Party Membership as a Signal in China’s Labor Market: A Resume Audit Experiment

We conduct a resume audit experiment using fictitious applications submitted to real internship vacancies posted on shixiseng.com. The experiment features a 2 × 2 within-job-post factorial design. We independently randomize two pieces of information contained in the applicant’s CV: (1) party membership (P), indicating whether the CV states that the applicant is a CCP member; and (2) a noncognitive-ability signal (S), indicating whether the CV includes a favorable summary of results from a Big Five personality assessment. The four experimental profiles are T1 (P = 0, S = 0), T2 (P = 1, S = 0), T3 (P = 0, S = 1), and T4 (P = 1, S = 1). Party membership is disclosed in the student-organization section of the CV. The noncognitive-ability signal is included in the self-assessment section and summarizes favorable Big Five traits. Other applicant characteristics are held constant or randomly assigned from predetermined sets of plausible résumé characteristics. All fictitious applicants are presented as third-year university students of the same age and with comparable academic performance. For each eligible job posting, one CV from each treatment cell is submitted, and the order of the four applications is randomized.

2022-04-09

2022-04-29

The primary outcome is a binary employer-callback indicator. An application is coded as receiving a callback if the employer responds favorably within two weeks of submission. Favorable responses may arrive by telephone, text message, email, or the platform’s messaging system. If no favorable response is received within two weeks after submission, the application is coded as receiving no callback. Research assistants will handle employer responses and, upon receiving a callback, will inform the employer that the fictitious applicant is no longer available.

1. Experimental Setting and Sampling
The field experiment takes place in Beijing from February to April 2022 using internship vacancies posted on shixiseng.com, a leading online platform for campus recruitment and internships in China. We select job postings that meet the following criteria: (1) the position is located in Beijing; (2) undergraduate students are eligible to apply; (3) when the same firm posts multiple eligible vacancies, only one posting is selected; and (4) the occupation is suitable for standardized applications, focusing on human resources, sales, and administrative assistance.
2. CV Construction and Treatment Design
Applicants on shixiseng.com are required to complete a standard CV template containing basic information such as name, gender, age, and education, together with optional sections including student organization experience, honors and awards, skills and languages, hobbies and interests, and self-assessment. We use these fields to construct a 2 × 2 within-job-post factorial design. Let P denote disclosure of party membership and S denote provision of a noncognitive-ability signal. The four CV types are:
T1: non-party member without a noncognitive-ability signal (P=0, S=0).
T2: party member without a noncognitive-ability signal (P=1, S=0).
T3: non-party member with a noncognitive-ability signal (P=0, S=1).
T4: party member with a noncognitive-ability signal (P=1, S=1).
Party membership is varied in the Student Associations and Organizations section of the CV. CVs assigned P=1 state that the applicant is a CCP member, while those assigned P=0 do not indicate party membership.
The noncognitive-ability signal is varied in the Self-assessment section of the CV. Research assistants take the NEO Personality Inventory-Revised (NEO-PI-R), which is based on the Big Five framework, and CVs assigned S=1 include a short summary of a favorable personality-test result. CVs assigned S=0 do not include this information.

To make the four profiles comparable on cognitive ability and other background characteristics, we use local mid-ranked public universities and randomly assign universities across CV types. GPA is set to “above average” for all applicants. All applicants are described as 21-year-old third-year college students, consistent with being available for an internship at least four days per week. Other resume content is drawn from predetermined banks based on real CVs and randomly assigned across profiles.

For each eligible job posting, four applications, one from each experimental cell, are submitted. The order of the four CV types within each job posting is randomized by computer. Applications are submitted every other day over an eight-day window, so that the same employer receives the four experimental applications sequentially rather than simultaneously.

Application/CV within a job-post block

No

500 job postings (4 applications per posting)

2,000 applications

Planned sample size is approximately 500 applications per treatment arm: 500 in T1 (P = 0, S = 0), 500 in T2 (P = 1, S = 0), 500 in T3 (P = 0, S = 1), and 500 in T4 (P = 1, S = 1). Each job posting receives one application from each treatment arm.