Evaluating NCA interventions through multivariate analysis with big data

This briefing note presents the key findings from a 2020i scoping study on the potential of combining Norwegian Church Aid (NCAii) data with external datasets for better evaluations, and what doing that would require.

Through its work, NCA produces a significant amount of data for monitoring and evaluation purposes. This includes geo-located, household-level survey data as well as project locations (e.g. wells and other water supply points,
schools). NCA also regularly conducts evaluations of a global nature (multi-country evaluations).

The evaluation type used by NCA depends on the evaluation’s overall purpose and questions, the nature of the intervention to be evaluated, the existing evidence and available resources.

One area of interest is to explore the potential of combining datasets – data produced by NCA together with open databases.

This includes:

• Traditional data sources (e.g. administrative data, survey data) made accessible with open data protocols, improving the possibilities for finding and reusing data files.
• Geographical data combining points (e.g. geolocated events), polygons (e.g. borders or water shores)
  and/or raster images (e.g. satellite photographs) that can be combined using powerful geographic information system (GIS) software.
• New data sources of increasing size, frequency and diversity (big data) collected organically by sensors (e.g. temperature or traffic measurements) or transaction machines (e.g. scanner or credit card data and mobile calls), with associated technologies for data storage and manipulation.

Evaluators’ interest in big data has grown considerably in recent years, reaching a new peak as a result of the Covid-19 pandemic.