A sociospatial model to predict the success of entrepreneurs in Mendocino County, California

Thesis (M.S.)--Humboldt State University, Natural Resources: Natural Resources Planning and Interpretation, 2012Predicting business success in rural areas solely on spatial characteristics is virtually nonexistent in literature. In this study I used logistic regression, geographical information systems (GIS), and spatial variables such as demographic and parcel land-use data to predict the success of entrepreneurs in Mendocino County, California. A total of 115 business surveys collected in 2010 were modeled at two scales, a county-wide coarse-scale model, and a within-city fine-scale, to determine what contributes to rural entrepreneurial success. Significant models were found at both scales, with approximately 80% of businesses correctly classified. Maps were created using GIS to visualize the probability of success over the landscape. Contributing variables to success were proportion of Latino population on census blocks for fine-scale modeling, and parcel acreage, parcel land value, civic engagement point count, and total population on census block for the coarse-scale model