigger in instigating negative evaluations of
United Kingdom comparing Bristol and South Welsh accents speakers, which may compound negative evaluations already
with RP accents, five of these seven negative effect sizes were experienced by minority group members, those members of
on the dimension of solidarity. All of these effect sizes except society living in poverty, and citizens who speak with a
one (i.e., quiet–talkative) were statistically significant. The data non-standard accent such as immigrants. The effect on the
that contributed to these results are scant, but we believed that third dimension examined, solidarity, was also significant,
they are worth noting because they contrast with the rest of though moderate (d = 0.52), and shows that speaking with a
the data and they suggest that there might be certain cultural standard accent very much leads to higher ratings on factors
factors that might lead to contrary results. Specifically in these such as attractiveness, benevolence, and trustworthiness. This
two studies, the respondents hearing the Bristol and South finding is consistent with our hypothesis but inconsistent with
Welsh accents may have upgraded these speakers on the basis some of the theoretical and empirical literature that has posited
of perceiving them as in-group versus out-group members. and found an effect favoring non–standard-accented speakers
on the dimension of solidarity.
The study also examined the magnitude of the effects of
Fail-Safe N accents by context examined and the country in which the
study took place. With respect to context, the average result
The Fail-Safe N determines how many studies exist in was strong across settings, with the standard accent being rated
researchers’ file drawers and were not found by us that would more favorably than non-standard accents. But the results also
be needed to undermine or contradict the results of this study show some variability of the effects, with the more formal and
(Rosenthal, 1991). It is important to make sure that the results high stakes settings, such as employment and sales situations,
were based on a thorough search of the literature and the generating the strongest effects favoring standard accents.
Fail-Safe N enables that possibility. The Fail-Safe N for this With respect to country, the average result was also strong
study is 10 547, which means that more than 10 000 studies across countries, with the standard accent being rated more
would have to be found with nil results to cause the current favorably than non-standard accents. However, there was also
study’s results to become non-significant statistically. considerable variability from country to country. Studies
conducted in the US generated the highest average effect size
(the study in Argentina contained only one effect size, whereas
the studies in the US generated 66 effect sizes), which was
DISCUSSION
double the effect size for studies conducted in the UK.
Although researchers typically associate RP as the accent
The current study set out to examine the statistical effect of generating the highest levels of prestige around the globe,
speakers’ accents on social evaluations. In many ways, and the current results show that at least in the US, the American
as highlighted previously, this enterprise was triggered by accent is much stronger in generating positive effects when
the lack of focal attention afforded to this vocal feature in compared with non-standard accents. There is evidence that
prominent treatises in social psychology and communication shows that the American Network accent has surpassed the
science. The first hypothesis examined the overall magnitude RP accent as the standard in other English-speaking countries,
of the effect of speakers’ accents on interpersonal judgments such as Australia (Giles & Billings, 2004).
made by the listener. The results revealed an effect size of There are applied and theoretical implications of this
d = 0.82 indicating that this effect is very strong. Across rating meta-analysis. That people have, or rather are attributed as
dimensions, speakers who use a standard accent are rated more having, a so-called “accent” has been a social issue and con-
positively than those using a non-standard accent, almost a full cern in everyday parlance, popular culture, and in many pro-
standard deviation higher. This effect may have considerable fessional contexts for a very long time. A recent Web search
consequences for those speakers being evaluated. For the on “reactions to people’s accents,” for instance, triggered
standard speaker, it represents a huge advantage, and for the well over four million hits. As Giles and Marlow (2011) have
non-standard speaker, it represents nothing less than a consid- argued, perhaps it is time to proactively utilize social net-
erable handicap. The reader should consider the fact that working and new communication resources to dispel and
evaluations have been shown to be shaped by single words, offer remedial options for what are often speech stereotypes
such as the speaker saying “hello.” The implications for with cumulatively disturbing repercussions, thereby alleviat-
non-standard speakers are considerable, as they are much less ing some of the social harm inflicting on those enduring lin-
likely to make positive impressions, even first impressions, guistic insecurities. Turning to models of social interaction,
when compared with standard speakers. the salience of accent has relevance for models such as com-
The effect on evaluations associated with status, such as munication accommodation theory (CAT; see, e.g., Gallois,
how educated, intelligent, or successful the speaker is, was Ogay, & Giles, 2005; Giles, Willemyns, Gallois, & Anderson,
even higher at d = 0.99. Furthermore, the effect of ratings 2007) that was originally formulated for explaining people’s
associated with the speaker’s dynamism was also strong at facility for changing their accents from situation to situation
d = 0.86. On two important dimensions, accented speech (Giles, 1973b).

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Effects of speakers’ accents 129

This theory examines the motivations for and consequences Another limitation pertains to the context in which ratings
arising from individuals, couples, and groups modifying are observed. On the one hand, many of the studies are not
their speech toward and away from each other. Besides the naturalistic but experimental or analogue in nature, and so a
ubiquity of underaccommodative and overaccommodative question remains as to whether these evaluations would be
strategies in conversation, CAT has devoted most of its em- replicated in real-life scenarios in which communication takes
pirical and theoretical attentions toward understanding how place. On the other hand, some of the contexts described in the
and why speakers converge and diverge away from each published studies are unclear or not at all specified. Because
other. The current meta-analysis underscores the social much evidence exists about behavior being context dependent,
rewards (e.g., increased status, education, confidence, com- it behooves researchers to pay closer to attention to context
petence) for people upwardly converging toward those with and to specify in the studies how the context was created or
more prestige accents and upwardly diverging away from selected. This is important to specify in studying the possible
non–standard-accented speakers. In fact, it would be interesting beneficial effects of non-standard accents on solidarity ratings,
to determine whether upward convergence is a more prevalent meaning settings where speaking with a non-standard accent
communicative tactic than downward convergence toward low may trump standard accent (e.g., in settings that are casual or
prestige speakers. Although an increase in social attractiveness, recreational, such as in sports settings or a street playground).
social identification, and sincerity might accompany the latter There is also a lack of consistency in the terms used to
in certain kinds of solidarity-stressing situations (see Giles describe personal characteristics that are assessed. Table 1 lists
& Ryan, 1982), it could also reap social costs in terms of an array of dimensions, which we classified given our under-
authenticity and lowered competence. standing of the dimensions of status, solidarity, and dynamism
The small number of negative effect sizes uncovered by our as empirically derived from the work by Zahn and Hopper
results is surprising. The literature in this field has suggested (1985). Future publications would benefit from reporting the
that non-standard accents might generate higher effect sizes characteristics of interest (e.g., friendliness or selfishness)
than standard accents, particularly when the judgments are under these three dimensions along with a rationale, to provide
made by similarly accented listeners. We expected a higher a common language among researchers. Alternatively,
number of negative effect sizes, particularly on the dimension researchers may seek to study evaluations based on personality
of solidarity. However, the overwhelming number of effect types that have been researched extensively in psychology,
sizes was positive, strong, and consistent across dimension, such as the 16 personality dimensions presented by Cattell
settings, and countries, in favor of standard-accented speakers. (e.g., warmth, dominance, sensitivity; Conn & Rieke, 1994).
This finding may have had to do with the type of subject/ There are also possibilities for grounding studies on language
participant typically included in the research studies reviewed: variability and the effects of speakers’ accents on principles of
educated college students who were probably standard- social psychology, such as studying the effects of language on
accented in their use of English and who thus rated as higher attitude convergence, on discrepancy or conflict resolution, on
on solidarity the standard-accented speakers who sounded to the construction of a shared social reality in relationships, and
them as similar and trustworthy. on the development of trust among couples or groups.
Many of the studies of interest reviewed for the meta-
analysis could not be included in the present study because
Limitations and Future Research of the inadequate reporting of statistics. Much of the data
included in the articles reviewed were presented as aggregated
We acknowledge some limitations in the current study. First, data, and often the current research team was unable to
the studies that were meta-analyzed did not provide evidence decipher statistics for specific groups or demographic variables
as to the accent of the evaluators. A research question that of interest. A more thorough reporting of statistics is needed to
remains and needs to be investigated is whether the effect of accurately evaluate differences between the overall evalua-
non-standard speech is as strong when evaluators use the same tions of accent conditions, as well as differences based on
type of English as the speakers. This question could not be the demographic differences of the evaluators. Researchers
examined on the basis of the current studies available in the are also encouraged to provide the statistical data needed to
literature. Also, given the realities of racism across the world, calculate effect sizes including means, standard deviations,
a potential confound that cannot be ascertained in the current numbers of participants by group, significance levels, T or Z
study is whether the downgrade of non-standard speech is scores, correlation coefficients, and R or F values.
because of the speech itself or the extent to which it is also a We also note the paucity of available empirical research
function of the race or ethnicity of the speaker or the listener that has been conducted on this topic in the last 10 years.
(see Ray, 2009). Another limitation pertains to sex differences Interpersonal evaluations based on accents continue to be an
in the ratings and evaluations of accented speech. Although important area of study and should continue to be studied in
the sex of the participants was noted in 15 of the studies, a variety of contexts. That being said, in different national
ratings were not presented by sex. So, it is impossible to contexts (e.g., Denmark, The Netherlands), subjective notions
discern the role of this variable in interpersonal evaluations of what is the “standard” (or even optional standards) is
based on accent (see Giles, Smith, Ford, Condor, & Thakerar, continually evolving (Grondelaers, van Hout, & Steegs,
1980). Future research in the area of interpersonal evaluations 2010; Kristiansen, 2001). Clearly, there is a need for more
based on speakers’ accents would benefit from research stud- research to explicitly document, objectively and subjectively,
ies that include detailed demographic data such as their sex, the nature of changes in socioeconomic circumstances within
ethnicity, race, education level, and accent. any given society, as well as the changing dynamic relations

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
130 Jairo N. Fuertes et al.

between ethnic and social groups within it (e.g., Bishop, Craighead, W. E., & Nemeroff, C. (Eds.). (2004). The concise Corsini
encyclopedia of psychology and behavioral science (3rd ed.). Chichester:
Coupland, & Garrett, 2003; Genesee & Holobrow, 1989) and Wiley & Sons.
relate these to accent evaluations. Cuddy, A. J. C., Fiske, S. T., & Glick, P. (2008). Competence and warmth as
In closing, evaluations based on accent appear to have a universal trait dimensions of interpersonal and intergroup perceptions: The
significant impact on individuals who do not speak with stereotype content model and the BIAS map. In M. P. Zanna (Ed.),
Advances in experimental psychology (Vol. 40, pp. 61–149). New York,
standard accents and are likely to lead to discrimination NY: Academic Press.
(Lippi-Green, 1997; Ryan & Giles, 1982; Ryan et al., 1982) Delamater, J. (Ed.). (2003). Handbook of social psychology. New York, NY:
and possibly other severe social consequences. For the Kluwer Academic/Plenum.
Derwing, T. M. (2003). What do ESL students say about their accents? The
accented speaker, the stigma attached to their speech can Canadian Modern Language Review, 59, 547–566.
create difficulties in communicating with others and lead to Donsbach, W. (Ed.). (2008). International encyclopedia of communication.
the increased perception that the listener is biased or will Oxford: Blackwell.
Edwards, J. R., & Jacobsen, M. (1987). Standard and regional standard
respond negatively (Gluszek & Dovidio, 2010a). These speech: Distinctions and similarities. Language in Society, 16, 369-80.
meta-perceptions related to discrimination increases when the Fiske, S. T., Cuddy, A. J. C., Glick, P., & Xu, J. (2002). A model of (often
speaker can be visibly categorized as a minority (Derwing, mixed) stereotype content: Competence and warmth respectively follow
from perceived status and competition. Journal of Personality and Social
2003; Lippi-Green, 1997). In addition, as noted by Gluszek Psychology, 82, 878–902.
and Dovidio (2010b), the stigma associated with accented *Fuertes, J. N., & Gelso, C. J. (2000). Hispanic counselors’ race and accent
speech can become intertwined with social identity and result and Euro Americans’ Universal Diverse Orientation: A study of initial per-
ceptions. Cultural Diversity and Ethnic Minority Psychology, 6, 211–219.
in a decreased sense of belongingness. Thus, carefully *Gallois, C., Callan, V. J., & Johnstone, M. J. (1984). Personality judgments
constructed and presented research is still needed in this of Australian aborigine and white speakers: Ethnicity, sex and context.
extremely important area. There are resources on the Internet Journal of Language and Social Psychology, 3, 39–57.
Gallois, C., Ogay, T., & Giles, H. (2005). Communication accommodation
that can be used to initiate the dialogue regarding evalua- theory. In W. Gundykunst (Ed.), Theorizing about intercultural communi-
tions/judgments/perceptions and accents in employment, cation (pp. 121–148). Thousand Oaks, CA: Sage.
professional settings, education, and social interactions. One Garrett, P., Coupland, N., & Williams, A. (2003). Investigating language
attitudes: Social meanings of dialect, ethnicity and performance. Cardiff:
such tool is the accent database that can be found at http:// University of Wales Press.
accent.gmu.edu. We hope that the size of the effects uncovered Genesee, F., & Holobow, N. E. (1989). Change and stability in intergroup
in this meta-analysis will stimulate further thinking, healthy perceptions. Journal of Language and Social Psychology, 8, 17–38.
criticisms and debate, and further research in this area. Giles, H. (1970). Evaluative reactions to accents. Educational Review, 22,
211–227.
Giles, H. (1973a). Communicative effectiveness as a function of accented
speech. Speech Monographs, 40, 330–331.
Giles, H. (1973b). Accent mobility: A model and some data. Anthropological
REFERENCES Linguistics, 15, 87–105.
Giles, H., & Billings, A. (2004). Language attitudes. In A. Davies, & E. Elder
(Eds.), Handbook of applied linguistics (pp. 187–209). Oxford: Blackwell.
*Arthur, B., Farrar, D., & Bradford, G. (1974). Evaluation reactions of college Giles H., & Bourhis, R. Y. (1973). Dialect perception revisited. The Quarterly
students to dialect differences in the English of Mexican-Americans. Journal of Speech, 59, 337–342.
Language and Speech, 17, 255–270. Giles, H., & Edwards, J. R. (2010). Attitudes to language: Past, present, and
Berger, C. R., & Bradac, J. J. (1982). Language and social knowledge. London: future. In K. Malmkjaer (Ed.), The Routledge linguistics encyclopedia
Academic Press. (3rd ed., pp. 35–40). London: Routledge.
(2000). Comprehensive meta-analysis—version 2. Englewood, NJ: Biostat. Giles, H., & Marlow, M. (2011). Theorizing language attitudes: Past frame-
Bishop, H., Coupland, N., & Garrett, P. (2003). Conceptual accent evaluation: works, an integrative model, and new directions. In C. Salmon (Ed.),
30 years of accent prejudice in the UK. Acta Linguistic Hefniensia, 37, Communication yearbook 35 (pp. 161–197). Thousand Oaks, CA: Sage.
131–154. Giles, H., & Niedzielski, N. (1998). German sounds awful, but Italian is
Blair, L. M., & Connor, H. S. (1978). Black and rural accents found to lessen beautiful. In L. Bauer, & P. Trudgill (Eds.), Language myths (pp. 85–93).
job opportunities. Monthly Labor Review, 101, 35–36. Harmondsworth: Penguin.
Borenstein, M., Hedges, L. V., Higgins, J. P. T., & Rothstein, H. (2009). Giles, H., & Powesland, P. F. (1975). Speech style and social evaluation. New
Introduction to meta-analysis. New York, NY: Wiley. York, NY: Academic Press.
Bourhis, R. Y., Giles, H., & Tajfel, H. (1973). Language as a determinant of Giles, H., & Ryan, E. B. (1982). Prolegomena for developing a social psycho-
Welsh Identity. European Journal of Social Psychology, 3, 447–460. logical theory of language attitudes. In E. B. Ryan, & H. Giles (Eds.),
Bradac, J. J. (1990). Language attitudes and impression formation. In H. Giles, Attitudes towards language variation (pp. 208–223). London: Edward
& W. P. Robinson (Eds.), Handbook of language and social psychology Arnold.
(pp. 387–412). Chichester, UK: Wiley. Giles, H., & Sassoon, C. (1983). The effect of speaker’s accent/ social class
*Brown, B. L., Giles, H., & Thakerar, J. N. (1985). Speaker evaluations as a background and message style on British listeners’ social judgments.
function of speech rate, accent and context. Language and Communication, Language and Communication, 3, 302–313.
5,207–220. *Giles, H., Baker, S., & Fielding, G. (1975). Communication length as a
Burleson, B., & Greene, J. (Eds.). (2003). Handbook of communication and behavioral index of accent prejudice. International Journal of the Sociology
social interaction skills. Mahwah, NJ: Erlbaum. of Language, 6, 73–81.
*Callan, V. J., Gallois, C., & Forbes, P. A. (1983). Evaluative reactions to Giles, H., Reid, S. A., & Harwood, J. (Eds.). (2010). The dynamics of
accented English. Journal of Cross-Cultural Psychology, 14, 407–426. intergroup communication. New York, NY: Peter Lang.
Cargile, A. C., & Bradac, J. J. (2001). Attitudes toward language: A review of Giles, H., Smith, P. M., Ford, B., Condor, S., & Thakerar, J. N. (1980). Speech
research and a general process model. In W. B. Gudykunst (Ed.), Commu- styles and the fluctuating saliency of sex. Language Sciences, 2, 260–282.
nication yearbook, 25, 347–382. Abingdon, Oxon, UK: Routledge. Giles, H., Willemyns, M., Gallois, C., & Anderson, M. C. (2007). Accommo-
*Cargile, A. C., & Giles, H. (1998). Language attitudes toward varieties of dating a new frontier: The context of law enforcement. In K. Fiedler (Ed.),
English: An American-Japanese context. Journal of Applied Communica- Social communication (pp. 129–162). New York, NY: Psychology Press.
tion Research, 26, 338–356. *Giles, H., Williams, A., Mackie, D. M., & Rosselli, F. (1995). Reactions to
Cohen, J. (1988). Set correlation and contingency tables. Applied Psychological Anglo- and Hispanic-American-accented speakers: Affect, identity, persua-
Measurement, 12, 425–434. sion, and the English-only controversy. Language and Communication, 15,
Conn, S. R., & Rieke, M. L. (1994). The 16PF Fifth Edition technical manual. 107–120.
Champaign, IL: Institute for Personality and Ability Testing, Inc. *Giles, H., Wilson, P., & Conway, A. (1981). Accent and lexical diversity as
Cooper, H. M. (1989). Integrating research: A guide for literature reviews detriments of impression formation and perceived employment suitability.
(2nd ed.). Thousand Oaks, CA: Sage. Language Sciences, 3, 91–103.

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Effects of speakers’ accents 131

*Gill, M. M. (1994). Accent and stereotypes: Their effect on perceptions of Marlow, M., & Giles, H. (2008). “Who you tink You, talkin propah?” Hawaiian
teachers and lecture comprehension. Journal of Applied Communication Pidgin demarginalized. Journal of Multicultural Discourses, 2, 53–69.
Research, 22, 348–361. Mulac, A. (1975). Evaluation of the speech dialect attitudinal scale. Speech
*Gill, M. M., & Badzinski, D. M. (1992). The impact of accent and status on Monographs, 42, 184–189.
information recall and perception formation. Communication Reports, 5, Mulac, A., Hanley, T. D., & Prigge, D. Y. (1974). Effects of phonological
99–106. speech foreignness upon three dimensions of attitude of selected American
Gluszek, A., & Dovidio, J. F. (2010a). Speaking with a nonnative accent: listeners. The Quarterly Journal of Speech, 60, 411–420.
Perceptions of bias, communication difficulties, and belonging to the Nazzi, T., Jusczyk, P., & Johnson, E. (2000a). Language discrimination by
United States. Journal of Language and Social Psychology, 29, 224–234. English-leaning 5-month-olds: Effects of rhythm and familiarity. Journal
Gluszek, A., & Dovidio, J. F. (2010b). A social psychological perspective on of Memory and Language, 43, 1–19.
the stigma of non-native accents in communication. Personality and Social Ng, S. H., & Bradac, J. J. (1993). Power in language. Thousand Oaks, CA: Sage.
Psychology Review, 14, 214–237. *Piche, G. L., Michlin, M., Rubin, D., & Sullivan, A. (1977). Effects of
Gluszek, A., Newheiser, A.-K., & Dovidio, J. F. (2011). Social psychological dialect-ethnicity, social class and quality of written compositions on tea-
orientations and accent strength. Journal of Language and Social Psychol- chers’ subjective evaluations of children. Communication Monographs,
ogy, 30, 28–45. 44, 60–72.
*Green, D. J., Scott, J. C., & Rosewarne, D. D. (2002). Perceptions about three *Powesland, P., & Giles, H. (1975). Persuasiveness and accent-message in-
indigenous English accents from prospective and practicing Argentine compatibility. Human Relations, 28, 85–93.
providers of business-related language services. Delta Pi Epsilon National Rakić, T., Steffens, M. C., & Mummendey, A. (2011). Blinded by the accent!
Conference, 65–70. The minor role of looks in ethnic categorization. Journal of Personality and
Grondelaers, S., van Hout, R., & Steegs, M. (2010). Evaluating regional accent Social Psychology, 100, 16–29.
variation in Standard Dutch. Journal of Language and Social Psychology, Ray, G. (2009). Language and interracial communication in the United
29, 101–116. States: Speaking in Black and White. New York, NY: Peter Lang.
Gudykunst, W. B., & Ting-Toomey, S. (1990). Ethnic identity, language Robinson, W. P. (1972). Language and social behavior. Harmonsworth:
and communication breakdowns. In H. Giles, & P. Robinson (Eds.), Penguin.
Handbook of language and social psychology (pp. 307–327). London: Robinson, W. P. (2003). Language in social worlds. Oxford: Blackwell.
Wiley. Rosenthal, R . (1991). Effect sizes: Pearson’s correlation, its display via the
*Hensley, A. (1972). Black high school students’ reactions to black speakers BESD, and alternative indices. American Psychologist, 46, 1086–1087.
of standard and black English. Language Learning, 22, 253–259. Rosenthal, R., & DiMatteo, M. R. (2001). Meta-analysis: Recent develop-
Hogg, M. A., & Cooper, J. (Eds.). (2003). The SAGE handbook of social psy- ments in quantitative methods for literature reviews. Annual Review of
chology. Thousand Oaks, CA: Sage. Psychology, 52, 59–82.
Holtgraves, T. M. (2002). Language as social action: Social psychology and Ryan, E. B. (1979). Why no low-prestige language varieties persist? In H.
language use. Mahwah, NJ: Erlbaum. Giles, & R. N. St. Clair (Eds.), Language and social psychology (pp.
Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis: Correcting 145–157). Oxford: Blackwell.
error and bias in research findings. Newbury Park, CA: Sage. Ryan, E. B., & Giles, H. (Eds.). (1982). Attitude towards language variation:
*Kamisli, S., & Dugan, S. (1997). Nonnative speakers’ speech perception of Social and applied contexts. London: Edward Arnold.
native speakers. Reports–Research, ERIC Clearinghouse on Language Ryan, E. B., & Sebastian, R. J. (1980). The effects of speech style and social
and Linguistics, Washington DC. class background on social judgments of speakers. The British Journal of
Kinzler, K. D., Dupoux, E., & Spelke, E. S. (2007). The native language of Social and Clinical Psychology, 19, 229–233.
social cognition. The Proceedings of the National Academy of the United Ryan, E. B., Carranza, M. A., & Moffie, R. W. (1977). Reactions towards
States of America, 104, 12577–12580. varying degrees of accentedness in the speech of Spanish-English. Lan-
Kinzler, K. D., Shutts, K., DeJesus, J., & Spelke, E. S. (2009). Accent guage and Speech, 20, 267–273.
trumps race in guiding children’s social preferences. Social Cognition, Ryan, E. B., Giles, H., & Sebastian, R. J. (1982). An integrative perspective
27, 623–624. for the study of attitudes toward language variation. In E. B. Ryan, & H.
Ko, S. J., Judd, C. M., & Stapel, D. A. (2009). Stereotyping based on voice in Giles (Eds.), Attitudes toward language variation (pp. 1–19). London:
the presence of individuating information: Vocal femininity affects Edward Arnold.
perceived competence but not warmth. Personality and Social Psychology Sanders, J., & Fitch, K. (Eds.). (2004). Handbook of language and social in-
Bulletin, 35, 198–211. teraction. Mahwah, NJ: Sage.
Krauss, R. M., Freyberg, R., & Morsella, E. (2002). Inferring speakers’ phys- Scherer, K. R. (1979). Personality markers in speech. In K. R. Scherer & H.
ical attributes from their voices. Journal of Experimental Social Psychol- Giles (Eds.), Social markers in speech (pp. 147–210). Cambridge:
ogy, 38, 618–625. Cambridge University Press.
Kristiansen, T. (2001). Two standards: One for the media and one for the Scherer, K. R., & Giles, H. (Eds.). (1979). Social markers in speech. Cambridge:
school. Language Awareness, 10, 9–24. Cambridge University Press.
Lambert, W. E. (1967). A social psychology of bilingualism. Journal of Social Stewart, M. A., & Ryan, E. B. (1985). Accent and social class effects on status
Issues, 23, 91–109. and solidarity evaluations. Personality and Social Psychology Bulletin, 11,
*Levin, H., Giles, H., & Garrett, P. (1994). The effects of lexical formality and 98–105.
accent of trait attributions. Language and Communication, 14, 265–274. Tajfel, H., & Turner, J. C. (1986). An integrative theory of intergroup conflict.
*Light, R. L., Richaard, D. P., & Bell, P. (1978). Development of children’s In S. Worchel, & W. G. Austin (Eds.), Psychology of intergroup relations
attitudes toward speakers of standard and non-standard English. Child (pp. 2–24). Chicago, IL: Nelson-Hall.
Study Journal, 8, 253–265. Trudgill, P. (1974). Sociolinguistics. Harmondsworth: Penguin.
*Lindemann, S. (2003). Koreans, Chinese or Indians? Attitudes and ideologies *Tsalikis, J., DeShields, O. W., & LaTour, M. S. (1991). The role of accent on
about non-native English speakers in the United States. Journal of Sociolin- the credibility and effectiveness of the salesperson. Journal of Personal
guistics, 7, 348–364. Selling and Sales Management, 6, 31–41.
Lippi-Green, R. (1997). English with an accent: Language, ideology, and Ventola, E., & Antos, G. (Eds.). (2008). Handbook of interpersonal communi-
discrimination in the United States. New York, NY: Routledge. cation. Berlin: Mouton de Gruyter.
Lipsey, M. W., & Wilson, D. W. (2001). Practical meta-analysis. Thousand Zahn, C. J., & Hopper, R. (1985). Measuring language attitudes: The speech
Oaks, CA: Sage. evaluation instrument. Journal of Language and Social Psychology, 4,
Manusov, V., & Patterson, M. L. (Eds.). (2006). Handbook of nonverbal 113–123.
behavior. Thousand Oaks, CA: Sage. *Study included in the meta-analysis.

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
132 Jairo N. Fuertes et al.

APPENDIX A SAMPLE ITEMS FROM THE CODING FORM

Bibliographic reference:
______________________________________________________________________________________
______________________________________________________________________________________

____ 1. Study ID #
____ 2. Publication year

Information specific to study

____ 3. Context
1 Employment 4 No situational contexts—general study
2 Educational settings 5 Other: __________________________
3 Professional

____ 4. Location of study


1 United States 5 Australia
2 England 6 New Zealand
3 Scotland 7 Singapore
4 Ireland 8. Other: __________________________

____ 5. Standard accent in location of study


1 Received Pronunciation (RP) 4 Standard New Zealand Accent
2 Standard American Accent 5 Standard Accent not specified
3 Standard Australian Accent 6 Other ___________________________

____ 6. Non-standard accent in location of study


1 Cockney or other regional accents in England 5 New Zealand regional accent
2 Spanish accented 6 Asian accent
3 Black American accent 7 Non-standard accent not specified
4 Australian regional accent 8 Other ___________________________

Sample descriptors

____ 7. N of sample
____8. Mean age of sample
____ 8. Gender of sample
1 Female________ 2 Male________ 3 Not given_______

____ 9. Ethnicity of sample


1 Australian 4 Irish
2 Japanese 5 Vietnamese
3 Greek 6 Other______________

____ 10. Accent of sample


1 Standard 2 Non-standard 3 Both

____ 11. Socioeconomic Status of sample


1 Wealthy 4 Lower Middle class
2 Upper-middle class 5 Poor
3 Middle class 6 Not given

Research design descriptors


____ 12. Type of study
1 Experiment 2 Quasi-experimental 3 Correlational
____ 13. Study design
1 Within subjects 2 Between subjects 3 Mixed design

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Effects of speakers’ accents 133

____ 14. Design of analysis


1 MANOVA factorial 2 ANCOVA factorial 3 ANOVA factorial
____ 15. What medium was used
1 Audio 2 Video 3 Other
____ 16. What accent guise was used
1 Matched guise 2 Verbal guise 3 Other

Nature of dependent variable descriptors (list)

Copyright © 2011 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 42, 120–133 (2012)
Copyright of European Journal of Social Psychology is the property of John Wiley & Sons, Inc. and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express
written permission. However, users may print, download, or email articles for individual use.

View publication stats

You might also like