The effects of non-native accented English on employability.

(1)

The effects of non-native accented English on

employability

Bachelor thesis

Mylène van de Wouw

International Business Communication Radboud University

Supervisor 1: Frank van Meurs Supervisor 2: Evelyne Fruit

(2)

Abstract

Due to globalization and migrations, multinational companies have often implemented English as a common corporate language (CCL) for its employees from different language backgrounds to communicate more efficiently. The current study examined the effect of non-native English accents on employability for a high-communication job. In an experiment, 211 Dutch and German non-native speakers of English, evaluated a Spanish English, Arabic English and British accent in terms of status, solidarity,

dynamism, employability and comprehensibility. Moreover, the current study examined whether stereotype images might be a possible explanation for differences in

evaluations on the accent conditions. The results showed that only the Arabic accent was downgraded on all dimensions. However, the Spanish accent was found to be evaluated similarly to the British accent. With the use of the stereotype content model, it was found that Arabs are stereotyped more negatively than Spaniards and English. For both Dutch and German companies with English as a CCL this study indicated that it is important to base employability on actual competences, and not focus on audio-cues only.

(3)

Introduction

Nowadays, organizations are becoming more global every day. In this global village, Piekkari, Vaara, Tienari, and Säntti (2005) discovered that organizations often use English as a Common Corporate Language (CCL), which can be used as a central medium of communication for interaction between employees from linguistically different backgrounds. Accordingly, the aim of a CCL is to facilitate communication between native and non-native speakers of the CCL (Tietze, 2008). Therefore, English as a CCL could be seen as a resource to facilitate communication, because of the increasing number of non-native speakers of English (Crystal, 2003). Due to the increasing globalization in business, many people are living and working in a foreign country. It is evident that people from different countries have different accents when speaking English as the shared language at the same workplace. Therefore, it is crucial to investigate the impact of such accents in the workplace, as it might be as pertinent as other personal factors like gender and skin color, and even evokes discrimination at the workplace (Deprez-Sims & Morris, 2010). Little is known about the effects of non-native accentedness in the workplace, especially when there is interaction in English in which both speaker and listener are non-native speakers of English. It has been noticed that in Germany and The Netherlands, many immigrants are from the Southern Europe and the Arabic region (Hagendoom & Hraba, 1987). Therefore, the current study proposes an experimental design to investigate the influence of a non-native Spanish-English accent, an Arabic-Spanish-English accent and a native British-Spanish-English accent on employability. Moreover, little is known about the effects of these specific accents in a European context (Gallardo del Puerto, García Lecumberri, & Gómez Lacabex, 2015; Grondelaers & van Gent, 2019; Hendriks, van Meurs, & de Groot, 2017). Hence, the current research aims to explore how Dutch and German listeners evaluate two non-native English accents. The results of this study could be of particular interest to Dutch and German multinational companies that hire people from different linguistic

(4)

Theoretical framework

Immigrants as potential employees in a multicultural society

Although a multicultural society is nowadays common, attitudes towards immigrants and refugees are found to be predominantly negative (Esses, Jackson, & Armstrong, 1998). Moreover, job seekers with a non-native accent are often discriminated when applying for a job (Krings & Olivares, 2007).

In the current study, the focus will be on native German and native Dutch as the evaluators of the accents because they are both Western European countries with large numbers of immigrants from various countries. In 2019, 267.738 persons immigrated to The Netherlands, of whom the largest group is from Turkey (Centraal Bureau voor Statistiek, 2019). Among these immigrants, Hagendoom and Hraba (1987) found a series of out-group preferences of the Dutch. After the Dutch, the so-called in-group, the Dutch inhabitants favor other Northern Europeans most, followed by Jews, Southern Europeans like Spaniards, individuals from previous colonial communities like

Surinamese and Indonesians and lastly members of Islamic regions such as the Mideast. Therefore, it can be hypothesized that Spanish-accented English is preferred to Arabic-accented English.

Labor migration in The Netherlands includes groups from Southern Europe, the largest being Spaniards; North Africa, and the Middle East (Hagendoom & Hraba, 1987). Following the Federal Statistical Office of Germany (2019), 13.1% of the total German population is foreign. However, of this foreign population only 1.62% is from Spain, and more than 35% is from an Arabic/Mideast country.

As Gluszek and Dovidio (2010) found that immigration increases the use of English, often with a non-native accent, the current study aims to investigate non-native accents of the foreign population from Spain and Arabic countries in The Netherlands and Germany.

Accents and negative consequences

An accent refers to speaking a second language (L2) differently than the ‘norm’, often associated with a specific group of people, commonly based on phonological or intonation differences across social groups or geographic regions (Lippi-Green, 2012). In other words, the phonology of their native language is kept. Therefore, the speech characteristics of this first language could interfere with the L2, which creates a

(5)

non-native accent. Having a non-non-native accent may evoke positive or negative reactions towards the listener. It has been found that speaking with a non-native accent reveals the person as not being native born and a native speaker of English, and even further, that this person does not master the language completely (Gluszek & Dovidio, 2010). Consequently, speaking with a non-native accent results in negative consequences, which are linked to the social identity theory of Tajfel (2010), considering that when a speaker’s accent is different than the accent of the listener, it creates the feeling that the speaker belongs to an outgroup of this second language and thus leads to social

categorization. For example, this process of social categorization activates stereotyping and prejudice in job interview processes (Biernat & Dovidio, 2000; Gluszek & Dovidio, 2010).

Other negative consequences of speaking with a non-native accent, is displayed in the similarity-attraction bias. This refers to the way that people may like or are more drawn to people with a similar attitude and values than to people who are more

dissimilar (Condon & Crano, 1988). In other words, being dissimilar from a person creates negative outcomes. The study of Deprez-Sims and Morris (2010) found evidence for the bias on evaluation processes in job interviews. The French-English accent was perceived as dissimilar to the recruiter from the US and therefore received an inferior evaluation than the Midwestern-US accent, who was found to be more similar.

As previous studies have shown that speaking with a non-native accents causes several negative consequences, the current study aims to examine the role of

stereotyping. For example, a previous study on accents and stereotypical associations by Hosoda, Nguyen and Stone-Romero (2010), showed that Spanish accents were

associated with warmth and kindness, but were not suitable for high-communication jobs. Moreover, they found that particularly when the accent is related to a minority group, such as Hispanics in the USA, the stereotypical images are stronger and more negative. In addition, the study of Ryan, Carranza and Moffie (1977) found in an

experiment that students in the US generally associated negative stereotypical images to Spanish accented speakers, independent of the accent strength. However, this study was carried out in the US, where the Hispanics are a minority group (Ramirez, 2004). Therefore, the current study aims to investigate the role of stereotyping in an European context by including the stereotype content model of Fiske, Cuddy, Glick, and Xu (2002) to examine whether Dutch and Germans have stereotypical images linked to

(6)

Spanish, Arabic and English people and whether these images influence their evaluations on non-native accentedness.

Evaluation of accents and determining variables

Status, solidarity and dynamism have been found to affect the evaluation of an accent (Giles & Billings, 2004). Fuertes, Gottdiener, Martin, Gilbert and Giles (2012) revealed in a meta-analysis on the effects of standard and non-standard accents, that speakers of standard English generally were evaluated more positively than non-native accented speakers on three dimensions (dynamism, solidarity and status). Furthermore, it was found that non-native accents had a strong effect on dynamism and status, but a less strong effect on solidarity. The current study aims to examine the three accents on these dimensions.

Other studies have found that non-native accents are negatively linked to power and status. For instance, the study by Śliwa and Johansson (2014) showed that native speakers of English are evaluated as having more power than non-native speakers. Likewise, Gluszek and Dovidio (2010) found in their analysis of empirical literature on accents, that Western European accents of English such as Dutch and German are often evaluated more positively than other accents, as these other accents such as Southern European and Middle Eastern languages are often related to less developed countries than Western European languages.

Similarly, the study of Giles and Billing (2004), revealed that English accents are linked to a higher social and economic status, and are evaluated higher, with opposite results for non-native accents, such as Arabic and Spanish accents. Russo, Islam and Koyuncu (2017) found that more prestigious accents such as native British are evaluated more positively. Therefore it is expected that the British accents are evaluated higher on the previous variables than the Arabic-English and Spanish-English accents.

A study by Hendriks, van Meurs and Reimer (2018) examined non-native English accents of Dutch teachers on several dimensions such as actual comprehension, attitude, familiarity, proficiency and perceived intelligibility. The accent strength of the speakers was taken into account, as moderate accents were perceived as less proficient than weak accents (Hendriks et al., 2018). It was concluded that strong accents lead to more negative evaluations and that identifying the country of origin of the speaker was easier than for other accent strengths. Therefore, this study focusses on strong accents,

(7)

to examine whether these conclusions also holds for other non-native accents and in the hiring genre.

Furthermore, the English proficiency of the listener to the accent is important (Beinhoff, 2014). It has been found that the proficiency has a significant effect on the perceived intelligibility of the speaker. When the proficiency is higher, the speaker is found to be more intelligible. The intelligibility is the competency of the listener of understanding a message correctly (Beinhoff, 2014). Therefore, the participants need to self-assess their English and participate in a LexTALE Test.

Previous research on non-native English accents in employment settings

Following Russo et al. (2017), speakers with a non-native accent are often discriminated in the workplace, which might influence employment-related decisions. They

hypothesized that the degree of communication requirements may affect the reactions of managers to the accent, such that the stronger the requirements, the more likely

stereotyping occurs. This is in line with the results of Adler (1987) and Hosoda and Stone-Romero (2010), who found that job applications with a non-native accent were not chosen for careers with a high degree of communication requirements. Hence, a high degree of communication requirements is included in the materials of this current study, to investigate whether the communication requirements influence the evaluation of the accentedness.

Spanish accents have already received some attention with regard to

accentedness in the hiring context (Carlson & McHenry, 2006; Deprez-Sims & Morris, 2010). However, most studies have been conducted only in the United States as there is a predominance of Hispanic people who live and work there (Ramirez, 2004).

For example, Deprez-Sims and Morris (2010) investigated on the dimensions accentedness, understandability and similarity whether Midwestern US, French and Colombian accents were evaluated differently in the US in the context of a job interview. It was found that the French accent obtained inferior evaluations to the Midwestern US accent, as it is was perceived as less understandable and less similar.

Moreover, a study by Carlson and McHenry (2006) investigated the influence of comprehensibility and ethnicity on the employability of the speaker in the US. Spanish, African, and Asian accents with different strengths were compared in a hiring context. They chose to compare Spanish, African and Asian accents with different strengths as these accents are most familiar in the area of Texas, where the study was conducted.

(8)

They showed a possible effect of familiarity; the Spanish-accented speaker scored higher on employability when speaking with a strong accent than the African- and the Asian- accented candidate. However, as previous studies mostly focus on Hispanics in The United States, the current study aims to investigate the Spanish-English accent in a European context, in the current study The Netherlands and Germany (Gallardo del Puerto et al., 2015; Hendriks et al., 2017).

Furthermore, the study by Kraut and Wulff (2013), compared Asian, Hispanic and Middle Eastern non-native accents in a multifactorial study with native English speakers as evaluators, and showed different results than Carlson and McHenry (2006). They revealed that Hispanic-accented speakers obtained lower evaluations than the Middle Eastern-accented speakers, which could imply that as Hispanic accents are more familiar in the US, they are more critically evaluated because of higher expectations (Nejjari, Gerritsen, Van der Haagen & Korzilius, 2012). Moreover, the study found that a higher familiarity with the accent results in a better comprehension (Kraut & Wulff, 2013). These previous studies showed that familiarity and similarity with the accents influences the evaluation on non-native accentedness.

The evaluation of non-native Arabic-English accents in a European context has not received much attention yet, while Arabs are a recognizable migration group both The Netherlands and Germany (Hagendoom & Hraba, 1987). One study found on Arabic-accented English is the study by Buckingham (2014) in the Gulf region. In an experiment, five qualified male English speakers with common accents of the Gulf region were used to examine whether university students prefer the English accent or the non-native accent English of the teachers (Buckingham, 2014). It became apparent that the speakers preferred the English accents over the non-native accents and that the students prefer the Standard British English accent over other English accents.

However, this was an evaluation made by native speakers of the accents and the design is therefore different than the current study with non-native evaluators. In the study of Grondelaers, van Gent, and van Hout (2015) evaluations of regional and ethnic

Standard Dutch accents such as Moroccan accent were examined. It was discovered that the Moroccan-accented Dutch was perceived as less superior than the regional native Dutch accents. However, as the study by Grondelaers et al. (2015) is based on the Arabic-Dutch accent and not on the Arabic-English accent, the current study aims to examine whether these perceptions also hold in a job interview context and when the accent is in English rather than in Dutch.

(9)

Scientific and practical relevance

Currently not many studies have examined the effects of a non-native accent on the evaluations of non-native English listeners with different native backgrounds. Given the increasing number of non-native English speakers, such research would make a

worthwhile contribution. Moreover, previous research mainly focused on native

environments (Deprez-Sims & Morris, 2010). It is of great importance to know whether there is a certain native “norm” of accented English in the workplace for job

interviewers, and whether the same attitude occurs towards two different non-native accents.

This paper is of practical relevance as the group of immigrants who are potential employees is increasing, and therefore using a CCL in a multinational company will be of great importance. The current study could provide a framework for multinational companies to avoid discrimination on non-native accents. By identifying possible positive and negative outcomes of having a non-native accent, solutions might be found to change the current negative consequences into positive consequences. In other words, the aim of the current study is to raise awareness about possible discrimination on the basis of non-native accents in English as a lingua franca.

The following research question is formulated in order to expand research in the field of accents and employability: To what extent do Dutch and German as non-native

English speakers evaluate a Spanish English, an Arabic English and a native British-English accent differently in terms of employability for a high communication job?

Method

Dutch and German individuals were asked to evaluate Spanish English, Arabic English and British accents in an experimental design.

Materials

The independent manipulated variable was ‘Accent’. The stimulus materials consisted of audios of six male speakers each with a different country of origin (Spain, Arabic, native British-English). Only male speakers were used to prevent a bias based on gender.

(10)

The speakers recorded a script (See Appendix I), in which they acted as a job interviewee. Moreover, no applicant name was used for all three accent conditions, so there was no possibility for biases based on names (Purkiss et al., 2006). It was aimed to simulate a job interview for a high-communication job (Junior Communication

Analysist).

Every participant was randomly assigned to one audio file. After listening to the audio file, a list of questions was asked, which results in a total of 29 questions per participant. The main questionnaire can be found in Appendix III.

Pre-test

Fourteen expert judges (age: M = 26.64, SD = 12.80; range 20-58) from Radboud University evaluated 12 audios (four Spanish-accented speakers, four Arabic-accented speakers and four British speakers) on voice characteristics, accent strength and

comprehensibility in a pre-test, to select a total of six suitable recordings. Twelve judges were undergraduate students from the International Business Communication bachelor study, and two judges were trained specialists in researching non-native English

accents. Based on these evaluations, for each accent condition two speech samples were chosen. Additionally, accent strength was measured, as the Spanish and Arabic accents needed to be perceived as ‘strong’ (Hendriks et al., 2018). Table 1 shows the means and standard deviations for accent strength, voice characteristics and comprehensibility for the selected speakers of each accent.

(11)

Table 1. Means and standard deviations (between brackets) for accent strength, voice characteristics and comprehensibility for the selected speakers of each accent condition by expert evaluation (N = 14) (1 = very weak/very low, 7 = very strong/very high). Accent condition

Spanish Arabic British

Sp1 Sp2 Sp1 Sp2 Sp1 Sp2 Accent strength 6.33 6.43 6.21 5.83 5.81 5.52 (.76) (.64) (.61) (.69) (.84) (.76) Voice characteristics 3.97 3.46 3.02 3.89 3.99 4.34 (.89) (.85) (.85) (.83) (.81) .61) High pitch 3.29 2.95 2.53 2.62 2.62 3.18 (1.47) (1.19) (1.21) (1.47) (1.61) (1.44) Slowness 3.83 5.3 5.19 3.73 3.70 3.29 (1.47) (1.32) (1.35) (1.51) (1.96) (1.72) Comprehensibility 4.59 3.53 4.25 5.15 4.51 5.00 (1.55) (1.42) (1.71) (1.35) (1.68) 1.54)

See Appendix I for the script and consent forms of the speakers and Appendix II for the pre-test questionnaire.

Subjects

A total of 131 Dutch participants (age: M = 31.72, SD = 14.38; range 17-63) and 80 German participants (age: M = 28.71, SD = 12.46; range 20-67) took part in the experiment. Of the 131 Dutch participants, 54 were male, 74 were female, one would rather not say their gender and two defined themselves as ‘other’. Of the 80 German participants, 26 were male, 51 were female and three would rather not say their gender.

With regard to educational level, the German participants’ most frequent obtained level of education was university (60%; no difference between bachelor and master), followed by high school (22%). The range was from less than high school to university. The Dutch participants’ most frequent obtained educational level was university (48.9%; no difference between bachelor or master), followed by high school (28.2%). The highest educational level of the Dutch participants’ ranged from less than high school to university. Both nationalities rated their English skills as moderately

(12)

good (Dutch: M = 21.11, SD = 4.19; German: M = 21.83, SD = 4.67). Dutch

participants had more experience as an applicant (M = 5.08, SD = 1.64) than interviewer (M = 3.26, SD = 2.14) and German participants had more experience as an applicant (M = 4.88, SD = 1.76) than interviewer (M = 2.93, SD = 2.13).

Distribution

Nationality (F (2, 208) < 1), gender (χ2 _{(6) = 7.86, p = .248), age (F (2, 207) = 1.63, p =}

.198), educational level (χ2 _{(8) = 9.57, p = .297), English proficiency (F (2, 208) < 1),}

self-assessment (F (2, 208) = 1.63, p = .198), job interview experience (F (2, 208) < 1), were equally distributed across all three accent conditions and therefore did not

influence possible differences. Gender (t (209) = 1.08, p = .283), age (t (185.39) = 1.06,

p = .111), educational level (t (209) = .90, p = .372), English proficiency (t (209) = .95, p = .342), self-assessment (t (208) = 1.14, p = .254), and job interview experience (t

(209) = 1.27, p = .206), were equally distributed between Dutch and German listeners and therefore did not influence possible differences.

Design

The study used a 3x2 between-subjects design, with accent (Spanish, Arabic, British), and listeners’ nationality (Dutch or German) as the independent variables. The

participants were randomly assigned to one of the six speech samples. The native-English accent condition was used as a control group.

Instruments

The instrument used for this experiment was an online questionnaire in which the German and Dutch participants evaluated one out of the six speech samples on: attitude towards the speaker, employability, perceived comprehensibility, accent strength, voice characteristics. Moreover, the participants needed to identify the origin of the speaker and indicate the familiarity with job interviews. Moreover, with the stereotype content model (SCM) stereotypical images were measured, and the English proficiency was tested with the LexTALE test. Composite means were calculated when the inter-item reliability was adequate or good.

Attitude towards the speaker was measured for the dimensions “Perceived status”,

“Solidarity” and “Dynamism” based on Giles and Billings (2004) with the statement: “The speaker sounds” for all three dimensions. For “Perceived status” the items

(13)

“confident, ambitious, competent, intelligent” were used (α = .793). For “Solidarity” the items “reliable, attractive, benevolent, like me” were used (α = .682). For “Dynamism” the items “enthusiastic, active, chatty, lively” were used (α = .873). All the items were anchored on 7-point Likert scales ranging from (1) completely disagree to (7)

completely agree.

Employability was measured by the statements “I think this speaker is suitable for

the position”, “I would hire the speaker” and “I would recommend to hire the speaker” anchored by a 7-point Likert scales ranging from (1) completely disagree to (7)

completely agree based on Deprez-Sims and Morris (2010) (α = .929).

Perceived comprehensibility was measured by using the statement “I found this

person easy to understand” followed by a 7-point Likert scale ranging from (1) completely disagree to (7) completely agree based on Carlson and McHenry (2006).

Manipulation check

Although accent strength was manipulated and checked in the pre-test, the perceived accent strength by the listeners is of great importance to see if the listeners agree on the accent being foreign.

The perceived accent strength was measured based on Deprez-Sims and Morris (2010) on a 7-point scale with the statement: “This speaker has a strong foreign accent in his English” ranging from (1) no foreign accent to (7) a strong foreign accent, and the statement “This speaker sounds like a native speaker of English” ranging from (1) totally disagree to (7) totally agree.

Identification of origin of the speaker can be examined by the following question

from Hendriks et al., (2018): “What country does this speaker come from?”. The listener was able to choose from a list containing all 193 countries. Similar to Hendriks et al. (2018), for the Spanish speakers all the official Spanish speaking countries were accepted as correct identifications: Argentina, Bolivia, Chile, Colombia, Costa Rica, Cuba, Dominican Republic, Ecuador, Equatorial Guinea, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Pero, Puerto Rico, Spain, Uruguay and Venezuela. For the identification of the Arabic accent, all the official Arabic speaking countries were accepted as correct identifications: Algeria, Bahrain, Chad, Comoros, Djibouti, Egypt, Eritrea, Iraq, Jordan, Kuwait, Lebanon, Libya, Mauritania, Morocco, Oman, Palestine, Qatar, Saudi Arabia, Somalia, Sudan, Syria, Tanzania, Tunisia, United Arab Emirates and Yemen. Turkey as the accent’s origin was

(14)

considered correct as well for the Arabic condition, due to possible unfamiliarity with the accent. For the British accent, all ‘inner-circle countries’ were accepted (Australia, Canada, New Zealand, UK, US).

Voice characteristics was evaluated by following the statements on a 7-point

Likert scale: “This speaker sounds natural”, “This speaker sounds fluent”, “This speaker sounds monotonous”, “This speaker speaks pleasant”, “This speaker sounds hesitant” (α = .607). The two other items “the speaker speaks high-pitched” and “the speakers speaks slowly” were analyzed separately, as they were anchored differently.

Background variables

Moreover, the participants were asked if they had experience with job interview

situations. This dependent variable was examined by the following statements “I have

experience in being a job interviewee” and “I have experience in being a job interviewer” anchored by Likert scales ranging from (1) completely disagree to (7) completely agree based on Deprez-Sims and Morris (2010).

To check whether the evaluation of non-native accents is linked to stereotypical

images, the dimensions “warmth” and “competence” of the stereotype content model

developed by Fiske et al. (2002) were measured with the items “confident”, “sincere”, “well-educated”, “warm”, “competent”, “economically successful”. Moreover two additional items were added: “If Spanish/Arabic/British people get special breaks, this is likely to make things more difficult for people like me” and “Resources that go to Spanish/Arabic/British people are likely to take away from the resources of people like me” on 7-point Likert scales anchored from (1) completely disagree to (7) completely agree (Fiske et al., 2002).

Moreover, the English proficiency of the participants was measured, as it has been shown that this might influence comprehension and evaluation (Beinhoff, 2014). Therefore, the participants have been asked to perform a LexTALE test, which

measures actual proficiency level by asking the participants to indicate whether the words shown are English words or not (Lemhöfer & Broersma, 2012).

The self-assessment of the English proficiency was measured by the following four items: writing, reading, listening and speaking on a 5 point scale ‘poor’ –

‘excellent’ based on Hendriks et al. (2018) (α = .91).

Finally, the participants were asked to fill in personal information such as gender, age, highest level of education completed, profession, nationality (country of

(15)

origin), and mother tongue. The composite means were calculated for all scales, since Cronbach’s alpha was >. 60 or higher.

Procedure

The participants were sampled based on convenience sampling via social media. The online questionnaire was available via mobile phone or laptop and was administered by Qualtrics. The experiment was conducted on an individual basis. No information was displayed about the purpose of the study and the speakers’ nationality to prevent biases. Moreover, the participants signed a consent form beforehand and were told that

participation was voluntary and that there was the possibility to stop at any time. The participants were thanked in advance. The questionnaire was published in English, to prevent biases based on language of instruction. Therefore, a language proficiency test was used to test to what extent differences in outcomes were due to a low English proficiency of the listeners. The participants were told that they were going to listen to a job applications which was followed by one speech sample with the including

evaluation scales. On average, completing the questionnaire took 15 minutes.

Statistical treatment

By using SPSS, several statistical tests were carried out to examine differences in the evaluation. A Chi-square test and ANOVA’s were used to ensure homogeneity among the participants of the three accent conditions. Moreover, to test for country of origin recognition of the speaker, a chi-square test was used. Lastly, two-way ANOVA’s with Bonferroni follow-up tests were used to examine the effects of the dependent interval variables on accent condition (Spanish English, Arabic English, British English) and listeners’ nationality (Dutch, German).

Results

The aim of this study was to investigate to what extent Dutch and German as non-native English speakers evaluate a Spanish English, an Arabic English and a native British-English accent differently in terms of employability for a high communication job. Differences between speakers within accent conditions

(16)

on the manipulation variables, t-tests were conducted. The significant differences are mentioned.

Spanish

An independent samples t-test showed a significant difference between the first Spanish speaker and the second Spanish speaker with regard to voice characteristics (t (82) = 2.69, p = .009). The first Spanish speaker (M = 3.97, SD = .89) was evaluated more positively on voice characteristics than the second Spanish speaker (M = 3.46, SD = .85). An independent samples t-test showed a significant difference between the first Spanish speaker and the second Spanish speaker with regard to slowness (t (82) = 4.85,

p < .001). The second Spanish speaker (M = 5.30, SD = 1.32) was found to be speaking

slower than the first Spanish speaker (M = 3.83, SD = 1.47). An independent samples t-test showed a significant difference between the first Spanish speaker and the second Spanish speaker with regard to comprehensibility (t (82) = 3.24, p = .002). The first Spanish speaker (M = 4.59, SD = 1.55) was found to be more comprehensible than the second Spanish speaker (M = 3.53, SD = 1.42).

Arabic

An independent samples t-test showed a significant difference between the first Arabic speaker and the second Arabic speaker with regard to perceived accent strength (t (60) = 2.29, p = .026). The first Arabic speaker (M = 6.21, SD = .61) was found to have a stronger accent than the second Arabic speaker (M = 5.83, SD = .69). An independent samples t-test showed a significant difference between the first Arabic speaker and the second Arabic speaker with regard to voice characteristics (t (60) = 4.02, p <.001). The second Arabic speaker (M = 3.89, SD = .83) was evaluated more positively on voice characteristics than the first Arabic speaker (M = 3.02, SD = .85). An independent samples t-test showed a significant difference between the first Arabic speaker and the second Arabic speaker with regard to slowness (t (60) = 4.01, p < .001). The first Arabic speaker (M = 5.19, SD = 1.35) was found to be speaking slower than the second Arabic speaker (M = 3.73, SD = 1.51). An independent samples t-test showed a significant difference between the first Arabic speaker and the second Arabic speaker with regard to comprehensibility (t (59.51) = 2.32, p = .024). The second Arabic speaker (M = 5.15,

SD = 1.35) was found to be more comprehensible than the first Arabic speaker (M =

(17)

No differences found between the two British speakers. Since the differences between the two speakers in each condition were relatively small, the following analysis will include all six speakers.

Manipulation check

The present study includes a manipulation check in order to see whether the speakers in each condition are similar in terms of perceived accent strength and voice

characteristics. In Table 2, the means and standard deviations of the manipulation checks can be found.

Perceived accent strength

A two-way analysis of variance with listeners’ nationality and accent as factors showed a significant main effect of accent on perceived accent strength (F (2, 205) = 12.98, p <.001). Listeners’ nationality was not found to have a significant main effect on perceived accent strength (F (1, 205) <1). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) = 2.28, p = .105). The Spanish accent (M = 6.38, SD = .70) was perceived as stronger than the Arabic (p = .020, Bonferroni-correction; M = 6.05, SD = .67) and the British accent (p <.001, Bonferroni-correction; M = 5.69, SD = .82). The Arabic accent was perceived as stronger than the British accent (p = .016, Bonferroni-correction). The manipulation check can therefore be seen as partly successful, as the Spanish accent should not have been perceived as stronger than the Arabic accent.

Voice characteristics

A two-way analysis of variance with listeners’ nationality and accent as factors showed a significant main effect of listeners’ nationality on voice characteristics (F (1, 205) = 11.80 p = .001) Accent was found to have a significant main effect on voice

characteristics (F (2, 205) = 11.61, p <.001 ). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) <1). The German participants (M = 4.01, SD = .92) evaluated the speakers in general higher on voice characteristics than the Dutch participants (M = 3.59, SD = .87). The British accent (M = 4.14, SD = .91) was evaluated higher on voice characteristics than the Spanish accent (p = .008, Bonferroni-correction; M = 3.71, SD = .90) and the Arabic accent (p <.001, Bonferroni-correction; M = 3.39, SD = .94).

(18)

High-pitched

A two-way analysis of variance with listeners’ nationality and accent as factors showed a significant main effect of listeners’ nationality on high pitch (F (1, 205) = 14.07, p

<.001) Accent was not found to have a significant main effect on high pitch (F (2, 205)

= 1.72, p = .181). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) = 1.77, p = .172). The Dutch participants (M = 3.17, SD = 1.38) found the speakers in general speaking with a higher pitch than the German participants (M = 2.40, SD = 1.35).

Slowness

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on slowness (F (1, 205) <1) Accent was found to have a significant main effect on slowness (F (2, 205) = 7.49, p = .001). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) = 1.04, p = .355). The British speakers (M = 3.52, SD = 1.86) was found to be speaking faster than the Spanish speakers (p < .001, correction; M = 4.58, SD = 1.57) and the Arabic speakers (p = .001, Bonferroni-correction; M = 4.58, SD = 1.58).

Table 2. Means and standard deviations (between brackets) for perceived accent strength and voice characteristics in function of the accent condition (1 = very weak/very low, 7 = very strong/very high)

Accent

n = 84 n = 62 n = 65

M (SD) M (SD) M (SD)

Perceived accent strength 4.83 (1.23) 4.15 (1.09) 4.83 (1.06) Voice characteristics 3.97 (.95) 3.67(.99) 3.89(.96) High pitch 3.93(1.26) 3.00(1.26) 3.55(1.44) Slowness 4.58 (1.257) 4.58(1.58) 3.52(1.86)

(19)

Identification of origin

Firstly, participants were asked if they were able to identify the country of origin. A Chi-square test showed a significant relation between accent and ability to recognize of the country of origin (χ2 (2) = 8.15, p = .017). Participants who listened to the Spanish accent (60.7%) thought they could identify the origin more than the participants in the Arabic (38.7%) and the British condition (43.1%). However, a total of 108 persons (51.2%) was not able to identify the country of origin.

Moreover, the participants who thought that they were able to identify the country of origin, were asked to fill in the specific country of origin. A Chi‐square test showed a significant relation between accent and correct association of country of origin (χ2 _{(2) =}

22.60, p <.001). Participants who listened to the British accent identified the country of origin gave more correct answers (82.8%) than incorrect (17.2%) than participants in the other two accent conditions. However, for the Spanish accent the participants gave relatively more incorrect answers (61.8%) than correct answers (38.2%). Moreover, for the Arabic accent the participants gave relatively more incorrect (73.1%) than correct answers (26.9%). Table 3 represents the count and percentages of correctly and incorrectly country of origin guesses for the three accent conditions.

Table 3. Counts and percentages of the correct and incorrect answers for country of origin identification in function of the accent.

Accent

n = 55 n = 27 n = 29

Origin Correct N 21a ₇ a ₂₄ b

% 38.2% 26.9% 82.8%

Incorrect N 34 a 20 a 5 b

% 30.9% 73.1% 17.2%

Note: Each subscript letter denotes a subset of Accent categories whose column proportions do not

differ significantly from each other at the ,05 level.

Note: The average total correct answers was 47.3% and the average total incorrect answers was

(20)

Evaluation of the accents

Status

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on status (F (1, 205) <1). Accent was found to have a significant main effect on status (F (2, 205) = 8.13, p <.001). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) <1). The Arabic accent (M = 4.15, SD = 1.09) was evaluated lower on status than Spanish (p = .001, Bonferroni-correction; M = 4.83, SD = 1.23) and the British accent (p = .002, Bonferroni-correction; M = 4.83, SD = 1.06). There was no difference between the latter ones.

Solidarity

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on solidarity (F (1, 205) <1). Accent was not found to have a significant main effect on solidarity (F (2, 205) = 1.51, p = .224). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) <1).

Dynamism

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on dynamism (F (1, 205) <1). Accent was found to have a significant main effect on dynamism (F (2, 205) = 7.02, p = .001). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) <1). The Spanish accent was evaluated higher on dynamism (M = 3.93, SD=1.26) than the Arabic accent (p <.001, Bonferroni-correction; M = 3.00, SD = 1.26). There was no significant difference found between the British accent condition and the other two conditions.

Employability

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on employability (F (1, 205) = 2.03, p

= .156). Accent was found to have a significant main effect on employability (F (2,

205) = 5.32, p = .006). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) <1). The Arabic accent was evaluated lower

(21)

on employability (M = 3.38, SD = 1.35) than the Spanish (p = .006, correction; M = 4.07, SD = 1.28) and the British accent (p = .043,

Bonferroni-correction; M = 3.96, SD = 1.36). However, there was no significant difference between the Spanish and the British accent.

Comprehensibility

A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on comprehensibility (F (1, 205) <1). Accent was not found to have a significant main effect on comprehensibility (F (2, 205) = 2.82, p = .062). The interaction effect between listeners’ nationality and accent was not statistically significant (F (2, 205) = 2.10, p = .125).

Table 4. Means and standard deviations (between brackets) for status, solidarity,

dynamism, employability and comprehensibility in function of the accent condition (1 = very negative evaluation, 7 = very positive evaluation).

Accent

Spanish Arabic British n = 84 n = 62 n = 65 M (SD) M (SD) M (SD) Status 4.83 (1.23) 4.15 (1.09) 4.83 (1.06) Solidarity 3.97 (.95) 3.67 (.99) 3.89 (.96) Dynamism 3.93 (1.26) 3.00 (1.26) 3.55 (1.44) Employability 4.07 (1.28) 3.38 (1.35) 3.96 (1.36) Comprehensibility 4.05(1.57) 4.63 (1.62) 4.72 (1.63)

Stereotype content model

Of great importance is the stereotype content model in this paper, as with this model the overall stereotypes of Spanish, Arabs and British in The Netherlands and Germany were investigated. A two-way analysis of variance with listeners’ nationality and accent as factors showed no significant main effect of listeners’ nationality on stereotypical images (F (1, 205) <1). Accent was found to have a significant main effect on stereotypical images (F (2, 205) = 10.94, p <.001). The interaction effect between

(22)

listeners’ nationality and accent was not statistically significant (F (2, 205) = 1.14, p = .321). Arabs were stereotyped more negatively (M = 4.18, SD = .79) in both The

Netherlands and Germany than British (p <.001, Bonferroni-correction; M = 4.76, SD = .64) and Spaniards (p = .002, Bonferroni-correction; M = 4.58, SD = .65). However, there was no significant difference between the latter ones.

Discussion and conclusion

Communication between non-native speakers is inevitable in the on-going process of globalization (Deprez-Sims & Morris, 2010). The aim of the current study was to investigate the effects of Spanish and Arabic non-native English accents on the evaluation of speakers for a high-communication job by non-native English speakers. This might have major implications for the job application processes in German and Dutch multinationals where English is used as a corporate language.

Evaluation of the candidate

With regards to status, solidarity and dynamism as dimensions of the attitude towards the speaker, mixed results were found.

With respect to status, the Arabic accent was evaluated significantly lower than the English accent, which is in line with previous studies (Fuertes et al., 2012; Giles & Billings, 2004; Gluszek & Dovidio, 2010; Śliwa and Johansson, 2014). These studies showed that non-native accented speakers are perceived as having less confidence, less ambition, less competence and less intelligence. However, no significant results were found for the non-native Spanish accent in comparison to the British accent, which makes the outcome striking.

There were no significant differences found with regard to solidarity between the three accent conditions, which is in agreement with the findings from Hendriks et al. (2018), who found that Dutch speakers with different accent strengths in English did not differ significantly on how French, German, and Spanish listeners evaluated the speaker on solidarity. Fuertes et al. (2012) found in a meta-analysis that in general non-native accents have inferior evaluations on solidarity, but that several studies also found higher

(23)

evaluations for non-native accents on solidarity. It might be that there are other factors confounding these results and that the solidarity dimension yields no clear results.

Moreover, the Spanish accent is evaluated significantly higher on dynamism than the Arabic accent with no significant differences in comparison to the British accent. These results are not in line with previous research examined in the meta-analysis of Fuertes et al. (2012), in which non-native accents were found to be downgraded on being enthusiastic, active, chatty and lively.

Moreover, the Arabic accent was evaluated more negatively on employability in comparison to the British accent, which is in line with the results of Deprez-Sims and Morris (2010), in which non-native accents were downgraded on employability. However, the non-native Spanish accent is not significantly different from the English accent, which is not in line with the study of Hosoda et al. (2012), who have found that Spanish-accented speakers were less likely to be hired in comparison to native speakers.

No differences have been found for comprehensibility, meaning that the results were not in line with previous research who found that non-native accents are found to be less comprehensible than native accents (Carlson & McHenry, 2006; Hendriks et al., 2018).

The Spanish accent was found to be stronger than the Arabic accent, but equally comprehensible. However, the Spanish accent did not have any significant inferior evaluation in comparison to the British accent, whereas the Arabic accent did got inferior evaluations. This is again most findings on accent strength, in which stronger accents are generally obtaining inferior evaluations on multiple dimensions than weaker accent strengths (Carlson & McHenry, 2006; Hendriks et al., 2018; Nejjari et al., 2012). Therefore, it might be concluded that the accent strength does not influence the accent evaluations.

The current study included the stereotype content model to measure

stereotypical associations (Fiske et al., 2002). Striking are the significant results that only Arabs have a more negative stereotypical image and that Spaniards had a similar strength of stereotypical images as British in the Netherlands and Germany, which are similar to the main results on attitude towards the speaker and the employability. Therefore, the stereotype content model could be a possible explanation for the significant differences in evaluations. Thus, these coherent results might indicate prejudice and stereotypes at the workplace (Biernat & Dovidio, 2000; Gluszek & Dovidio, 2010; Krings & Olivares, 2007).

(24)

Another plausible explanation might be that these negative stereotypical images of the Dutch and the German are related to the social identity theory of Tajfel (2010), in which the Dutch and the German possible perceive Arabs as an out-group, as being a large group of immigrants (Hagendoom & Hraba, 1987). Moreover, immigrants are often minority groups in the host countries. In line with Hosoda et al. (2012), a minority group has stronger stereotypical images, which might occur with Arabs as immigrants in The Netherlands and Germany. Moreover, the similarity-attraction bias might have a possible link to the results, where Arabs might be perceived as more dissimilar than Spaniards and British (Condon & Crano, 1988), which could be possibly based on geographical or linguistical associations. These biases might have led to the inferior evaluation for Arabs and the similar evaluations for Spaniards and British on the dimensions measured. However, as the present study did not measure explicitly for similarity and familiarity such as Deprez-Sims and Morris (2010) and Nejjari et al., (2012), the link is limited.

In conclusion, several effects have been found of Spanish and Arabic non-native English accents on the evaluation of speakers for a high-communication job by non-native English speakers. It has been found that the inferior evaluations for non-non-native accents only hold for the Arabic-English accent. The Spanish accent was mostly not significantly different from the British accent, which is not in line with previous studies. As the stereotype content model shows the same inferior results for the Arabic accent as the main results and previous studies, stereotypical images might be a plausible

explanation for the differences in evaluations.

Limitations and suggestions for further research

The first limitations are found in the manipulation checks, which showed several significant differences on voice characteristics and accent strength between the two speakers within each accent condition. These discrepancies might have confounded the evaluation of the accentedness. It is recommended for further research to select more similar speakers to prevent confounding results. Another limitation was that the current study only recorded male speakers. Male speakers could have been evaluated differently as previous studies have found gender bias in the hiring genre (Swim, Borgida,

Maruyama, & Myers, 1989). Therefore, further research should include female speakers in order to examine the effect of gender on hiring decisions.

(25)

The large number of students among the participants, combined with the little experience of being a job interviewer, was a clear limitation of the current study as it has led to a low ecological validity. As the obtained results are of particular interest for job recruiters, it is suggested for future research to examine evaluations of professional recruiters who are experienced with job interviews in order to be able to make

generalizations to real-life settings. Moreover, the external validity can be seen as rather low, as only Dutch and German participants were included. Following Hendriks et al., (2018) these countries are not only geographically, but also linguistically and culturally close. Further research should include countries who are linguistically and culturally different, to examine whether non-native accentedness is perceived differently in other countries, or whether the results are rather universal. Lastly, the questionnaire was found to be exhaustive, as many participants stopped before the end. It is recommend for further research to make the speech samples shorter or to ask fewer questions.

Regarding the results, a high number of participants incorrectly guessed the country of origin, which might have led to different stereotypical images when

evaluating the accents. However, Lindemann (2003) mentions that the identification of foreignness itself leads to downgraded evaluations on accentedness, than rater the correct guessing of the country of origin. Moreover, the Spanish and British accent might be perceived as ‘less’ foreign, as part of Europa, and therefore obtained higher evaluations than the Arabic accent (Cargile, Maeda, Rodriguez, & Rich, 2010).

An additional limitation addresses the possible influence of familiarity and similarity. Although the present study has found significant differences on the stereotype content model, it did not measure for similarity nor familiarity. Further research should include these dimensions in order to get a clearer picture on accent evaluations and the influence of stereotypical associations.

Implications

This study offers insight into the evaluations of non-native English speakers on the employability of Spanish and Arabs in the business-world. By using a European and an Arabic accent, an English accent control group and the measurement of the stereotype content model, the study gives a different perspective of the effects of accent for hiring success than previous research. In line with previous studies, the presence of an accent when speaking English influences the perceived employability and thus job

(26)

implements the stereotype content model of Fiske et al. (2002). Therefore, the current study contributes to the theory that evaluations on non-native accents might be related to stereotype images. Practical implications are of particular interest for recruiters of a multinational company in native or non-native environment, who should be aware of the evidence that non-native accents are often negatively evaluated in job interview settings and therefore not base evaluations on audio-cues only. For example, workshops could be given to raise awareness on stereotyping during job interviews. In conclusion, it is recommended to be conscious of possible stereotyping or discrimination during job interviews, especially in international workplace contexts, so fair application processes could be made possible.

(27)

References

Adler, S. (1987). Bidialectalism? Mandatory or elective? Asha, 29, 41–42.

Beinhoff, B. (2014). Perceiving intelligibility and accentedness in non-native speech: A look at proficiency levels. Concordia Working Papers in Applied Linguistics, 5, 58-72.

Biernat, M., & Dovidio, J. F. (2000). Stigma and stereotypes. In T. F. Heatherton, R. E. Kleck, M. R. Hebl, & J. G. Hull (Eds.), The social psychology of stigma (pp. 88 125). New York: Guilford.

Buckingham, L. (2014). Building a career in English: Users of English as an additional language in academia in the Arabian Gulf. Tesol Quarterly, 48(1), 6-33.

Carlson, H. K., & McHenry, M. A. (2006). Effect of accent and dialect on employability. Journal of Employment Counseling, 43(2), 70-83.

Cargile, A. C., Maeda, E., Rodriguez, J., & Rich, M. (2010). " Oh, You Speak English So Well!": US American Listeners' Perceptions of" Foreignness" among Nonnative Speakers. Journal of Asian American Studies, 13(1), 59-79.

Centraal Bureau voor de Statistiek. (2019, October 17). Hoeveel immigranten komen naar Nederland? Retrieved from https://www.cbs.nl/nl-nl/dossier/dossier-asiel migratie-en-integratie/hoeveel-immigranten-komen-naar-nederland-

Condon, J. W., & Crano, W. D. (1988). Inferred evaluation and the relation between attitude similarity and interpersonal attraction. Journal of Personality and Social

Psychology, 54(5), 789

Crystal, D. (2012). English as a global language. Cambridge Uiversity Press.

Deprez-Sims, A. S., & Morris, S. B. (2010). Accents in the workplace: Their effects during a job interview. International Journal of Psychology, 45(6), 417-426. Esses, V. M., Jackson, L. M., & Armstrong, T. L. (1998). Intergroup competition and

attitudes toward immigrants and immigration: An instrumental model of group conflict. Journal of Social Issues, 54, 699-724.

Federal Statistical Office of Germany. (2019, April 15). Retrieved March 4, 2020, from https://www.destatis.de/EN/Themes/SocietyEnvironment/Population/Migration- Integration/Tables/foreign-population-laender.html

Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed)

stereotype content: competence and warmth respectively follow from perceived status and competition. Journal of personality and social psychology, 82(6), 878.

(28)

Fuertes, J. N., Gottdiener, W. H., Martin, H., Gilbert, T. C., & Giles, H. (2012). A meta- analysis of the effects of speakers' accents on interpersonal evaluations

European Journal of Social Psychology, 42(1), 120-133.

Gallardo del Puerto, F., García Lecumberri, M. L., & Gómez Lacabex, E. (2015). The assessment of foreign accent and its communicative effects by naïve native judges vs. experienced non‐native judges. International Journal of Applied

Linguistics, 25(2), 202-224.

Giles, H., & Billings, A. C. (2004). Assessing language attitudes: Speaker evaluation studies. In A. Davies, & E. Elder (Eds.), The handbook of applied linguistics, 187-209.

Gluszek, A., & Dovidio, J. F. (2010). The way they speak: A social psychological perspective on the stigma of nonnative accents in communication. Personality

and Social Psychology Review, 14(2), 214-237.

Grondelaers, S. & van Gent, P. (2019). How “deep” is Dynamism? Revisiting the evaluation of Moroccan-flavored Netherlandic Dutch. Linguistics Vanguard, 5(s1).

Grondelaers, S., van Gent, P., & van Hout, R. (2015). Is Moroccan-flavoured Standard Dutch standard or not? On the use of perceptual criteria to determine the limits of standard languages. Responses to Language Varieties, 191–218.

https://doi.org/10.1075/impact.39.09gro

Hagendoom, L., & Hraba, J. (1987). Social distance toward Holland's minorities: Discrimination against and among ethnic outgroups. Ethnic and Racial

Studies, 10(3), 317-333.

Hendriks, B., van Meurs, F., & de Groot, E. (2017). The effects of degrees of Dutch accentedness in ELF and in French, German and Spanish. International Journal

of Applied Linguistics, 27(1), 44-66.

Hendriks, B., van Meurs, F., & Reimer, A. K. (2018). The evaluation of lecturers' nonnative-accented English: Dutch and German students' evaluations of

different degrees of Dutch-accented and German-accented English of lecturers in higher education. Journal of English for Academic Purposes, 34, 28-45.

Hosoda, M., & Stone-Romero, E. (2010). The effects of foreign accents on

employment-related decisions. Journal of Managerial Psychology, 25(2), 113-132.

(29)

Hosoda, M., Nguyen, L. T., & Stone-Romero, E. F. (2012). The effect of Hispanic accents on employment decisions. Journal of Managerial Psychology, 27(4), 347-364.

Kraut, R., & Wulff, S. (2013). Foreign-accented speech perception ratings: A multifactorial case study. Journal of Multilingual and Multicultural

Development, 34(3), 249-263.

Krings, F., & Olivares, J. (2007). At the doorstep to employment: Discrimination against immigrants as a function of applicant ethnicity, job type, and raters’ prejudice. International Journal of Psychology, 42(6), 406–417.

Lemhöfer, K., & Broersma, M. (2012). Introducing LexTALE: a quick and valid lexical test for advanced learners of English. Behavior Research Methods, 44(2), 325-343.

Lippi-Green, R. (2012). English with an accent: Language, ideology and discrimination

in the United States. Routledge.

Lindemann, S. (2003). Koreans, Chinese or Indians? Attitudes and ideologies about non‐native English speakers in the United States. Journal of

sociolinguistics, 7(3), 348-364.

Nejjari, W., Gerritsen, M., Van der Haagen, M., & Korzilius, H. (2012). Responses to Dutch‐accented English. World Englishes, 31(2), 248-267.

Piekkari, R., Vaara, E., Tienari, J., & Säntti, R. (2005). Integration or disintegration? Human resource implications of the common corporate language decision in a cross-border merger. International Journal of Human Resource Management,

16(3), 330–344.

Purkiss, S. L. S., Perrewé, P. L., Gillespie, T. L., Mayes, B. T., & Ferris, G. R. (2006). Implicit sources of bias in employment interview judgments and

decisions. Organizational Behavior and Human Decision Processes, 101(2), 152-167.

Ramirez, R. R. (2004). We the people: Hispanics in the United States. Washington, DC: US Department of Commerce, Economic and Statistics Administration, US Census Bureau.

Russo, M., Islam, G., & Koyuncu, B. (2017). Non-native accents and stigma: How self- fulfilling prophesies can affect career outcomes. Human Resource Management

Review, 27(3), 507-520.

(30)

of accentedness in the speech of Spanish-English bilinguals. Language and

Speech, 3, 267-273. https://doi.org/10.1177/002383097702000308.

Swim, J., Borgida, E., Maruyama, G., & Myers, D. G. (1989). Joan McKay versus John McKay: Do gender stereotypes bias evaluations? Psycho- logical Bulletin, 105, 409–429. doi:10.1037/0033-2909.105.3.409

Tajfel, H. (2010). Social identity and intergroup relations. Cambridge: Cambridge University Press.

(31)

Appendices

Appendix 1 – Script for the participants and consent forms Script

Well, I’d like to start by telling you something about my career. After I finished high school, I directly went to university to study Communication and Information Studies. I graduated in three years. The programme included an internship in a large organisation. I learnt a lot about marketing and other aspects of communication. I want to get more experience, so I am looking for a job.

A little bit about myself... I can work well in teams and by myself. I have a great sense of responsibility and I always want to learn more. I think I am a team player and I can work with everyone.

If I had to describe myself in three words, I would say “responsible, open-minded, happy”. I am a quick learner and open for everything. I push myself to the limits and I like to get to know other cultures. I like to make others enthusiastic and I think outside the box. In that sense, you could say I am creative as well.

That is why I think I am a good candidate for the position of junior communication analyst in your organization.

(32)

(33)

CONSENT FORM

Title of the research study: Effects of non-native accents on employability Researcher responsible: Myléne van de Wouw

Statement of participant

The aim of the research study has been outlined to me. I was given the opportunity to ask questions regarding the research study. I participate voluntarily in the research study. I understand that I can stop at any point during the research study, should I wish to do so. I understand how the data of the research study will be stored and how they will be used. I consent to participating in the research study as described in the information document.

Permission for audio/video recordings

I give permission to (please check all that apply): Yes No

¨¨ have audio recordings made of me for this research study and save these recordings, following the applicable regulations of the Radboud University

¨¨ share the audio recordings with participants from Germany and The Netherlands other remarks: ………..

Name: Date of birth:

Signature: Date:

Statement of executive researcher

I declare that I have informed the above-mentioned person correctly about the research study and that I abide by the guidelines for research as stated in the protocol of the Ethics Assessment Committee Humanities.

Name: Mylène van de Wouw

(34)

CONSENT FORM

Title of the research study: Effects of non-native accents on employability Researcher responsible: Myléne van de Wouw

¨ ¨ have audio recordings made of me for this research study and save these recordings, following the applicable regulations of the Radboud University

¨ ¨ share the audio recordings with participants from Germany and The Netherlands

other remarks: ………..

Name: ………... Date of birth:……….... Signature: ... Date:………...

Name: Mylène van de Wouw

(35)

(36)

(37)

(38)

(39)

(40)

Consent forms British

CONSENT FORM Title of the research study: Effects of non-native accents on employability Researcher responsible: Manou Gomlich

I give permission to (please check all that apply):

Yes No

have audio recordings made of me for this research study and save these recordings, following the applicable regulations of the Radboud University

x share the audio recordings with participants from Germany and The Netherlands

o her remarks

Name: Viliyan Donchev Date of birth: 11-03-1996

Signature: Viliyan Donchev Date: 17-04-2020

Name: Manou Gomlich

(41)

CONSENT FORM

Title of the research study: Effects of non-native accents on employability Researcher responsible: Manou Gomlich

have audio recordings made of me for this research study and save these recordings, following the applicable regulations of the Radboud University

share the audio recordings with participants from Germany and The Netherlands

other remarks: ………..

Name: ………... Date of birth:……….... Signature: ... Date:………...

Name: Manou Gomlich

(42)

(43)

(44)

Appendix II – Pre-test questionnaire

Dear Participants, We would be grateful if you could help us select the most suitable speech fragments for an experiment. In the experiment, listeners will be asked to evaluate an applicant for a job with a high communication demands based on a speech sample spoken in English In this pretest, we would kindly like to ask you as language professionals to assess the accent strength and other speech characteristics of the speakers in the following speech samples. Thank you for your help!

All the 12 audio fragments are formed together with the following set of questions: 1. Accent strength

1.1 The speaker has a strong foreign accent

Completely disagree 1 2 3 4 5 6 7 Completely agree 1.2 the speaker sounds like a native speaker of English

2. Identification of country of origin 2.1 where do you think the speaker is from?

▼ Afghanistan (1) ... Zimbabwe (1357) 3. Voice characteristics

The speaker has a high-pitched voice

Completely disagree 1 2 3 4 5 6 7 Completely agree The speaker speaks fluently

Completely disagree 1 2 3 4 5 6 7 Completely agree The speaker sounds monotonous

Completely disagree 1 2 3 4 5 6 7 Completely agree The speaker has a pleasant voice

Completely disagree 1 2 3 4 5 6 7 Completely agree The speaker sounds natural

Completely disagree 1 2 3 4 5 6 7 Completely agree The speaker speaks slowly