for the analysis (measurement variables) with the information in the data map describing the study design (research variables) creates original data sets that are ready for analysis, as shown in figure 6.1. Doing so is difficult, creative work, and it cannot be reproduced by someone who lacks access to the detailed records and explanations of how the data were interpreted and modified. The chapter stressed that code must be well organized and well documented to allow others to understand how research outputs were created and used to answer the research questions. The next chapter of this book provides a guide to assembling the raw findings into publishable work and describes methods for making data, code, documentation, and other research outputs accessible and reusable alongside the primary outputs.
FIGURE 6.1 Data analysis tasks and outputs Design Acquisition Processing Analysis Clean data
Treat missing observations and distribution patterns
Integrate data sources
Construction documentation
Analysis data dictionary
Merged data
Create analysis indicators
Analysis data
Conduct exploratory analysis
Include result in final output? Yes Format and export outputs
No
Analysis archive
Analysis code Raw outputs
Publication
Source: DIME (Development Impact Evaluation), World Bank. 148
DEVELOPMENT RESEARCH IN PRACTICE: THE DIME ANALYTICS DATA HANDBOOK