Title of Dataset:
Contextualised Accent Bias in Professional Hierarchies, Study 1 – Evaluations of University Graduate Job Candidates

Creator:
Yang Ye, University of Greenwich
Katie Reeves, University of Greenwich

Date of Data Collection:
2021

Geographic Coverage:
United Kingdom

⸻

Contact:
y.ye@greenwich.ac.uk

⸻

Description of the Study

This dataset contains responses from a study examining how listeners evaluate a university graduate job candidate when the candidate speaks in different English accents. The study investigates whether evaluations vary depending on whether the candidate speaks Received Pronunciation (RP) or Multicultural London English (MLE) during recorded interview responses.

Participants listened to two audio clips, completed general evaluation ratings, assessed the candidate’s future prospects across several UK job roles, and responded to questionnaires on prejudice control, perceived discrimination, diversity of social contact, and accent attitudes. Two attention check items were included.

The data were collected using an online Qualtrics survey. Audio recordings were voiced by professional actors. Ethical approval was obtained from the Psychology and Counselling Research Ethics Committee, University of Greenwich.

⸻

Dataset Structure

The dataset is provided as a UTF-8 encoded CSV file (study1_data.csv), containing one row per participant.
Variables correspond to:

• General candidate evaluations
• Job prospect likelihood (near-term 0–5 years; long-term 15–20 years)
• Job success likelihood
• MCPR (Motivation to Control Prejudiced Reactions) items + reverse-coded items
• Perceived discrimination items + reverse-coded items
• Diversity of contact items
• Accent attitude items
• Attention checks
• Composite evaluation variables
• Demographic variables

A complete variable list and item wording are provided in study1_codebook.csv.

⸻
Attention Check Variables

• Q43 – Degree subject attention check
• Q44 – Interview-question memory attention check
⸻

Derived Variables (Composite Scores)

• MCPR – Total score from MCPR scale (Q46_1–Q46_17), incorporating reverse-coded items
• P_Discrimination – Total score from perceived discrimination scale (Q48_1–Q48_8), incorporating reverse-coded items
• Diversity_of_Contact – Total score from Q50_1–Q50_5
• eval_overall – General evaluation composite score (Q24–Q28)
• eval_bartender
• eval_barmanager
• eval_clerk
• eval_govmanager
• eval_analyst
• eval_bankmanager
Each combines likelihood of obtaining (Q31_x) and likelihood of success (Q33_x).

⸻

Reverse-Coded Variables

Reverse-coded variables are prefixed with R_, including:

• R_Q46_2
• R_Q46_4
• R_Q46_5
• R_Q46_9
• R_Q46_16
• R_Q46_17
• R_Q48_2
• R_Q48_6
• R_Q48_7

These are used in composite scoring.

⸻

File List
	1.	study1_data.csv – UTF-8 CSV dataset
	2.	study1_codebook.csv – Complete ASCII-only codebook
	3.	study1_survey.pdf – Original Qualtrics survey instrument
	4.	README_study1.txt – Documentation

⸻

Methodology Summary

Participants were recruited online through the University of Greenwich participant pool (n = 106) and various social media platforms (n = 14). All participants were at least 18 years old and residing in the UK.

Procedure:
	1.	Participants completed a sound check and listened to two interview recordings (one “teamwork” prompt, one “problem-solving” prompt).
	2.	Participants rated the candidate on general impressions using 10-point Likert scales.
	3.	Participants evaluated the candidate’s likelihood of obtaining and succeeding in six job roles across hospitality, public sector administration, and banking.
	4.	Participants completed questionnaires on:
	•	MCPR (17 items)
	•	Perceived discrimination (8 items)
	•	Diversity of contact (5 items)
	•	Accent attitudes (6 items)
	5.	Participants answered two attention check questions and provided demographic information.

The survey took approximately 10–15 minutes to complete.

⸻


Encoding and Compatibility

• All data and codebook files use plain ASCII or UTF-8 encoding.
• Files can be opened in R, Stata, SPSS, Python, or Excel.
• No proprietary formatting is required to parse data.

⸻

Citation

Ye, Y., Withey, C., Sharma, D., & Levon, E. (2025). The Accent ceiling: Contextualised accent bias in professional hierarchies (Manuscript submitted for publication).
