Among the 246 financial institutions responding, there were 507 robberies from 1968 to 1977, an average of two per institution. To ascertain the characteristics of banking institutions which contribute to their being robbed, three separate methodological approaches were used. Ordinary least squares regressions was used to analyze the rate of robberies per year, and two logit regression models were used to analyze the probability of not being robbed compared to the probability of being robbed one or more times. After controlling for the amount of time a facility had been open, savings and loans had the highest probability of being robbed, followed by banks and then credit unions. Facilities with the higher number of entrances were more likely to be robbed. Facilities with only direct entrances were robbed more often than those with other entrance arrangements, while facilities with only corridor entrances were less likely to be robbed. Facilities with larger numbers of teller stations tended to have higher robbery rates. Overall, the larger, more easily accessed facilities had a higher probability of being robbed. Facilities with security guards tended to be robbed less often, while findings related to the number of visible surveillance cameras were inconclusive. Since the extent of security measures was somewhat confounded with the physical and financial size of the institutions, this study could not provide much useful knowledge on the efficacy of security devices. Suggestions for future research are offered, along with tabular data and four references.