We are conducting an online survey experiment that complements analyses of media coverage of domestic violence. The main goal of the survey experiment is to evaluate the effect of providing information about the prevalence of domestic violence on how people perceive and react to cases of domestic violence.
Experimental design
Participants will be randomized into either a treatment group or a control group.
Design details
Randomization method
The randomization will be done on Prolific, a website for conducting online surveys.
Randomization unit
The unit of randomization will be the individual subject.
Clusters
N/A
Observations
For this intervention, we plan to collect roughly 900 observations.
Arms
We plan to collect roughly 450 observations per treatment arm.
Primary outcomes
We are interested in the following outcomes:
1) Respondents’ assessments of severity of the situation in the vignette,
2) Allocation of responsibility within the couple for the conflict presented in the vignette,
3) Willingness to seek help when faced with situations like in the vignette.
4) Willingness to support anti-violence organizations.
Secondary outcomes
After exposure to the treatment (infographic or control), participants will be asked how they feel at that moment. For each of the following emotions, they will rate the intensity on a 5-point Likert scale:
Sad, Worried, Optimistic, Determined to act, Uncomfortable, Powerless, Angry
Scale: 1 = Not at all, 5 = Extremely
This outcome was not included in the pilot but is added in the final study to capture emotional mechanisms underlying treatment effects.
We will also collect baseline sociodemographic data on respondents (gender, education, employment, cohabitation with partner, political orientation, beliefs in gender norms, past donation behaviour, trust in the government), and information on prior knowledge of prevalence of domestic violence in France or knowledge of the control group message. Compared to the pilot we added one question on past donation behaviour and one on trust in the government. We will explore if responses change by sociodemographic characteristics, prior knowledge, past donation behaviour or emotional response.
Cited by
No citation data available for this registration yet.