x

We will perform our scheduled update at 10:30 AM EST on DEC 13, 2017, for 20 minutes.
Discrimination Against the Entrepreneurs and Ethnic Minorities: Field Experiment in Russia
Last registered on October 16, 2017

Pre-Trial

Trial Information
General Information
Title
Discrimination Against the Entrepreneurs and Ethnic Minorities: Field Experiment in Russia
RCT ID
AEARCTR-0001308
Initial registration date
June 06, 2016
Last updated
October 16, 2017 4:57 PM EDT
Location(s)
Region
Primary Investigator
Affiliation
University of Kassel
Other Primary Investigator(s)
PI Affiliation
University of Kassel
Additional Trial Information
Status
Completed
Start date
2017-03-20
End date
2017-08-31
Secondary IDs
Abstract
Entrepreneurs stay in business despite both lower initial earnings and slower growth of income than in paid employment (Hamilton et al., 2000; Moskowitz and Vissing-Jorgensen, 2002; Astebro et al., 2004). This puzzling behavior is commonly explained by nonpecuniary taste-based factors like preferences for autonomy or control (see for review Astebro et al., 2014). Though specific preferences, of course, play a role, the entrepreneurs can stay self-employed simply because they can not find a job since prospective employers could prefer candidates with corporate experience.

Some studies show discrimination of entrepreneurs at the labor market ( e.g. Failla et al., 2017). However, these evidence either correlation or inconclusive. Moreover, little is known why employees discriminate the entrepreneurs: is it taste-based discrimination (Becker, 1971) or statistical one (Phelps, 1972; Arrow, 1973)? We provide correspondence experiment (Bertrand and Mullainathan, 2004) in Russia to fill this gap Specifically, we try to understand (1) if entrepreneurial discrimination exists; (2) if it exists, is it statistical or taste based; (3) if providing a goal for applying can reduce the level of discrimination. In addition, we measure the level of ethnic discrimination.

We send 12.000 fictitious resumes in response to 3000 real job advertisement in Russia and measure the rate of call-backs as an indicator of interest in the applicant. To elicit discrimination towards entrepreneurs we send identical CV's, except that we randomly assign if a person was self-employed or worked for a company. To assess if the entrepreneurial discrimination is statistically based, we apply for the occupations with different skill-levels: High or low skill level job according to International Standard Classification of Occupations (2008).
We assess racial discrimination randomly assigning the Slavic and ethnic minorities sounding names to the resumes with equal quality e.g. Alexey Lebedev vs. Ansar Juraev.
External Link(s)
Registration Citation
Citation
Asanov, Igor and Maria Mavlikeeva. 2017. "Discrimination Against the Entrepreneurs and Ethnic Minorities: Field Experiment in Russia." AEA RCT Registry. October 16. https://www.socialscienceregistry.org/trials/1308/history/22368
Experimental Details
Interventions
Intervention(s)
We provide the correspondence experiment in Russia. Overall, we send 12.000 fi ctitious resumes in response to 3.000 real job advertisement in Russia and we measure the rate of callbacks and response on e-mails as an indicator of interest in the applicant.


Intervention Start Date
2017-03-20
Intervention End Date
2017-08-31
Primary Outcomes
Primary Outcomes (end points)
Response Rate: Call-back and e-mail answers
Primary Outcomes (explanation)
1. Call-back and e-mail answers
2. Call-back
3. E-mail answers
Secondary Outcomes
Secondary Outcomes (end points)
Secondary Outcomes (explanation)
Experimental Design
Experimental Design
To elicit discrimination towards entrepreneurs we send 8.000 CV's in response to 2.000 vacancies. We randomly assign if (1) a person was self-employed or worked for a company and (2) whether motivation section in CV is provided. To assess if the entrepreneurial discrimination is statistically based, we apply for the similar occupations with different skill-levels: High (Skill level 4) or low(skill level 2) skill level job according to International Standard Classi cation of Occupations (2008). Namely, we send resume on the next vacancies in Skill level 4: Finance Managers (1211), Advertising and Public Relations Managers (1222), Information and Communications Technology Services Managers (1330). In skill level 2: Accounting Associate Professionals (3313); Conference and Event Planners (3332); Information and Communications Technology Operations Technicians (3511).





To assess racial discrimination we send 4.000 CV's in the response of 1.000 vacancies.We randomly vary the Slavic and ethnic minorities sounding names for the resumes with equal quality e.g. Alexey Lebedev vs. Ansar Juraev. Also, we vary if the person has work experience (on average 1.5 year) or (s)he recently graduated. We choose a broad range of occupations in different industries with the highest demand and that requires or does not face-to-face interaction with clients. Thus, we chose next six occupations for this part of experiment: Advertising and Marketing Professionals (2431, skill level 4), Real Estate Agents and Property Managers (3334, skill level 3), General Office Clerks (4110, skill level 2), Answering Service Operator (4223, skill level 2), Receptionists (general) (4226, skill level 2), Messengers, Package Deliverers and Luggage Porters (9621, skill level 1).


Computer software for correspondence study.
We develop a computer program for this experiment. It contains three main functions:
1. Resumes generating
2. Mailing
3. E-mail Responses tracking



CV characteristics database and CV Generation. To create the resumes, we analyzed a large number of real CVs posted online. With the help of these CVs, we created a database for resumes' sections. We also made a database of names, surnames, and middle names based on most popular Slavic and ethnic minorities names.

All the resumes contain the following sections: Heading (name, mobile telephone, e-mail, birth date); Work experience; Education; Professional skills; Personal qualities; Advanced training, Language Skills. The program randomly chooses gender of the applicant and construct their full name from corresponding part of names database (according to the type of CV). Next, it takes randomly lines from every section (according to the type of CV) in order to create a resume. The formatting of the resumes is also chosen at random. All the characteristics of the resume stored in MySQL database with unique CV identifier number.


All this helps us to address the Heckman critique (Heckman, 1998) as suggested by Carlsson et al. (2014). Thus, it gives us the possibility to interpret the ratios as suggested by Neumark (2012) and to estimate wage differentials in lines with Lanning (2013).



Mailing. We apply within-subject design: For each vacancy, computer program sends four types of the resumes in random order.

All applications contain a cover letter, where candidates express an interest in the posted job advertisement. The exact wording of the cover letter is chosen from a set of cover letters at random.To make the application realistic and targeted to specific vacancy the program contains an electronic form that is filled for every vacancy. The form contained following lines: job vacancy, email of a recipient, web source and the name of contact person (if given). This information is used to generate the cover letter.




E-mail Responses tracking. The program automatically tracks the responses on e-mails that are unique for each resume.

Call-back tracking. The unique phone number is assigned to each treatment and job
skill level. We plan to track phone calls on each type of treatment and job skill level. That
allows us to have clear results with respect to the main treatments. To get the detailed
picture of the characteristics of resume, we plan to match the call-back and corresponding
CV.
Experimental Design Details
Randomization Method
Random number generation within computer mailing program.
Randomization Unit
Vacancy. We send different types of CV for the same vacancy in random order.
Was the treatment clustered?
Yes
Experiment Characteristics
Sample size: planned number of clusters
The sample size was determined before the study and adjusted after the pilot experiment:
"The sample size of 1200 resumes and the low callback rate of the pilot experiment cannot provide sufficient data for clear support or disproof of all the presented hypotheses...The experiment is currently running and in total 12, 000 resumes are to be sent out and the callbacks processed until July, 2017."(Maria Mavlikeeva, Master Thesis. Submitted: 10 of April 2017)

3000 vacancies (12 000 CV)

Sample size: planned number of observations
12 000 CV (aprrox. 1 000 responses: calls and e-mails)
Sample size (or number of clusters) by treatment arms
Discrimination against Entrepreneurs Part of Experiment

2000 cvs with entrepreneurial experience and motivation section
2000 cvs with entrepreneurial experience and no motivation section
2000 cvs with corporate experience and motivation section
2000 cvs with corporate experience and no motivation section

Discrimination against Ethnic Minorities

1000 cvs with slavic sounding names and and with working experience
1000 cvs with slavic sounding names and with no working experience
1000 cvs with non-slavic sounding names and with working experience
1000 cvs with non-slavic sounding names and with no working experience

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)
IRB
INSTITUTIONAL REVIEW BOARDS (IRBs)
IRB Name
IRB Approval Date
IRB Approval Number
Analysis Plan
Analysis Plan Documents
PAP_FINAL.pdf

MD5: 0398270ce63e06a49c41990072fb62e2

SHA1: d911b0cd70e150512c4c415a1b10bdae407eee77

Uploaded At: October 16, 2017

Post-Trial
Post Trial Information
Study Withdrawal
Intervention
Is the intervention completed?
No
Is data collection complete?
Data Publication
Data Publication
Is public data available?
No
Program Files
Program Files
Reports and Papers
Preliminary Reports
Relevant Papers