To identify factors related to adequate knowledge of stroke symptoms and risk factors, logistic regression ana- lyses were performed. Candidate variables in the binary model were age, sex, educational level, previous cerebro- vascular disease, having at least 1 risk factor of stroke, living situation, and having received information on stroke at an earlier stage. Variables reaching a significance level of P , .20 were subjected to the multivariate anal- ysis. Stepwise elimination was performed, removing vari- ables with a significance level of P $ .05. Results are presented as odds ratios (OR) and 95% confidence inter- vals (CI).