Using Kappa methodology to consider volunteered geographic information in official land administration systems in developing countries

Abstract

Developing countries have ineffective and outdated land administration systems (LAS) which make it a challenge to properly register, manage and secure tenure for local communities. Moreover, limited maintenance budgets prevalent in these countries make it difficult for organizations to conduct regular systematic updates of geographic information. For a LAS to remain useful, it must reflect realities on the ground, and this can only be achieved if land information is reported regularly. However, if changes in the land are not captured in properly administered land registers, LAS lose societal relevance and are eventually replaced by informal systems. Volunteered Geographic Information (VGI) can address these challenges by providing a timely, affordable, flexible, inclusive, up-to-date, and fit for purpose land information to support the limited current systems. However, the consideration and integration of such external sources relies on the quality and credibility of contributed datasets by the public. The public whom in most cases are non-geographic experts can contribute datasets of varying quality and detail, resulting in the reluctance of officials to utilize VGI in their databases. Therefore, this study proposes a Kappa statistical measure methodology: a consensus-based measure of agreement to establish ‘proxy’ quality and credibility of VGI. The methodology was successfully implemented in a case study of Botswana to demonstrate its effectiveness in establishing trust and confidence of contributed datasets by the public. The innovative approach to Kappa presented here, minimizes the need for ground truthing of VGI data.