In this paper we leverage a canonical nine dimensional model of offline risk perception to better understand online risk perceptions. Understanding risk perception facilitates the development of better risk communication and mitigation technologies. We conducted a classic off-line survey to identify the dimensions of online risk perceptions of end users. These results were different from those observed for offline risks. When investigating the original nine dimensional model we found that severity of a risk was the biggest factor in shaping risk perception. We note that technically similar dimensions were not clustered together by participants. We further reduced the nine dimensional model to a four dimensional framework. The variance explained by the original framework was 13.76%. This increased to 83.6% under the reduced model. Under the new model time was the most significant determinant of perceived risk. We found that risks that had a physical analogue were considered more risky and were much better understood. Finally, we discuss the experimental challenges in using offline risk models for online risks.