Four student projects on the gymnasium floor. Their hypotheses, materials, results, and the judge's ribbon — so you can vote for the best experiment.
💼 Can One Platform Do Everything?
By: LinkedIn · Microsoft Academy · Grade: Professional · Since 2003
🥇 1st Place — Most Complete
Hypothesis
If a single platform combines networking, job search, content, learning, and sales tools, then it will become the dominant professional network with 1 billion+ users across 200+ countries.
Hypothesis: CONFIRMED. 1B+ users, $15B+ revenue, 200+ countries. The experiment worked — one platform CAN do everything. But the content feed variable is contaminated with AI noise, and the premium cost variable keeps increasing. Further research needed on long-term content quality.
🏅 Judge's Notes
Most ambitious project at the fair. Impressive breadth of materials and strong quantitative results. Deducted points for declining content quality and escalating cost variables. Still the clear winner for overall completeness.
Areas for Improvement
→ Control the AI noise contaminating the content variable
→ Improve organic visibility for institutional test groups
→ Disclose data sharing with Microsoft parent lab
🔍 How Fast Can You Find a Job?
By: Indeed + Glassdoor · Recruit Academy · Grade: Employment · Since 2004
🥈 2nd Place — Fastest Results
Hypothesis
If a platform focuses exclusively on job search with zero social features, then it will achieve the fastest employment results with 350M+ monthly users. Glassdoor control group adds employer transparency.
Hypothesis: CONFIRMED. Fastest time-to-employment in the fair. Glassdoor adds the employer intel the control group needs. However, removing all social variables means zero networking capability. The experiment proves speed but sacrifices breadth.
🏅 Judge's Notes
Excellent focused methodology. Fastest measurable outcome in the employment category. Glassdoor control adds genuine rigour. Marked down for zero networking results and narrow experimental scope.
Areas for Improvement
→ Add networking variables for broader applicability
→ Verify Glassdoor control data freshness
→ Reduce sponsored variable interference
→ Extend experiment beyond single job-search cycle
→ Address data visibility to paying observers
🚀 Does Transparency Change Outcomes?
By: Wellfound · Indie Academy · Grade: Startup · Since 2010
🏆 Special Award — Most Innovative
Hypothesis
If salary and equity data are displayed openly on every listing, then 8M+ candidates will make better-informed startup career decisions across 150K+ startups.
Hypothesis: CONFIRMED (within scope). Salary and equity transparency scored 9.8 — highest single variable at the fair. But the experiment only works in startup conditions. Results go to zero in corporate test environments.
🏅 Judge's Notes
Most creative hypothesis. The transparency variable is genuinely novel — no other project measures it. Special Innovation Award for methodology. Limited by narrow test conditions and small sample size.
Areas for Improvement
→ Expand test conditions beyond startup environments
→ Increase sample size from 8M to broader population
→ Address early-stage startup data reliability
→ Develop networking variable beyond current baseline
→ Increase visibility at mainstream science fairs
🤝 Is In-Person Better Than Digital?
By: Meetup + Lunchclub · IRL Academy · Grade: Connection · Since 2002
🌟 People's Choice Award
Hypothesis
If professional networking is conducted in person rather than digitally, then 52M+ participants will form more authentic connections across 2M+ annual events. AI pairing (Lunchclub) as catalyst variable.
Hypothesis: CONFIRMED. Authenticity scored 9.9 — the single highest result at the entire fair. LinkedIn is the poster board. This project IS the science. But zero employment variables means this experiment can't answer the job question.
🏅 Judge's Notes
The project that made the judges look up from their clipboards. Authenticity score of 9.9 is the highest single measurement recorded today. People's Choice winner by a landslide. No employment data — but that wasn't the hypothesis.
Areas for Improvement
→ Add employment variables for broader application