evidoria

What works in Stanford

All catalogues · United States of America profile

3
evaluated practices
58/100
average quality score
0%
strong evidence, of 2 assessed

Quality scores are 0–100 rubric scores; “strong evidence” reflects study design; adoptions count documented replications elsewhere. Know more

Evidence snapshot for Stanford

Stanford's Office of Technology Licensing has generated $88m+ in annual licensing revenue and launched 28 new startups in FY2025, though only an estimated 20–25% of the 541 invention disclosures received are ever licensed, indicating highly concentrated returns from a small number of blockbuster patents. In educational technology, a randomised trial of GPT-4 in Stanford's Code in Place course with 5,831 learners found that offering the tool reduced average exam participation and attendance, though students who actually adopted it scored better on exams, with benefits concentrated among learners from low-HDI countries. In response to concerns about generative-AI-enabled cheating, Stanford voted in May 2026 to end 105 years of unproctored exams, expanding a pilot of equitable proctoring practices—though the evidence linking the policy shift directly to an AI-driven cheating surge rests on faculty testimony rather than published, AI-specific violation counts.

Drawn from: Stanford OTL — From Cohen-Boyer to Google, a Self-Funding University Tech-Transfer Engine · The GPT Surprise — Stanford's Code in Place Randomised Trial of GPT-4 Chat for 5,831 Learners in 146 Countries · Stanford Ends 105 Years of Unproctored Exams — A Governance Response to Generative-AI Cheating

Generated from this place's cited practices only. Always check the linked sources.

Strengths & weaknesses vs peers

Strengths

Evidence of impact / measured results +43 pts 100 here · 57 average

Transferability / demonstrated replication +40 pts 100 here · 60 average

Multi-stakeholder collaboration +26 pts 100 here · 74 average

Sustainability / continuity +22 pts 100 here · 78 average

Transparency, ethics & data protection +20 pts 67 here · 47 average

Scalability +16 pts 78 here · 62 average

Weaknesses

Teacher capability & pedagogy -21 pts 17 here · 37 average

Inclusiveness / equity -43 pts 0 here · 43 average

Score profile vs peers

Evidence of impact / measured results
100 · 57
Transferability / demonstrated replication
100 · 60
Multi-stakeholder collaboration
100 · 74
Sustainability / continuity
100 · 78
Transparency, ethics & data protection
67 · 47
Scalability
78 · 62
Evidence of learning impact
33 · 26
Innovation
67 · 62
Equity & inclusion
50 · 50
Teacher capability & pedagogy
17 · 37
Inclusiveness / equity
0 · 43

Coloured bar: Stanford. Grey bar: peer average across the corpus. Both on a common 0–100 scale.

Points compare this place's average on each dimension with the average across all evaluated places, on a common 0–100 scale. Know more

Top good practices here

Stanford OTL — From Cohen-Boyer to Google, a Self-Funding University Tech-Transfer Engine · 81

The GPT Surprise — Stanford's Code in Place Randomised Trial of GPT-4 Chat for 5,831 Learners in 146 Countries · 60

Stanford Ends 105 Years of Unproctored Exams — A Governance Response to Generative-AI Cheating · 33

Each number is the practice's overall 0–100 quality score against its catalogue's rubric. Know more

Recommended to adopt or replicate

Transferable practices from elsewhere that address this place's gaps.

Smart City Wien Framework Strategy · fit 82/100

  • Rated highly transferable
  • Addresses this place's weak dimensions

Ciudad Mujer — El Salvador's Integrated Women's Services Programme · fit 100/100

  • Rated highly transferable
  • Backed by strong evidence

Already adopted in 3 other places

SASA! Community Mobilisation Programme against Intimate Partner Violence — Uganda · fit 100/100

  • Rated highly transferable
  • Backed by strong evidence

Already adopted in 4 other places

Digital Green — Video-Mediated Agricultural Extension · fit 100/100

  • Backed by strong evidence

Ushahidi — Kenya's Crowdsourced Crisis-Mapping Platform · fit 82/100

  • Rated highly transferable
  • Addresses this place's weak dimensions

Fit estimates how well a practice matches this place, 0 (poor) to 100 (excellent); the list is ordered to favour practices that address this place's gaps. Know more