Dark Mode
Racial Bias in Job Callbacks Experiment
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Explaining data derived from a field experiment conducted to study the influence of perceived race and gender on job application callback rates. The study involved submitting randomly generated resumes to job postings, where first names were strategically chosen to predominantly signal the applicant's race (Black or White) and gender. The design of the experiment, which randomly assigned race and gender attributes independently of resume quality, allows users to draw causal conclusions about labour market discrimination. The data provides insights into key resume attributes that impact whether an applicant receives a callback.
Columns
The dataset contains 30 columns detailing both the job advertisement characteristics and the experimental resume attributes:
job_ad_id: Unique identification for the advertisement.job_city: City where the job was located (either Chicago or Boston).job_industry: The job sector (e.g., other service, business and personal service).job_type: The nature of the role (e.g., secretary, retail sales).job_fed_contractor: Indicates if the employer is a federal contractor.job_equal_opp_employer: Indicates if the employer is an Equal Opportunity Employer.job_ownership: The type of company (e.g., private, nonprofit).job_req_any: Indicates if any requirements were listed for the job.job_req_communication: Indicates if communication skills were required.job_req_education: Indicates if some level of education was required.job_req_min_experience: The minimum experience amount required (56% are missing/not listed).job_req_computer: Indicates if computer skills were required.job_req_organization: Indicates if organization skills were required.job_req_school: The required level of education (e.g., none listed, some college).received_callback: Binary indicator (1 or 0) showing if a callback was received.firstname: The first name used on the resume (e.g., Tamika, Anne).race: The inferred race associated with the first name (White or Black).gender: The inferred gender associated with the first name (f or m).years_college: Years of college education listed (mean is 3.62 years).college_degree: Indicator for listing a college degree.honors: Indicator for listing academic honours.worked_during_school: Indicator for listing working while in school.years_experience: Years of professional experience listed (mean is 7.84 years).computer_skills: Indicator if computer skills were listed on the resume.special_skills: Indicator if any special skills were listed.volunteer: Indicator for listed volunteering experience.military: Indicator for listed military experience.employment_holes: Indicator for holes in employment history.has_email_address: Indicator for listing an email address.resume_quality: Resume quality classification (high or low, split evenly at 50% each).
Distribution
The data file, provided in CSV format (
resume.csv), contains 4870 valid records across 30 columns. The variables are primarily categorical or binary indicators, though some, such as years_college and years_experience, are numerical.Usage
Ideal applications include conducting analysis and prediction of job application callback rates. Users can model the causal linkage between inferred racial and gender attributes and hiring outcomes due to the experimental design. This data is suitable for investigating the existence and magnitude of discrimination in the labour market, particularly how race interacts with other factors like resume quality, experience, and educational attainment.
Coverage
The geographic coverage spans job postings monitored in Boston and Chicago. The time range of data collection was several months during 2001 and 2002. The demographic scope primarily focuses on the inferred race (Black or White) and gender (Female or Male) associated with applicants' first names, which were randomly assigned during the study.
License
CC0: Public Domain
Who Can Use It
- Economists and Social Scientists: For field experiments analysis, labour market studies, and measuring racial inequality.
- Policy Makers: To assess the effectiveness of anti-discrimination laws and inform future equal opportunity policies.
- Machine Learning Engineers: To build models predicting employment outcomes or to examine bias in algorithmic hiring tools.
Dataset Name Suggestions
- Racial Bias in Job Callbacks Experiment
- US Labour Market Discrimination Data (2001-2002)
- Field Experiment on Hiring Bias
- Resume Discrimination Study
Attributes
Original Data Source: Racial Bias in Job Callbacks Experiment
QUALITYLoading...