Sure they can but then they're just upside down numbers. Ed Bott posted on his personal blog a link to a post by blogger Long Zheng asking (tongue-in-cheek) if the numbers in the report can be turned on their heads? For example, does the fact that 25% of the sites surveyed in the research try to infect connecting systems with malware mean that the other 75% were clean? Well, yes actually. The question then becomes for anyone willing to take those odds, how do you tell before you connect to a site whether it's bad or not? In other words, you have a three in four chance of getting pirated software and a one in four chance of getting attacked. Hmmm, lemme think...