In late 2005, in Khayelitsha (a large African township in Cape Town, South Africa), crime was perceived to be the dominant hindrance keeping the unemployed from entering self-employment. This finding is robust to a series of alternative methods of ranking one hindrance as more severe than another. Other severe hindrances include the risk of business failure from one unlucky month, a lack of access to start-up capital, high transport costs, and jealousy within the community individuals face if selfemployment is successful. These issues ranked above or on par with profitability concerns and were generally found to dominate concerns over government regulations, intimidation from existing self-employed workers, variable income streams, a lack of specific job skills and/or general business skills, disincentives within the family/kin networks, lowering the probability of getting a better job, and other potential hindrances, even though all were presented on equal terms. In a country facing persistent broad unemployment rates of approximately forty-five percent for Africans, these findings have immediate policy implications. This methodology may also be useful in many developing countries facing high open or disguised unemployment rates or where governments (or practitioners) seek to assist the micro-enterprise sector. The paper presents some of the concerns associated with deriving policy based on these on-the-ground perceptions and addresses other aggregation and statistical issues associated with the Likert scale approach adopted.