Eliciting Social Norms: A Survey Experiment on Descriptive and Normative Beliefs

This study employs a vignette-based survey experiment to examine the measurement of social norms related to grit and ambition in the labor market. The core objective is to compare the elicitation of normative expectations (what ought to be done) with empirical expectations (what is commonly done or believed) and assess how these differences influence perceptions of gender norms.
The intervention consists of randomly assigning participants to different framing conditions when eliciting their beliefs about social norms:
Normative frame – Participants report what they believe ought to be done in a given scenario.
Empirical frame – Participants report what they believe others commonly do or think in a given scenario.
Participants are presented with hypothetical workplace scenarios involving a central character whose gender is randomly varied. They are asked to evaluate the character’s career-related choices, allowing us to test: i)Whether normative expectations and empirical expectations systematically differ in the context of gender norms related to grit and ambition; ii)Whether misperceptions of social norms (i.e., the difference between second-order beliefs and actual first-order beliefs) vary depending on whether normative or empirical expectations are elicited. iii)How these differences contribute to gender gaps in career ambition and perseverance, particularly in how men and women perceive and internalize workplace expectations.
In the second wave of the study, participants will be exposed to informational feedback based on the responses collected in the first wave. This feedback will highlight either the actual normative expectations or empirical expectations reported by their peers, allowing us to examine whether social norm misperceptions can be corrected and whether this correction works differently for normative and empirical expctetations
The survey will be administered through Computer-Assisted Web Interviews (CAWI), ensuring a randomized and controlled exposure to different elicitation frames.

2025-02-09

2025-04-30

The key outcome variables of interest in this experiment focus on measuring perceived social norms related to grit and ambition in the labor market, as well as the extent of misperceptions in these norms. Specifically, we will assess:

1. First-Order Beliefs (Individual Views)
Self-reported beliefs about what ought to be done (normative expectations) vs. what is commonly done or believed (empirical expectations).
1.2. Differences in first-order beliefs based on the framing condition (normative vs. empirical).

2. Second-Order Beliefs (Perceived Beliefs of Others)
Perceptions of what others believe regarding grit and ambition in career choices.
2.1 Differences in second-order beliefs across framing conditions.

3. Misperceptions of Social Norms
Defined as the gap between second-order beliefs (perceptions of others’ beliefs) and the actual distribution of first-order beliefs (self-reported beliefs of others).
3.1 Whether misperceptions vary systematically depending on whether normative or empirical expectations are elicited.

4. Gender Differences in Norm Perception and Misperception
Systematic differences between men and women in first- and second-order beliefs.
4.1 Whether gender gaps in ambition and perseverance are reflected in differences in perceived norms.

5. Effect of Informational Feedback on Belief Updating (Second Wave)
Whether exposure to peer-reported normative vs. empirical expectations leads to changes in first- or second-order beliefs.
Whether belief updating differs based on framing (normative vs. empirical) and whether misperceptions are corrected differently under the two conditions.

1. Job Market Expectations and Preferences
Expected future earnings and career trajectories.
Preferences for job security, work-life balance, flexibility, and remote work.
Willingness to negotiate salaries, request promotions, or take on leadership roles.
2. Social Norm Feedback and Belief Updating
Magnitude of belief updating in response to social norm feedback (e.g., whether exposure to others’ first-order beliefs leads to stronger belief revisions among certain subgroups).
Whether individuals resist or adopt new social norm information based on pre-existing attitudes.
3. Risk Preferences and Decision-Making
Individual risk tolerance, competitiveness, and confidence in career-related decisions and their interaction with the first- and second-order beliefs.
4. Geographic and Job Mobility
Willingness to relocate for job opportunities, particularly in relation to perceived safety concerns.
Gender differences in choosing job security vs. higher earning potential.
5. Psychological and Behavioral Traits
Personality traits (TIPI), patience, and competitiveness, which may mediate career choices and perception of the norm.
Self-confidence in career decisions and perceived control over professional success.
6. Impact of Experimental Treatments
Changes in beliefs and career expectations between Wave 1 and Wave 2, due to exposure to social norm feedback.
Effects of different framing strategies (descriptive vs. normative) in shaping gender-related perceptions.
Interaction between educational background, gender, and exposure to role models in shifting career expectations.

This study employs a two-wave vignette-based survey experiment on a representative sample of the Italian population aged 19-29, stratified by gender, education level (high school vs. university graduates), geographic region, and metropolitan vs. non-metropolitan area. The experiment is designed to analyze how social norms and personal values shape early career decisions, focusing on the distinction between normative expectations (what ought to be done) and empirical expectations (what is commonly done or believed).

Wave 1: Baseline Elicitation of Norms and Beliefs
In the first wave, participants are presented with two career decision vignettes, with the gender of the protagonist randomized as the key experimental variation:
Ambition Vignette – The protagonist must choose between a prestigious long-term job opportunity abroad (potential for career advancement) and a less ambitious but stable local position. This scenario captures perceptions of ambition and career commitment and examines how these judgments differ by gender.
Grit Vignette – The protagonist chooses between a high-risk, high-reward career path and a low-risk, stable career path, allowing us to assess perceptions of perseverance and resilience in career decisions and potential gender biases in these perceptions.
The primary treatment variation in this study is the framing of belief elicitation, which systematically varies between:
Normative Frame ("Should") – Participants are asked what ought to be done in the given scenario (normative expectations).
Empirical Frame ("Would") – Participants are asked what they believe others would do in the given scenario (empirical expectations).
For each vignette, we elicit:

First-Order Beliefs – Participants' own views about what ought to be done (normative) or what they believe others would do (empirical).
Second-Order Beliefs – Participants' perceptions of what others believe ought to be done (normative) or what others commonly do (empirical).
By randomizing both the gender of the protagonist and the framing of belief elicitation (should vs. would), we can test:

Whether normative expectations and empirical expectations systematically differ.
Whether gendered perceptions of ambition and grit influence these differences.
The extent to which individuals misperceive social norms based on how beliefs are elicited.
At the end of the survey, one of the two vignettes is randomly selected, and 10% of respondents receive an incentive based on the accuracy of their second-order beliefs, encouraging engagement and truthful reporting.

Wave 2: Norm Updating and Informational Treatment
The second wave follows the same structure as Wave 1, but introduces an informational feedback intervention before participants respond to the vignettes again. Specifically, participants are shown:

Feedback on Normative Expectations – Aggregate responses from Wave 1 about what participants believe ought to be done in the given scenarios.
Feedback on Empirical Expectations – Aggregate responses from Wave 1 about what people actually reported doing or believing in the given scenarios.
This allows us to test:

Whether exposure to social information shifts beliefs, and whether the shift differs between normative and empirical expectations.
Whether perceptions of gendered career norms can be altered through informational interventions.
Whether individuals update their beliefs differently depending on whether they initially responded to a normative ("should") or empirical ("would") framing.
By comparing responses across the two waves, we can assess the role of informational interventions in correcting norm misperceptions and examine whether social norms systematically reinforce or mitigate gender gaps in ambition and perseverance.

Not available

The random assignment of participants to the different experimental conditions is handled by the survey administration company by a software that assigns the participants randomly to one of the conditions. Likewise, the company will randomly select a subset of participants eligible for an additional payment which will be based on the accuracy of their second-order belief estimates.

The survey administration company is responsible for the random assignment of participants to the different experimental conditions at the individual level. This approach ensures that the assignment of experimental conditions remains representative of the overall sample in terms of key stratification

No

NA

We plan to collect data from 5,000 individuals, split equally across two waves (2,500 per wave). Each wave will be independently representative based on gender, age groups (19-24 and 25-29 years), education level (high school graduates vs. university graduates), geographic region, and metropolitan vs. non-metropolitan area to ensure consistency in sample composition across both waves.

In each vignette, we implement four treatment variations, defined by the gender of the main character (male/female) and the frame used in the elicitation of first- and second-order beliefs. This results in 625 observations per wave and treatment.

This study aims to detect meaningful differences in the main outcomes using a two-sample t-test. Given a total sample size of 1,250 observations (625 per group), a significance level of 0.05, and assuming a common standard deviation of 0.8, the minimum detectable effect (MDE) is 0.1269 units at 80% power. This calculation is made for each wave within each gender variation.