University Education and Students' Willingness-To-Pay for Green Attributes of Prospective Jobs

Throughout the survey, participants will respond to ten stated-preference experimental questions, each presenting two job offers, A and B. Each pair of job offers will vary randomly in monthly salaries and in the values of ten attributes. Six of these ten attributes can take one of two values: green (eco-friendly) or brown (not eco-friendly). The remaining four attributes can take on two values referring to some dimensions of working conditions, more desirable and less desirable respectively.

In each stated preference question, participants will indicate their probabilistic choice between JOB OFFER A and JOB OFFER B. This approach enables precise measurement of respondents' Willingness-to-Pay (WTP) for job attributes, free from confounding factors related to equilibrium conditions in the job market.

The survey involves ten stated-preference experimental questions. Each of these questions displays a pair of hypothetical job offers, namely JOB OFFER A and JOB OFFER B. Each job offer is described by ten attributes characterizing both the occupation and the company providing it, along with the offered monthly salary. Of these ten attributes, six can take on two values: one brown and one green (brown = not environmentally sustainable; green = environmentally sustainable). Below is the list of these six non-wage attributes and their corresponding brown and green values:

Employee benefits: BROWN = Not specified | GREEN = Annual public transport pass
Certifications: BROWN = Not specified | GREEN = Company certified ISO 14001 (environmental standards management)
On the job training: BROWN = Not specified | GREEN = Training on environmental topics and standards
Goods and services produced by the company: BROWN = Not specified | GREEN = Eco-friendly
Energy sources used by the company: BROWN = Not specified | GREEN = 50% from renewable energy
Performance bonus: BROWN = Granted if profit goals are met | GREEN = Granted if sustainability goals are met

The remaining four non-monetary attributes refer to some of the most important dimensions of working condition, namely schedule flexibility, working from home, pace of work, and autonomy at work. Even these attributes can take on two values, as follows:

Work schedule: Schedule set by your manager | You can set your own schedule
Option to work from home: No | Yes
Pace of work: Fast-paced | Moderate
Autonomy at work: Your tasks and procedures are defined by your manager | You can manage your work autonomously

First of all, we define the baseline job, around which job attributes would vary. It is constituted by the brown values as regards the attributes characterized by the green-brown dualism, and by the less desirable values as regards the attributes referring to working conditions. The baseline salary is €1,586, which represents the average salary in Italy for graduates three years after obtaining a master’s degree.
In each stated-preference question, the two job offers would have identical attribute values except for two attributes, which are selected randomly. These two non-wage attributes would randomly take on one of the two potential values, without replacement (i.e., the two job offers would not share the same value for the changing attributes). All other attributes remain equal between JOB OFFER A and JOB OFFER B and are set at their brown or less attractive values depending on attributes’ category.
The offered monthly salary is always randomly assigned across JOB OFFER A and JOB OFFER B and computed as follows. Given the baseline wage w = €1,586, the hypothetical wages are computed as θA*w and θB*w, where both θA and θB are normally distributed. To ensure that the wage variation remains within a plausible range, both θA and θB are truncated to the interval [0.75, 1.25], thereby restricting the wage difference between the two monthly salaries to a maximum of ±25% relative to the baseline wage. To limit the number of cases when one job offer entirely dominates the other on all varying characteristics (i.e., when one job offer displays both the green/more desirable values in the two varying attributes and the higher salary), the wage is recalculated using the aforementioned formula. If one job remains dominant, the values of the two varying attributes are re-randomized. At this stage, the new selections are adopted without further modifications.
The two varying attributes and the offered monthly salary are highlighted in yellow for improved visibility.
Below the table displaying attributes and salaries of the two hypothetical job offers, each stated-preference experiment includes a multiple-choice question with four options to gather the respondents’ preference for JOB OFFER A or JOB OFFER B. Specifically, the options are:

I strongly prefer JOB OFFER A
I prefer JOB OFFER A
I prefer JOB OFFER B
I strongly prefer JOB OFFER B

2025-03-15

2025-04-15

The key outcome is the participants' Willingness-To-Pay (WTP) for each green attribute of hypothetical jobs.

To derive each individual's WTP for each green attribute, we first need to estimate, using a probabilistic model, two parameters in the indirect utility function: the individual's marginal utility of each green attribute and the individual's marginal utility of the wage.

The secondary outcome is the participants' Willingness-To-Pay (WTP) for each desirable attribute of hypothetical jobs.

To derive each individual's WTP for each desirable attribute, we first need to estimate, using a probabilistic model, two parameters in the indirect utility function: the individual's marginal utility of each desirable attribute and the individual's marginal utility of the wage.

In the first section of the survey, participants are asked to respond to a set of questions on demographic characteristics, including gender, age, ethnicity, region of residence, education level, employment status, unemployment status, and occupation. Also, for those attending or who have attended university courses, additional questions cover the academic year of enrolment, course name, and university attended.
In the second section, we elicit information on a range of behavioral traits, that is risk tolerance, time preferences, competitiveness, and social preferences –altruism, reciprocity, and trust– using experimentally validated survey questions.

Afterwards, the survey involves ten stated-preference experimental questions. Each of these questions displays a pair of hypothetical job offers, namely JOB OFFER A and JOB OFFER B. Each job offer is described by ten attributes characterizing both the occupation and the company providing it, along with the offered monthly salary. Of these ten attributes, six can take on two values: one brown and one green (brown = not environmentally sustainable; green = environmentally sustainable). The remaining four attributes can take on two values referring to some dimensions of working conditions, more desirable and less desirable respectively. The offered monthly salary is always randomly assigned across JOB OFFER A and JOB OFFER B.

In the first section of the survey, participants are asked to respond to a set of questions on demographic characteristics, including gender, age, ethnicity, region of residence, education level, employment status, unemployment status, and occupation. Also, for those attending or who have attended university courses, additional questions cover the academic year of enrolment, course name, and university attended.
In the second section, we elicit information on a range of behavioral traits, that is risk tolerance, time preferences, competitiveness, and social preferences –altruism, reciprocity, and trust– using experimentally validated survey questions.

Afterwards, the survey involves ten stated-preference experimental questions. Each of these questions displays a pair of hypothetical job offers, namely JOB OFFER A and JOB OFFER B. Each job offer is described by ten attributes characterizing both the occupation and the company providing it, along with the offered monthly salary. Of these ten attributes, six can take on two values: one brown and one green (brown = not environmentally sustainable; green = environmentally sustainable). Below is the list of these six non-wage attributes and their corresponding brown and green values:

Employee benefits: BROWN = Not specified | GREEN = Annual public transport pass
Certifications: BROWN = Not specified | GREEN = Company certified ISO 14001 (environmental standards management)
On the job training: BROWN = Not specified | GREEN = Training on environmental topics and standards
Goods and services produced by the company: BROWN = Not specified | GREEN = Eco-friendly
Energy sources used by the company: BROWN = Not specified | GREEN = 50% from renewable energy
Performance bonus: BROWN = Granted if profit goals are met | GREEN = Granted if sustainability goals are met

The remaining four nonmonetary attributes refer to some of the most important dimensions of working condition, namely schedule flexibility, working from home, pace of work, and autonomy at work. Even these attributes can take on two values, as follows:

Work schedule: Schedule set by your manager | You can set your own schedule
Option to work from home: No | Yes
Pace of work: Fast-paced | Moderate
Autonomy at work: Your tasks and procedures are defined by your manager | You can manage your work autonomously

First of all, we define the baseline job, around which job attributes would vary. It is constituted by the brown values as regards the attributes characterized by the green-brown dualism, and by the less desirable values as regards the attributes referring to working conditions. The baseline salary is €1,586, which represents the average salary in Italy for graduates three years after obtaining a master’s degree.
In each stated-preference experimental question, the two job offers would have identical attribute values except for two attributes, which are selected randomly. These two non-wage attributes would randomly take on one of the two potential values, without replacement (i.e., the two job offers would not share the same value for the changing attributes). All other attributes remain equal between JOB OFFER A and JOB OFFER B and are set at their brown or less attractive values depending on attributes’ category.
The offered monthly salary is always randomly assigned across JOB OFFER A and JOB OFFER B and computed as follows. Given the baseline wage w = €1,586, the hypothetical wages are computed as θA*w and θB*w, where both θA and θB are normally distributed. To ensure that the wage variation remains within a plausible range, both θA and θB are truncated to the interval [0.75, 1.25], thereby restricting the wage difference between the two monthly salaries to a maximum of ±25% relative to the baseline wage. To limit the number of cases when one job offer entirely dominates the other on all varying characteristics (i.e., when one job offer displays both the green/more desirable values in the two varying attributes and the higher salary), the wage is recalculated using the aforementioned formula. If one job remains dominant, the values of the two varying attributes are re-randomized. At this stage, the new selections are adopted without further modifications.
The two varying attributes and the offered monthly salary are highlighted in yellow for improved visibility.
Below the table displaying attributes and salaries of the two hypothetical job offers, each stated-preference experiment includes a multiple-choice question with four options to gather the respondents’ preference for JOB OFFER A or JOB OFFER B. Specifically, the options are:

I strongly prefer JOB OFFER A
I prefer JOB OFFER A
I prefer JOB OFFER B
I strongly prefer JOB OFFER B

For each of the ten experiment questions, the randomization of the salaries offered and the values of the changing attributes is ensured through a programming language (JavaScript).

Individual.

No

There is no cluster.

The sample consists of 2,000 individuals aged between 18 and 30 residing in all Italian regions. Each individual is expected to respond to ten stated preference questions.

The treatment is the same for all the individuals.