If you don't collect some numbers that skews your data away from the truth. If you collect false numbers (many ways to do that) it also skews your data. If you collect numbers but then exclude them in some way that skews results. Thus the first question needs do be do we even have correct valid numbers to work with. If the source data is wrong in anyway then no amount of statistics can correct for that (statistics can correct for sample bias in some situations, but that is different from completely missing or intentionally wrong data by someone who knows how to fool that correction).
Once we have accurate data, then there are a lot of ways to lie with statistics. Is shoplifting a "serious crime" - different people will have different answers. There are many ways to slice up the data we have and if you want to get any particular result you will slice in many different ways until you find a result you like and then work the questions backward to make it seem like they were correct. I'm sure someone better trained in statistics can come up with other ways to make the data show whatever they want.