Declaring and Diagnosing Research Designs.

Area of Science:

• Research methodology
• Scientific communication

Background:

• Selecting and communicating high-quality research designs presents significant challenges for researchers.
• Clear articulation of research designs is crucial for study reproducibility and validity.

Purpose of the Study:

• To introduce a formal framework for declaring the analytically relevant features of research designs.
• To provide a method applicable to qualitative, quantitative, and mixed-methods research for enhanced design transparency.

Main Methods:

• The proposed framework involves defining four key components: a model of the world (M), an inquiry (I), a data strategy (D), and an answer strategy (A).
• Declaration of these design features in code enables the use of Monte Carlo simulations for diagnosing properties like statistical power and bias.
• The approach supports both ex ante (for design improvement and preregistration) and ex post (for description and reanalysis) declarations.

Main Results:

• The formal declaration of research designs allows for the diagnosis of crucial properties such as power, bias, and accuracy of causal inferences.
• Ex ante declarations aid in improving study designs, facilitating preregistration, and reconciling intended versus actual analyses.
• Ex post declarations enhance the description, sharing, reanalysis, and critique of existing research designs.

Conclusions:

• The framework provides a demonstrably complete method for declaring research designs, applicable across diverse methodologies.
• Open-source software, DeclareDesign, is available to implement this approach, promoting standardized and transparent research practices.
• This formal declaration process improves research quality, communication, and the overall scientific process.