RACIAL EQUITY DEEP DIVE: Leveraging Machine Learning to Fully Surface Insights From Qualitative Data

Air Force Inspection Agency (AFIA) articulated the need to proactively identify "where there's smoke there's fire" issues among Airmen, their families, and other interested parties that might not be fully realized by leadership. These issues – such as privatized housing on bases and racial disparity – have potential to become significant negatives to the Air Force, airmen, and families. They also may become major topics in Congress or the media if not proactively identified and addressed. We explored possible data sets for a Phase I POC. AFIA identified open text responses to the recent Air Force Racial Disparity Review as a potentially viable candidate. Over 116,000 individuals took this recent survey. While a large majority of the data from this survey is in quantifiable Likert scales, the review did produce 30,000 pages of freeform text responses. With racial equity and inclusion now a major societal focus, it's naturally imperative that Air Force similarly assess its historic, current, and future role in these conversations. This vital work is underway with the aforementioned Air Force Racial Disparity Review. AFIA recognize the importance of a full analysis of all responses, including the vast amount of freeform data that does not lend itself to traditional quantitative or qualitative analysis techniques.