Methods
This was a cross sectional study of a sample of female students grade four in Rasht (a
metropolitan in north Iran) in 2012. A systematic random sampling method was applied to
recruit students. They were asked to respond to a designed questionnaire containing items on
brushing behavior based on the health belief model. In this study for the first time perceived
severity and perceived barriers were divided into two parts, perceived subjective and
objective severity and perceived physical and psychological barriers and were treated as
independent variables. Logistic regression analysis was performed in order to identify the
variables that predict the desirable behavior (brushing twice a day or more).