Full Library

The Method Positive Deviance (PD) is based on the observation that in every community or organization, there are a few individuals who achieve significantly better outcomes than their peers, despite having similar challenges and resources. These individuals are referred to as positive deviants, and adopting their solutions is what is referred to as the PD approach¹. The method described in this Handbook follows the same logic as the PD approach but uses pre-existing, non-traditional data sources instead of — or in conjunction with — traditional data sources. Non-traditional data in this context broadly refers to data that is digitally captured (e.g. mobile phone records and financial data), mediated (e.g. social media and online data), or observed (e.g. satellite imagery). The integration of such data to complement traditional data sources generally used in PD is what we refer to as Data Powered Positive Deviance² (DPPD). The digital data opportunity Recent developments in the availability of digital data provide an opportunity to look for positive deviants³ in new ways and in unprecedented geographical and on temporal scales. A number of studies⁴ have described the challenges related to the application of the PD approach in development. Given these challenges, there are obvious opportunities for innovation in PD and our particular interest here is in the innovative opportunities offered by non-traditional data, following the increasing “datafication” of development and the growing availability of big datasets in a variety of development sectors⁵. DPPD builds on this and expands our ability to extract value from non-traditional digital data while providing a systematic process for leveraging local know-how and the collective wisdom of communities. Data Powered Positive Deviance The DPPD method described in this Handbook emerged from a process of research and testing and follows the same stages as the PD approach. The difference is that DPPD integrates pre-existing, non-traditional data across the five stages, requiring a series of new and specific methods and practices that are not required in the PD approach. The first stage is also somewhat different because it not only defines the problem, but it also checks if it is suitable and feasible to use the DPPD method for the proposed project. Table 1 lists the five stages of the DPPD method. This Handbook dedicates a section to each stage. Stage 1 Assess problem-method fit Stage 2 Determine positive deviants Stage 3 Discover underlying factors Stage 4 Design and implement interventions Stage 5 Monitor and evaluate