A reanalysis of data obtained in a cross sectional study of the effectiveness of environmental design in deterring robberies of Florida convenience stores was performed. Data obtained in a study of 110 convenience stores that have been robbed between 1984 and 1986 were reviewed. The stores had been visited in 1987 and their environmental designs evaluated. For the reanalysis, the stores were classified into high and low robbery groups. Each store was evaluated according to 14 environmental design factors and five crime risk factors, which included environmental design factors related to store location such as light traffic on an adjacent street and proximity to a major street, the inside and physical characteristics of the store, and store policies. Examples of physical store factors were poor exterior lighting and the opportunity for concealed access or escape routes. The store policy factors involved cash handling policies, number of clerks, and hours of operation. The local crime factors included a high county crime rate, being located in a large county, having few sworn law enforcement officers, having a clearance rate of less than 25%, and being located within a municipal jurisdiction. Associations between environmental design and crime risk factors and the store robbery rates were analyzed. Having a concealed access or escape route, cash register at the back or side of the store, high county crime rate, and a large county population were significantly associated with an increased risk of being robbed, odds ratios (ORs) 9.5, 6.2, 6.3, and 2.8, respectively. Poor cash handling policies were significantly associated with a decreased robbery rate, OR 0.1. A logistic model incorporating all of the factors agreed well with the actual robbery data; the model correctly classified the identified robbery rate of 26% in 70% of the stores. Some confounding was indicated when local crime factors were removed from the model. The authors conclude that studies of the effectiveness of environmental designs for deterring robbery must take into account local crime risk factors.