Preference for the Workplace and Youth Unemployment

This study examines the impact of providing labor market and peer information on college students' job preferences, as well as their subsequent job search behavior and outcomes.

This study examines the impact of providing labor market and peer information on college students' job preferences, as well as their subsequent job search behavior and outcomes. Our experiment employs a 2x2 factorial design, with two types of intervention information measured in baseline surveys:
(1) information on the market prices of job attributes;
(2) information on peers' reservation prices for job attributes.

To reduce cognitive load, each student will receive information on only one job attribute. For each attribute, the sample for that attribute will be randomly divided into four groups: an active control group, a market information intervention group, a peer information intervention group, and a combined market and peer information intervention group. The intervention is personalized, with all students in these four groups receiving their individual reservation price for the job attribute. Additionally, we include a passive (blank) control group, where students receive no information.

2024-10-31

2024-11-15

There are two types of primary outcomes in this study:

1.Job preferences, including reservation prices for job attributes, which will be measured through an incentivized choice experiment.

2.Job search behavior and outcomes, including post-graduation employment status, as well as the attributes of both the current job and any jobs that were forgone during the job search process.

1.Fertility intention, including the percentage likelihood of having children and the preferred gender composition.

2.Social media usage, including the types and intensity of activities, as well as reactions to social media recommendation algorithms.

Details of the experimental design will remain undisclosed until the trial is completed.

This study focuses on three key job attributes: (1) job stability, measured by the probability of being fired over a one-year period (%); (2) working hours, measured by weekly working hours; and (3) promotion prospects, measured by the annual percentage increase in base salary (%). These three attributes are selected for two reasons: First, they are among the most widely studied job attributes in the literature, with substantial research highlighting their importance. Second, results in baseline surveys indicate that these attributes are of particular concern to the target population and could significantly influence their job search behavior and outcomes.

The study examines the impact of providing two types of information on college students' preferences for these three attributes and their subsequent job search behavior and outcomes: (1) information on the market prices of job attributes, and (2) information on peers' reservation prices for job attributes. These two types of information are derived from our baseline surveys. In June 2024, we surveyed graduates from the class of 2024 to collect information on their post-graduation employment status and the job attributes of their current jobs, which was used to estimate the market prices of job attributes. In September 2024, we surveyed students from the class of 2025, using an incentivized choice experiment to gather data on job preferences. This data was used to estimate peers' reservation prices for job attributes.

To reduce cognitive load, each student will receive information about only one job attribute. For each attribute, the sample will be randomly divided into four groups: an active control group, a market information intervention group, a peer information intervention group, and a combined market and peer information intervention group. The intervention is personalized, with all students in these four groups receiving their individual reservation price for the job attribute, as collected in the September baseline survey. This allows students receiving market or peer information to compare their own reservation price with either the market price or their peers' reservation price. The study's focus on three job attributes results in 12 experimental groups. Additionally, a blank control group, in which students receive no information (i.e., no individual reservation price), is included, bringing the total number of groups to 13.

The trial will be conducted at eight colleges in China. Of the eight colleges, six are higher vocational colleges and two are universities. Within each college, we randomly select one class per major. For majors with five or more classes, we randomly select two classes. A full-class survey will be administered to all selected classes. Using this sampling approach, we expect to survey approximately 10,000 students who will graduate in June 2025, representing about 40% of the total student population in the class of 2025 across these colleges.

We will conduct several rounds of surveys. First, in October 2024, the information intervention will be applied to the 13 groups, immediately followed by an incentivized choice experiment to collect updated data on students' job preferences. All students will complete the same version of the choice experiment used in the September survey. In June 2025 and beyond, as students graduate and enter the labor market, we will conduct follow-up surveys to gather data on their actual job search behavior and outcomes.

Based on data from our baseline survey in September, we will randomly assign all participating students into 13 groups. To account for differences across study fields, randomization will be stratified by study field. The sample consists of approximately 10,000 students from 279 classes across 8 schools, covering nearly 200 majors. We group these majors into study fields according to the classification system established by the Ministry of Education of China. In total, there are 27 study fields, with 17 from higher vocational colleges and 10 from universities (compared to a total of 19 in higher vocational colleges and 12 in universities).

Within each study field, we will generate a random integer between 1 and 13 using a uniform distribution, and students will be assigned to different experimental groups based on the number they receive. For students receiving both market and peer information, the order of information presentation will be further randomized: half will receive market information first, and the other half will receive peer information first.

Individual

Yes

The sample consists of approximately 10,000 students from 279 classes across 8 schools, representing 27 study fields.

The sample consists of approximately 10,000 students from 279 classes across 8 schools, representing 27 study fields.

The planned number of observations will be randomly and evenly distributed across 13 groups.