|
Abstract:
|
There are numerous grants and funding opportunities available for communities who seek to improve and protect their environmental resources. In practice, unfortunately, funding is often provided to communities with the resources to craft the most strongly phrased proposals rather than the most critical or impactful projects, leading to “winner takes all” biases. This analysis seeks to determine if language choices are contributing to this disparity. In partnership with The Nature Conservancy Maryland/DC Chapter, funding solicitations were collected from several organizations and evaluated using text-mining techniques. Domain experts evaluated solicitations based on their ability to draw equitable proposals. Profiling and categorization techniques were used to identify terms and phrasing unique to “good” and “bad” classifications. These differences in language were then evaluated to determine areas of bias. A model was also created to identify inequitable solicitations. Based on these results, changes are recommended to the way such solicitations are crafted and to the methods used to evaluate the resulting proposals in order to ensure a more equitable distribution of resources.
|
