for example the BRFSS data used in this study measure body weight and height through self-reporting where body weight tends to be undersetimated and height overestimated hence the predictive effectiveness of the statistical analysis of this model is limited as are any conclusions drawn from its regression parameter values finally it must be noted that due to the use of state-level data conclusions can be drawn only about aggregate seatbelt usage at the state level and cannot be directly translated into conclusions about seatbelt usage at the individual level