Table 1 Pillars of the Metaketa Initiative.

Challenges for
cumulative learning
The Metaketa approach
1. Confounding in observational
research
1. Randomized controlled trials
(RCTs)
2. Limited external validity
of single RCTs
2. Multiple studies in diverse
contexts
3. Heterogeneous, scattered findings3. Meta-analysis with overall finding
4. Diversity of interventions4. Coordination on common arm
intervention
5. Noncomparable measurement
that impedes aggregation
5. Harmonized measurement of
inputs, outcomes, and controls
6. Researcher incentives for
innovation over replication
6. Study-specific interventions
preserve innovation and allow
analysis of comparative
effectiveness
7. Private data7. Open data and replication code
8. Errors in data or code8. Third-party data analysis
9. Fishing (data mining,
specification searching, and
failure to account for multiple
hypothesis tests)
9. Preanalysis plans with limited
number of specified
hypotheses
10. Publication bias10. Publication of all registered
analyses