Australia made history in late 2025 as the first nation to ban social media access for children under 16. As governments around the world watch closely—with France, the United Kingdom, Denmark, Spain, Greece, Indonesia, and Mexico among the many countries considering similar restrictions—one question looms large: What will the impact be on families?

Researchers at The Kids Research Institute Australia and the University of Western Australia approached NORC to partner on a study capturing family experiences before and after the ban. We had about a month to get a survey into the field before the policy took effect, a challenge familiar to those working at the intersection of research and policy.

An easy option would have been to use an opt-in survey panel, including only people who self-select or volunteer to participate in surveys and are ready to respond immediately.  But policymakers need reliable, trustworthy data, not just opinions from whoever happened to volunteer to complete a survey.

There’s a difference between asking people who volunteer to take surveys what they think and ensuring you’re hearing from a representative cross-section of an entire population. Opt-in panels—also known as convenience panels—introduce a fundamental problem: the people who choose to respond may be very different from those who don’t.

For example, parents who are particularly concerned about social media—or those who have strong opinions either way—might be more likely to sign up for panels and complete surveys about this topic. This means you’re hearing disproportionately from certain types of families, not getting a balanced picture of how all Australian families are experiencing the policy.

Probability sampling solves this problem by giving everyone in the internet-connected population a chance of being selected. For the Australia Social Media Ban Family Survey, we randomly generated mobile phone numbers to create a representative sample of Australian households with children ages 9-16. We then sent text message invitations asking parents to complete an online survey. This approach allowed us to reach families beyond those already signed up for survey panels, ensuring our findings reflect the experiences of all types of Australian families, not just those who typically participate in surveys.

Using a probability sample ensured we could include and represent hard-to-reach demographic groups who might opt out of convenience panels but are critical to understanding how all kinds of families experience these restrictions.

We also decided to track the same families over time. This allows us to distinguish policy effects from normal developmental changes in the affected children and measure the true impact of the ban nationwide.

Despite the tight time constraint, we made important design choices that strengthened the study. We included parents and caretakers of 16-year-olds—who are not affected by the ban—which will allow us to compare how families with children under the age threshold experience the ban relative to those over it.

We also made a pragmatic choice not to survey children in Wave 1 of our survey because our tight deadline did not allow for all necessary ethical clearances. Personally, I would have loved to include youth voices from the start because children are the ones directly affected by the ban. For Wave 2, we will add telephone interviews with 12- to 16-year-olds, capturing children’s perspectives directly, something that’s rare in research on minors but critical for understanding how they’re coping and adapting to the policy.

As more nations consider age-based social media restrictions, our survey will generate data that inform whether these policies work as intended. A probability-based longitudinal study tracking the same families before and after policy implementation creates a rich evidence base. We’ll learn:

• Whether parenting practices and family dynamics change
• If children’s well-being—mental health, social connectedness, and exposure to online risks—improves, stays the same, or worsens
• Whether the ban has unintended consequences
• How effects vary across family types, socioeconomic groups, or regions
• Whether and how children are circumventing the ban

To this end, we are developing a modular survey instrument that includes a core set of indicators comparable across countries and add-on modules that can be customized for each countries’ needs and priorities.

For instance, the modular approach allows countries to explore different research questions aligned with their policy concerns, whether that’s examining links between social media use and educational attainment, investigating effects on specific dimensions of child well-being such as social connectedness, or addressing locally relevant policy priorities. This flexibility ensures that while core measures remain consistent for cross-national comparison, each country can pursue the substantive questions that matter most to their policymakers and communities.

We’re also developing an interactive dashboard to incorporate comparable data from multiple countries as their policies evolve. For example, a policymaker in France could see how their country’s implementation challenges, family experiences, and impacts on children’s well-being compare with those in Australia or elsewhere. The dashboard will provide invaluable visualizations of findings at national and subnational levels, providing insights into how different communities and demographic groups experience policy impacts.

This study opens doors to broader questions about young people’s digital lives. There is enormous potential to combine probability surveys with emerging methods such as data donation, in which participants can choose to share their social media usage data with researchers—voluntarily and with strong privacy protections in place.

Data donation enables researchers to study users’ time spent on different platforms, the content they view, and their exposure to advertising. Combining data on platform usage patterns with self-reported survey responses will give us unprecedented insight into how digital environments affect children.

The methodological infrastructure we’re building extends beyond social media policy evaluation. The same probability-based approach, longitudinal design, and causal inference frameworks can be applied to other digital policy questions—from screen time regulations and online safety interventions to uses of artificial intelligence and educational technology implementation. As digital policy challenges multiply, having adaptable, rigorous research tools becomes increasingly valuable.

• Probability sampling yields representative findings by giving everyone the chance to be selected into the survey. It is superior to non-probability samples such as opt-in panels.
• Random digit dialing via SMS enables rapid deployment while maintaining methodological rigor and the ability to reach hard-to-reach populations.
• Probability-based longitudinal designs tracking the same families over time enable causal inference by distinguishing policy effects from normal developmental changes.
• Strategic choices—such as including 16-year-olds as a comparison group—allow for causal inference analysis to inform future policies.
• Including both parent and youth perspectives provides a comprehensive understanding of how policies affect families.
• Modular, adaptable research tools enable cross-country comparisons while accommodating each country’s particular conditions, priorities, and needs.

• To measure policy impact, countries considering social media restrictions should collect baseline data before policy restrictions go into effect. Probability-based longitudinal designs with comparison groups enable researchers to attribute impacts to policies.
• As more nations experiment with age-based restrictions, standardized measurement approaches and cross-national data sharing will enable comparative learning. Interactive dashboards can make findings accessible and actionable for diverse stakeholders, from national policymakers to community leaders.
• Policymakers benefit most from research designs that balance speed with scientific rigor, generating credible, actionable evidence without sacrificing the representativeness needed for sound policy decisions.