Aggregating evidence for a central theory from diverse studies using the generalized order-restricted information...

Area of Science:

• Social and behavioral sciences
• Statistical methodology
• Evidence synthesis

Background:

• Reproducibility crisis in social sciences necessitates new evidence aggregation methods.
• Current meta-analysis is limited by study design and context, leaving valuable data unexploited.
• Need for a tool that can aggregate evidence from heterogeneous studies.

Purpose of the Study:

• Introduce Generalized Order-Restricted Information Criterion (approximation) (GORIC[A]) evidence aggregation as an alternative to traditional meta-analysis.
• Develop a statistical tool to combine evidence across studies with varying designs and contexts.
• Quantify the overall relative support for a shared central theory.

Main Methods:

• Utilizes GORIC(A), an information criterion capable of evaluating equality and inequality/order restrictions.
• Applies GORIC(A) to individual studies and aggregates the resulting evidence across multiple studies.
• Validates the method through a simulation study assessing its performance with heterogeneous study designs.

Main Results:

• GORIC(A) evidence aggregation is not affected by study-design heterogeneity.
• The method successfully synthesizes evidence from a diverse set of studies.
• Demonstrates increased available information for theory investigation and enhanced robustness of results.

Conclusions:

• GORIC(A) evidence aggregation is a viable tool for evidence synthesis across diverse research.
• The method enhances the robustness and confidence in scientific findings by incorporating all relevant studies.
• Facilitates a more comprehensive understanding of central theories by leveraging a wider range of evidence.