Workplace Hostility

We plan to conduct a survey experiment to currently enrolled students and past students from various departments of a Canadian public university to evaluate the impact of hostile environment on career choices.
The recruitment of students and alumni will take place by email (university emails) and using various newsletters.
Participants will be presented with thirteen job scenarios with varying job characteristics and will be asked for each scenario to indicate their preferred choice. They will also answer a short questionnaire. The survey should not take more than 15-20 minutes. Participants will receive a 5 CAD payment for completing the full survey. In addition, they will be able to enter a lottery and 30 participants chosen at random will earn a 250 CAD gift card.

2023-07-05

2023-10-06

Binary variable indicating the subjects' choice between pairs of hypothetical job offers for 13 scenarios.

Our primary outcome variable is respondents’ willingness-to-pay for non-wage amenities in particular the WTP to avoid hostile attributes of the workplace. For each choice pair t, we will look at a binary variable Y = 1 if the respondent chooses job A over job B, = 0 otherwise. We will investigate heterogeneity by gender, race, parental education, majors, seniority, work location, and amount of team work.

Binary variable indicating the subjects' choice between hypothetical job offers.

Our outcome variable is respondents’ willingness-to-pay for non-wage amenities in particular the WTP to avoid hostile attributes of the workplace. For each choice pair t, we will look at a binary variable Y = 1 if the respondent chooses job A over job B, = 0 otherwise. We will investigate heterogeneity by participants' past experience of hostile environments during studies and at workplace, and participants' perceived risk of experiencing hostility in the next two years at workplace. For experience of hostility during their studies and at work, we will use dichotomous variables with 5 values (“Never”, “Rarely”, “Sometimes”, “Often”, “Always”). For perceived risk, we will ask the percent chance that they will experience hostility in the next two years and will use a continuous variable between 0 and 100. For scenario 12 and 13, we ask participants to predict the percentage of respondents in this study who will choose Job A over Job B, and will use a continuous variable between 0 and 100.

In the first part of the survey, respondents will be asked about their current situation. For students currently enrolled, this will include year of program, their current GPA and intended major. For alumni, it will include job title, industry, current earnings, and hours worked.
We will collect information on sociodemographic characteristics such as gender, age, race, parental education, and their preferred industries. The survey will contain a series of attention checks to control for the quality of responses.
After the survey, each respondent will see a series of thirteen hypothetical scenarios in which they will have to select between two jobs. Each scenario will present variation in compensation (hourly wage), and other non-monetary attributes. Respondents will be instructed that the jobs vary only in the listed characteristics and are otherwise identical. We will then ask three question about job search behaviors.

Not available

Randomization will be done by a computer.

The survey contains across and within subject randomization. We randomly assign whether the workplace location and amount of group-work varies for the participant. That is, either, the job add will vary on three characteristics (workplace hostility, workplace location, amount of group-work) or it will only vary on one characteristics (workplace hostility). This design choice has been made to decrease the amount of noise we may obtain by always varying all characteristics since the main focus is on workplace hostility. The following factor vary within-participant. We have a total of 30 scenarios. The first eleven scenarios which the participants is presented with are randomly selected among 28 of those 30 scenarios. That is, the scenario and the order of those eleven scenarios are randomized at the individual level. Note that every participant sees the other two out of 28 scenarios. Those two scenarios are shown to every participant as scenario 12 and 13 out of the total 13 scenarios they are presented with. The order of these two scenarios is fixed. That is, every participants sees the same two scenarios as scenario 12 and scenario 13. What is randomized for each job is the payment amount. The process is as follows. First, we use a benchmark annual salary for each participant. By default, the benchmark annual salary is equal to the median annual full-time salary for workers employed with the relevant major, aged between 25 and 64 with a bachelor’s degree or higher. For alumni and students who indicated that they have already accepted a job, we are able to update the benchmark annual salary by either using the median annual full-time salary for workers employed with their relevant major in their relevant sector, aged between 25 and 64 with a bachelor’s degree or higher *if* they participant provided us with the industry they work in. If the participant shares with us their current salary (or salary of their job which they have accepted but may not have started), their current salary will be used as their benchmark annual salary. Should the participant indicate that their annual salary is below CAD 10,000, we use the initial default salary. This design choice is based on the assumption that this individual may reference the wage of a student job or similar. We compute the participant’s benchmark annual salary after the participant completed the pre-questionnaire which contains industry and wage-related questions. To construct the job-specific wage, we follow Maestas et al. (2023). That is, we will use the benchmark annual salary and for each job randomly varied the wage to lie between 0.75 and 1.25 times the benchmark annual salary. We do this for each of the total 26 jobs (13 scenarios of which each scenario consists of two jobs).

No

We will invite about 19,000 subjects to participate in the survey experiment. We hope to be able to collect data from at least 3,000 subjects. Each subject will make 13 choices between pairs of fictitious job offers.

We will invite about 19,000 subjects to participate in the survey experiment. We hope to be able to collect data from at least 3,000 subjects. Each subject will make 13 choices between pairs of fictitious job offers.

There are no treatment arms in our design, but the job offers' characteristics vary randomly in each individual choice experiment.