Evidence and Learning

Evidence-aware and honest about what each state engagement must validate.

GI is grounded in adaptive problem-solving, institutional memory, service-delivery accountability and knowledge brokering. Each state engagement is designed to test what can be made durable inside government systems.

Partner on evaluation Read evidence note

Starting assumptions

Implementation and coordination capacity shape whether policy intent becomes public value.
Embedded actors generate practical knowledge that does not always travel.
Adaptive problem-solving needs diagnosis, experimentation, review and learning capture.
Knowledge infrastructure should be curated and honest about conditions for adoption.

What a state engagement should validate

Whether embedded governance talent can anchor structured Action Labs across districts or departments.
Whether review cadence improves continuity and adoption.
Whether curated pathways reduce repeated diagnosis and reinvention.
Whether GovCap products can make field learning reusable for new actors.

Evidence areas

The model draws from several bodies of public-systems learning.

The website should be clear about the intellectual basis while staying humble about claims that require field validation.

State capability

Implementation capacity, cooperation and accountability shape whether policy intent becomes service delivery and public value.

Adaptive problem-solving

Problem-driven, iterative and locally fitted work is better suited to complex public problems than one-size-fits-all replication.

Institutional memory

Useful evidence must be captured, organised and made available in forms that public actors can actually use during decisions.

Frontline diagnosis

Field knowledge and citizen-facing realities matter when they are connected to response capacity, not treated as consultation alone.

Knowledge brokering

Intermediaries can help translate evidence, connect communities and make practical knowledge travel across contexts.

Capability building

Training is stronger when it is tied to live problems, role expectations, routines and institutional use, not only classroom delivery.

Learning products

The output is not only a report. It is reusable public problem-solving knowledge.

Evidence notes, playbooks, failure notes, solution pathways, learning capsules and evaluation questions should all feed the Collective and GovCap.

Reference base

Public references support the design assumptions, not untested outcome claims.

The starter evidence inventory keeps a small set of public references connected to the website while the GI-specific evidence base is still being built.

Governance and service delivery

World Bank reports on governance and service delivery support the focus on implementation, accountability and public-system incentives.

WDR 2017
WDR 2004

Adaptive problem-solving

PDIA and Doing Development Differently support problem-driven, iterative and locally fitted approaches to complex public problems.

PDIA overview
DDD Manifesto

Evidence use

Evidence-use networks and knowledge-brokering practice support the need to make evidence usable, timely and connected to decision routines.

Alliance for Useful Evidence

Evaluation posture

GI should not overstate proof before a state engagement is operating. The evaluation posture is to define claims, gather evidence during implementation and publish learning only when it has passed review.

Measure change

Track whether Action Labs show credible progress on named problems, including what changed, what did not and why.

Measure durability

Look for routines, counterpart ownership and reusable knowledge that remain useful beyond a single project cycle.