Measures of each concept are developed and evaluated in terms of their ability to explain individual's risks of residential burglary, personal theft, and personal violence. Using data from 5,271 respondents in the British Crime Survey, mixed empirical support is found for the structural-choice model underlying current victimization theories. Specifically, the structural components of these theories (proximity and exposure) are more consistently related to victimization experiences than the target-selection or choice components (attractiveness and guardianship). Multiplicative models that test for interaction among these major variables also produce results that are contrary to theoretical predictions. The authors conclude that present theories do not adequately describe individuals' risks of victimization and discuss some alternative directions for future research on criminal victimization. 3 tables, 11 notes, 33 references. (Author abstract)