Experimental Design
Using an online survey, we will combine an information provision experiment with a factorial survey experiment (vignettes). We will invite employed men aged 20 to 49 who live with a partner in the same household and intend to have children in the future to participate in the study. Our design involves the following three steps (in chronological order).
The first step is a background survey. We will collect basic information on participants’ socio-demographic background, workplace characteristics, attitudes toward gender norms, involvement in domestic and childcare duties, and their beliefs about the level of support they would receive from significant others and employers regarding their own parental leave take-up. We will also elicit participants’ second-order beliefs about fathers taking leave in Poland by asking, first, what proportion of fathers actually take paternity leave, and second, what share of men support fathers taking parental leave.
The second step is the information experiment. We will randomly assign participants to either the treatment or control group. Participants in the treatment group will be shown information about how many eligible fathers take paternity leave and how many men support the idea of fathers taking parental leave. Participants in the control group will not see any information.
In the third step, participants will be presented with a series of vignettes describing parental leave scenarios that vary in duration, wage replacement rate, transferability, and the possibility of taking leave simultaneously with a partner. After each scenario, participants will be asked to indicate their willingness to take parental leave under these conditions. In the vignettes, we will consider four particular policy design features ( attributes) and the following options, which reflect the range within each feature observed in OECD countries:
• Leave duration: 2 weeks, 8 weeks, 16 weeks, 24 weeks
• Wage replacement rate: 25%, 50%, 75%, 100%
• Transferability to the partner: Yes, No
• Possibility of taking leave simultaneously with a partner: Yes, No
The design will yield 64 possible vignette combinations (4 × 4 × 2 × 2). Each participant will be randomly assigned four vignettes.
Our primary outcome of interest is leave-taking intention, measured as the self-reported willingness to take leave on a 0–100 scale. First, we will analyze how various leave features affect this willingness. Second, we will examine whether the willingness to take leave differs between participants assigned to the treatment group (who receive information on leave uptake and support) and those in the control group (who do not receive any information). Third, we will investigate interaction effects, i.e. whether the relevance of each of the four design features for the willingness to take leave is equally affected by the information experiment. Fourth, we will conduct heterogeneity analyses. We will investigate whether the information experiment affects the willingness to take parental leave differentially across subgroups, e.g., by degree of (mis)perception, own leave-taking experience, expected support, by significant others, socio-demographic characteristics, and attitudes toward gender roles. Fifth, we will explore potential underlying mechanisms of the belief updating process.