Parental Leave and Teen Family Cohesion

ISEF Category: Behavioral and Social Sciences

Subcategory: Sociology and Anthropology  ·  Difficulty: Advanced  ·  Setup: Home Setup  ·  Time: 1 to 2 Months

The Hook

A country's leave policy can shape more than newborn care. It can also change how teens describe family life years later. You can test that idea with public OECD and PISA data. This turns a big policy question into a student-friendly data project.

What Is It?

Parental-leave policy generosity means how much paid time off parents can get, how much of their pay they keep, and who qualifies. Family-cohesion scores tell you how strongly teens feel that their family stays connected, supports one another, and spends time together. Think of the policy as the rules of the game and the cohesion score as the scoreboard.

This project asks whether countries with more generous leave rules also have higher teen reports of family cohesion. You are not proving that one policy causes a family to feel a certain way. You are testing whether the patterns move together across countries after you account for other country differences. Multilevel modeling helps here because students are nested inside countries, like players are nested inside teams.

Why This Is a Good Topic

This is a strong science fair topic because the data already exist, the question is clear, and the result can be measured with country and student numbers instead of opinions. It connects to a real policy issue that many people care about, including work-life balance, family support, and teen well-being. You can learn how to code variables, compare countries, and read a multilevel model without needing a wet lab.

Research Questions

• How does parental-leave generosity relate to average adolescent family-cohesion scores across countries?
• What is the effect of paid leave duration on family-cohesion scores after adjusting for GDP per capita?
• Does wage replacement rate predict family-cohesion scores more strongly than leave length?
• To what extent do parental-leave policy measures explain between-country differences in teen family cohesion?
• Which parental-leave indicator, total weeks, pay replacement, or eligibility rules, best predicts family-cohesion scores?
• Does the policy link stay similar after you split countries by region or income level?

Basic Materials

• Laptop or desktop computer with internet access.
• OECD family policy tables and PISA public-use data files.
• Spreadsheet software for cleaning and merging country-year data.
• A statistics notebook or document for tracking variables, code, and assumptions.
• Basic graphing software for checking trends and making quick charts.
• Access to a school library or open web sources for policy background.

Advanced Materials

• Laptop with R or Python installed.
• Institutional access to PISA microdata or extended country files, if available.
• OECD API access or downloadable country-year tables.
• A secure folder for raw data, code, and versioned outputs.
• University library access for methods papers and policy sources.
• Access to a statistics mentor or methods lab for feedback on multilevel modeling.

Software & Tools

• R: Fits multilevel models, reshapes country-year data, and makes publication-style charts.
• RStudio Desktop: Gives you a clean workspace for scripts, tables, and model output.
• Python: Helps merge OECD and PISA files and check missing values with pandas.
• JASP: Lets you test regression models with a simpler interface before you code.
• Google Sheets: Helps you screen source tables and keep a variable log.

Experiment Steps

1. Define the student-level outcome and the country-level policy measure you will compare.
2. Choose one PISA wave and one OECD policy table so your data come from the same policy window.
3. Build a country-year dataset and decide how you will handle missing values and outliers.
4. Fit a null multilevel model, then add parental-leave generosity and basic country controls.
5. Test alternate policy measures and subgroup splits to see whether the pattern stays stable.
6. Check the size, direction, and uncertainty of the policy effect, then turn the result into clear visuals.

Common Pitfalls

• Matching a country's leave policy to the wrong PISA year, which makes the policy timing and survey timing line up badly.
• Using the average student score without a multilevel model, which treats clustered country data like independent cases.
• Comparing countries before checking score scales and response formats, which can hide survey design differences.
• Collapsing paid leave, eligibility, and wage replacement into one vague variable, which makes the policy effect hard to interpret.
• Keeping only countries with complete data, which can change the sample in ways that favor richer countries.

What Makes This Competitive

A stronger entry does more than compare averages. You separate student-level patterns from country-level policy effects, then check whether the result survives alternate leave measures and control variables. You can raise the bar by reporting effect sizes, confidence intervals, and model fit, not just p-values. A clear policy story, plus careful handling of nested data, makes the project read like real social-science research.

Project Variations

• Swap family-cohesion scores for adolescent life satisfaction and test whether the pattern holds.
• Compare total parental leave with paternity-leave share to see which policy measure matters more.
• Restrict the sample to one region, such as Europe or East Asia, and test whether the cross-national pattern changes.

Learn More

• OECD Family Database: Find country-level parental-leave indicators and family policy tables on the OECD site.
• OECD PISA Data Explorer: Look up public PISA well-being indicators and country comparisons on the OECD site.
• NCES PISA resources: Read plain-language guides to PISA structure and survey variables on the National Center for Education Statistics site.
• NIH Office of Behavioral and Social Sciences Research: Search for methods pages on survey analysis, multilevel models, and social determinants of health.
• MIT OpenCourseWare: Search for free statistics and regression lectures that explain hierarchical modeling.